diff options
Diffstat (limited to 'parse.y')
| -rw-r--r-- | parse.y | 15517 |
1 files changed, 11620 insertions, 3897 deletions
@@ -3,314 +3,1295 @@ parse.y - $Author$ - $Date$ created at: Fri May 28 18:02:42 JST 1993 - Copyright (C) 1993-2003 Yukihiro Matsumoto + Copyright (C) 1993-2007 Yukihiro Matsumoto **********************************************************************/ %{ +#if !YYPURE +# error needs pure parser +#endif #define YYDEBUG 1 #define YYERROR_VERBOSE 1 -#ifndef YYSTACK_USE_ALLOCA #define YYSTACK_USE_ALLOCA 0 -#endif +#define YYLTYPE rb_code_location_t +#define YYLTYPE_IS_DECLARED 1 + +#include "ruby/internal/config.h" -#include "ruby.h" -#include "env.h" -#include "intern.h" -#include "node.h" -#include "st.h" -#include <stdio.h> -#include <errno.h> #include <ctype.h> +#include <errno.h> +#include <stdio.h> -#define YYMALLOC rb_parser_malloc -#define YYREALLOC rb_parser_realloc -#define YYCALLOC rb_parser_calloc -#define YYFREE rb_parser_free -#define malloc YYMALLOC -#define realloc YYREALLOC -#define calloc YYCALLOC -#define free YYFREE -static void *rb_parser_malloc _((size_t)); -static void *rb_parser_realloc _((void *, size_t)); -static void *rb_parser_calloc _((size_t, size_t)); -static void rb_parser_free _((void *)); +struct lex_context; + +#include "internal.h" +#include "internal/compile.h" +#include "internal/compilers.h" +#include "internal/complex.h" +#include "internal/error.h" +#include "internal/hash.h" +#include "internal/imemo.h" +#include "internal/io.h" +#include "internal/numeric.h" +#include "internal/parse.h" +#include "internal/rational.h" +#include "internal/re.h" +#include "internal/symbol.h" +#include "internal/thread.h" +#include "internal/variable.h" +#include "node.h" +#include "probes.h" +#include "regenc.h" +#include "ruby/encoding.h" +#include "ruby/regex.h" +#include "ruby/ruby.h" +#include "ruby/st.h" +#include "ruby/util.h" +#include "ruby/ractor.h" +#include "symbol.h" + +enum shareability { + shareable_none, + shareable_literal, + shareable_copy, + shareable_everything, +}; -#define yyparse ruby_yyparse -#define yylex ruby_yylex -#define yyerror ruby_yyerror -#define yylval ruby_yylval -#define yychar ruby_yychar -#define yydebug ruby_yydebug - -#define ID_SCOPE_SHIFT 3 -#define ID_SCOPE_MASK 0x07 -#define ID_LOCAL 0x01 -#define ID_INSTANCE 0x02 -#define ID_GLOBAL 0x03 -#define ID_ATTRSET 0x04 -#define ID_CONST 0x05 -#define ID_CLASS 0x06 -#define ID_JUNK 0x07 -#define ID_INTERNAL ID_JUNK - -#define is_notop_id(id) ((id)>tLAST_TOKEN) -#define is_local_id(id) (is_notop_id(id)&&((id)&ID_SCOPE_MASK)==ID_LOCAL) -#define is_global_id(id) (is_notop_id(id)&&((id)&ID_SCOPE_MASK)==ID_GLOBAL) -#define is_instance_id(id) (is_notop_id(id)&&((id)&ID_SCOPE_MASK)==ID_INSTANCE) -#define is_attrset_id(id) (is_notop_id(id)&&((id)&ID_SCOPE_MASK)==ID_ATTRSET) -#define is_const_id(id) (is_notop_id(id)&&((id)&ID_SCOPE_MASK)==ID_CONST) -#define is_class_id(id) (is_notop_id(id)&&((id)&ID_SCOPE_MASK)==ID_CLASS) -#define is_junk_id(id) (is_notop_id(id)&&((id)&ID_SCOPE_MASK)==ID_JUNK) - -#define is_asgn_or_id(id) ((is_notop_id(id)) && \ - (((id)&ID_SCOPE_MASK) == ID_GLOBAL || \ - ((id)&ID_SCOPE_MASK) == ID_INSTANCE || \ - ((id)&ID_SCOPE_MASK) == ID_CLASS)) - -NODE *ruby_eval_tree_begin = 0; -NODE *ruby_eval_tree = 0; - -char *ruby_sourcefile; /* current source file */ -int ruby_sourceline; /* current line no. */ - -static int yylex(); -static int yyerror(); - -static enum lex_state { - EXPR_BEG, /* ignore newline, +/- is a sign. */ - EXPR_END, /* newline significant, +/- is an operator. */ - EXPR_ARG, /* newline significant, +/- is an operator. */ - EXPR_CMDARG, /* newline significant, +/- is an operator. */ - EXPR_ENDARG, /* newline significant, +/- is an operator. */ - EXPR_MID, /* newline significant, +/- is an operator. */ - EXPR_FNAME, /* ignore newline, no reserved words. */ - EXPR_DOT, /* right after `.' or `::', no reserved words. */ - EXPR_CLASS, /* immediate after `class', no here document. */ -} lex_state; -static NODE *lex_strterm; - -#ifdef HAVE_LONG_LONG -typedef unsigned LONG_LONG stack_type; -#else -typedef unsigned long stack_type; +struct lex_context { + unsigned int in_defined: 1; + unsigned int in_kwarg: 1; + unsigned int in_argdef: 1; + unsigned int in_def: 1; + unsigned int in_class: 1; + BITFIELD(enum shareability, shareable_constant_value, 2); +}; + +#include "parse.h" + +#define NO_LEX_CTXT (struct lex_context){0} + +#define AREF(ary, i) RARRAY_AREF(ary, i) + +#ifndef WARN_PAST_SCOPE +# define WARN_PAST_SCOPE 0 #endif -#define BITSTACK_PUSH(stack, n) (stack = (stack<<1)|((n)&1)) -#define BITSTACK_POP(stack) (stack >>= 1) -#define BITSTACK_LEXPOP(stack) (stack = (stack >> 1) | (stack & 1)) -#define BITSTACK_SET_P(stack) (stack&1) +#define TAB_WIDTH 8 + +#define yydebug (p->debug) /* disable the global variable definition */ + +#define YYMALLOC(size) rb_parser_malloc(p, (size)) +#define YYREALLOC(ptr, size) rb_parser_realloc(p, (ptr), (size)) +#define YYCALLOC(nelem, size) rb_parser_calloc(p, (nelem), (size)) +#define YYFREE(ptr) rb_parser_free(p, (ptr)) +#define YYFPRINTF rb_parser_printf +#define YY_LOCATION_PRINT(File, loc) \ + rb_parser_printf(p, "%d.%d-%d.%d", \ + (loc).beg_pos.lineno, (loc).beg_pos.column,\ + (loc).end_pos.lineno, (loc).end_pos.column) +#define YYLLOC_DEFAULT(Current, Rhs, N) \ + do \ + if (N) \ + { \ + (Current).beg_pos = YYRHSLOC(Rhs, 1).beg_pos; \ + (Current).end_pos = YYRHSLOC(Rhs, N).end_pos; \ + } \ + else \ + { \ + (Current).beg_pos = YYRHSLOC(Rhs, 0).end_pos; \ + (Current).end_pos = YYRHSLOC(Rhs, 0).end_pos; \ + } \ + while (0) +#define YY_(Msgid) \ + (((Msgid)[0] == 'm') && (strcmp((Msgid), "memory exhausted") == 0) ? \ + "nesting too deep" : (Msgid)) + +#define RUBY_SET_YYLLOC_FROM_STRTERM_HEREDOC(Current) \ + rb_parser_set_location_from_strterm_heredoc(p, &p->lex.strterm->u.heredoc, &(Current)) +#define RUBY_SET_YYLLOC_OF_NONE(Current) \ + rb_parser_set_location_of_none(p, &(Current)) +#define RUBY_SET_YYLLOC(Current) \ + rb_parser_set_location(p, &(Current)) +#define RUBY_INIT_YYLLOC() \ + { \ + {p->ruby_sourceline, (int)(p->lex.ptok - p->lex.pbeg)}, \ + {p->ruby_sourceline, (int)(p->lex.pcur - p->lex.pbeg)}, \ + } + +enum lex_state_bits { + EXPR_BEG_bit, /* ignore newline, +/- is a sign. */ + EXPR_END_bit, /* newline significant, +/- is an operator. */ + EXPR_ENDARG_bit, /* ditto, and unbound braces. */ + EXPR_ENDFN_bit, /* ditto, and unbound braces. */ + EXPR_ARG_bit, /* newline significant, +/- is an operator. */ + EXPR_CMDARG_bit, /* newline significant, +/- is an operator. */ + EXPR_MID_bit, /* newline significant, +/- is an operator. */ + EXPR_FNAME_bit, /* ignore newline, no reserved words. */ + EXPR_DOT_bit, /* right after `.' or `::', no reserved words. */ + EXPR_CLASS_bit, /* immediate after `class', no here document. */ + EXPR_LABEL_bit, /* flag bit, label is allowed. */ + EXPR_LABELED_bit, /* flag bit, just after a label. */ + EXPR_FITEM_bit, /* symbol literal as FNAME. */ + EXPR_MAX_STATE +}; +/* examine combinations */ +enum lex_state_e { +#define DEF_EXPR(n) EXPR_##n = (1 << EXPR_##n##_bit) + DEF_EXPR(BEG), + DEF_EXPR(END), + DEF_EXPR(ENDARG), + DEF_EXPR(ENDFN), + DEF_EXPR(ARG), + DEF_EXPR(CMDARG), + DEF_EXPR(MID), + DEF_EXPR(FNAME), + DEF_EXPR(DOT), + DEF_EXPR(CLASS), + DEF_EXPR(LABEL), + DEF_EXPR(LABELED), + DEF_EXPR(FITEM), + EXPR_VALUE = EXPR_BEG, + EXPR_BEG_ANY = (EXPR_BEG | EXPR_MID | EXPR_CLASS), + EXPR_ARG_ANY = (EXPR_ARG | EXPR_CMDARG), + EXPR_END_ANY = (EXPR_END | EXPR_ENDARG | EXPR_ENDFN), + EXPR_NONE = 0 +}; +#define IS_lex_state_for(x, ls) ((x) & (ls)) +#define IS_lex_state_all_for(x, ls) (((x) & (ls)) == (ls)) +#define IS_lex_state(ls) IS_lex_state_for(p->lex.state, (ls)) +#define IS_lex_state_all(ls) IS_lex_state_all_for(p->lex.state, (ls)) + +# define SET_LEX_STATE(ls) \ + parser_set_lex_state(p, ls, __LINE__) +static inline enum lex_state_e parser_set_lex_state(struct parser_params *p, enum lex_state_e ls, int line); + +typedef VALUE stack_type; -static stack_type cond_stack = 0; -#define COND_PUSH(n) BITSTACK_PUSH(cond_stack, n) +static const rb_code_location_t NULL_LOC = { {0, -1}, {0, -1} }; + +# define SHOW_BITSTACK(stack, name) (p->debug ? rb_parser_show_bitstack(p, stack, name, __LINE__) : (void)0) +# define BITSTACK_PUSH(stack, n) (((p->stack) = ((p->stack)<<1)|((n)&1)), SHOW_BITSTACK(p->stack, #stack"(push)")) +# define BITSTACK_POP(stack) (((p->stack) = (p->stack) >> 1), SHOW_BITSTACK(p->stack, #stack"(pop)")) +# define BITSTACK_SET_P(stack) (SHOW_BITSTACK(p->stack, #stack), (p->stack)&1) +# define BITSTACK_SET(stack, n) ((p->stack)=(n), SHOW_BITSTACK(p->stack, #stack"(set)")) + +/* A flag to identify keyword_do_cond, "do" keyword after condition expression. + Examples: `while ... do`, `until ... do`, and `for ... in ... do` */ +#define COND_PUSH(n) BITSTACK_PUSH(cond_stack, (n)) #define COND_POP() BITSTACK_POP(cond_stack) -#define COND_LEXPOP() BITSTACK_LEXPOP(cond_stack) #define COND_P() BITSTACK_SET_P(cond_stack) +#define COND_SET(n) BITSTACK_SET(cond_stack, (n)) -static stack_type cmdarg_stack = 0; -#define CMDARG_PUSH(n) BITSTACK_PUSH(cmdarg_stack, n) +/* A flag to identify keyword_do_block; "do" keyword after command_call. + Example: `foo 1, 2 do`. */ +#define CMDARG_PUSH(n) BITSTACK_PUSH(cmdarg_stack, (n)) #define CMDARG_POP() BITSTACK_POP(cmdarg_stack) -#define CMDARG_LEXPOP() BITSTACK_LEXPOP(cmdarg_stack) #define CMDARG_P() BITSTACK_SET_P(cmdarg_stack) +#define CMDARG_SET(n) BITSTACK_SET(cmdarg_stack, (n)) + +struct vtable { + ID *tbl; + int pos; + int capa; + struct vtable *prev; +}; + +struct local_vars { + struct vtable *args; + struct vtable *vars; + struct vtable *used; +# if WARN_PAST_SCOPE + struct vtable *past; +# endif + struct local_vars *prev; +# ifndef RIPPER + struct { + NODE *outer, *inner, *current; + } numparam; +# endif +}; -static int class_nest = 0; -static int in_single = 0; -static int in_def = 0; -static int compile_for_eval = 0; -static ID cur_mid = 0; -static int command_start = Qtrue; - -static NODE *cond(); -static NODE *logop(); -static int cond_negative(); - -static NODE *newline_node(); -static void fixpos(); - -static int value_expr0(); -static void void_expr0(); -static void void_stmts(); -static NODE *remove_begin(); -#define value_expr(node) value_expr0((node) = remove_begin(node)) -#define void_expr(node) void_expr0((node) = remove_begin(node)) - -static NODE *block_append(); -static NODE *list_append(); -static NODE *list_concat(); -static NODE *arg_concat(); -static NODE *arg_prepend(); -static NODE *literal_concat(); -static NODE *new_evstr(); -static NODE *evstr2dstr(); -static NODE *call_op(); -static int in_defined = 0; - -static NODE *negate_lit(); -static NODE *ret_args(); -static NODE *arg_blk_pass(); -static NODE *new_call(); -static NODE *new_fcall(); -static NODE *new_super(); -static NODE *new_yield(); - -static NODE *gettable(); -static NODE *assignable(); -static NODE *aryset(); -static NODE *attrset(); -static void rb_backref_error(); -static NODE *node_assign(); - -static NODE *match_gen(); -static void local_push(); -static void local_pop(); -static int local_append(); -static int local_cnt(); -static int local_id(); -static ID *local_tbl(); -static ID internal_id(); - -static struct RVarmap *dyna_push(); -static void dyna_pop(); -static int dyna_in_block(); -static NODE *dyna_init(); - -static void top_local_init(); -static void top_local_setup(); - -#define RE_OPTION_ONCE 0x80 - -#define NODE_STRTERM NODE_ZARRAY /* nothing to gc */ -#define NODE_HEREDOC NODE_ARRAY /* 1, 3 to gc */ -#define SIGN_EXTEND(x,n) (((1<<(n)-1)^((x)&~(~0<<(n))))-(1<<(n)-1)) -#define nd_func u1.id -#if SIZEOF_SHORT == 2 -#define nd_term(node) ((signed short)(node)->u2.id) +enum { + ORDINAL_PARAM = -1, + NO_PARAM = 0, + NUMPARAM_MAX = 9, +}; + +#define NUMPARAM_ID_P(id) numparam_id_p(id) +#define NUMPARAM_ID_TO_IDX(id) (unsigned int)(((id) >> ID_SCOPE_SHIFT) - tNUMPARAM_1 + 1) +#define NUMPARAM_IDX_TO_ID(idx) TOKEN2LOCALID((tNUMPARAM_1 + (idx) - 1)) +static int +numparam_id_p(ID id) +{ + if (!is_local_id(id)) return 0; + unsigned int idx = NUMPARAM_ID_TO_IDX(id); + return idx > 0 && idx <= NUMPARAM_MAX; +} +static void numparam_name(struct parser_params *p, ID id); + +#define DVARS_INHERIT ((void*)1) +#define DVARS_TOPSCOPE NULL +#define DVARS_TERMINAL_P(tbl) ((tbl) == DVARS_INHERIT || (tbl) == DVARS_TOPSCOPE) + +typedef struct token_info { + const char *token; + rb_code_position_t beg; + int indent; + int nonspc; + struct token_info *next; +} token_info; + +typedef struct rb_strterm_struct rb_strterm_t; + +/* + Structure of Lexer Buffer: + + lex.pbeg lex.ptok lex.pcur lex.pend + | | | | + |------------+------------+------------| + |<---------->| + token +*/ +struct parser_params { + rb_imemo_tmpbuf_t *heap; + + YYSTYPE *lval; + + struct { + rb_strterm_t *strterm; + VALUE (*gets)(struct parser_params*,VALUE); + VALUE input; + VALUE prevline; + VALUE lastline; + VALUE nextline; + const char *pbeg; + const char *pcur; + const char *pend; + const char *ptok; + union { + long ptr; + VALUE (*call)(VALUE, int); + } gets_; + enum lex_state_e state; + /* track the nest level of any parens "()[]{}" */ + int paren_nest; + /* keep p->lex.paren_nest at the beginning of lambda "->" to detect tLAMBEG and keyword_do_LAMBDA */ + int lpar_beg; + /* track the nest level of only braces "{}" */ + int brace_nest; + } lex; + stack_type cond_stack; + stack_type cmdarg_stack; + int tokidx; + int toksiz; + int tokline; + int heredoc_end; + int heredoc_indent; + int heredoc_line_indent; + char *tokenbuf; + struct local_vars *lvtbl; + st_table *pvtbl; + st_table *pktbl; + int line_count; + int ruby_sourceline; /* current line no. */ + const char *ruby_sourcefile; /* current source file */ + VALUE ruby_sourcefile_string; + rb_encoding *enc; + token_info *token_info; + VALUE case_labels; + VALUE compile_option; + + VALUE debug_buffer; + VALUE debug_output; + + ID cur_arg; + + rb_ast_t *ast; + int node_id; + + int max_numparam; + + struct lex_context ctxt; + + unsigned int command_start:1; + unsigned int eofp: 1; + unsigned int ruby__end__seen: 1; + unsigned int debug: 1; + unsigned int has_shebang: 1; + unsigned int token_seen: 1; + unsigned int token_info_enabled: 1; +# if WARN_PAST_SCOPE + unsigned int past_scope_enabled: 1; +# endif + unsigned int error_p: 1; + unsigned int cr_seen: 1; + +#ifndef RIPPER + /* Ruby core only */ + + unsigned int do_print: 1; + unsigned int do_loop: 1; + unsigned int do_chomp: 1; + unsigned int do_split: 1; + unsigned int keep_script_lines: 1; + + NODE *eval_tree_begin; + NODE *eval_tree; + VALUE error_buffer; + VALUE debug_lines; + const struct rb_iseq_struct *parent_iseq; #else -#define nd_term(node) SIGN_EXTEND((node)->u2.id, CHAR_BIT*2) + /* Ripper only */ + + struct { + VALUE token; + int line; + int col; + } delayed; + + VALUE value; + VALUE result; + VALUE parsing_thread; #endif -#define nd_paren(node) (char)((node)->u2.id >> CHAR_BIT*2) -#define nd_nest u3.id - -/* Older versions of Yacc set YYMAXDEPTH to a very low value by default (150, - for instance). This is too low for Ruby to parse some files, such as - date/format.rb, therefore bump the value up to at least Bison's default. */ -#ifdef OLD_YACC -#ifndef YYMAXDEPTH -#define YYMAXDEPTH 10000 +}; + +#define intern_cstr(n,l,en) rb_intern3(n,l,en) + +#define STR_NEW(ptr,len) rb_enc_str_new((ptr),(len),p->enc) +#define STR_NEW0() rb_enc_str_new(0,0,p->enc) +#define STR_NEW2(ptr) rb_enc_str_new((ptr),strlen(ptr),p->enc) +#define STR_NEW3(ptr,len,e,func) parser_str_new((ptr),(len),(e),(func),p->enc) +#define TOK_INTERN() intern_cstr(tok(p), toklen(p), p->enc) + +static st_table * +push_pvtbl(struct parser_params *p) +{ + st_table *tbl = p->pvtbl; + p->pvtbl = st_init_numtable(); + return tbl; +} + +static void +pop_pvtbl(struct parser_params *p, st_table *tbl) +{ + st_free_table(p->pvtbl); + p->pvtbl = tbl; +} + +static st_table * +push_pktbl(struct parser_params *p) +{ + st_table *tbl = p->pktbl; + p->pktbl = 0; + return tbl; +} + +static void +pop_pktbl(struct parser_params *p, st_table *tbl) +{ + if (p->pktbl) st_free_table(p->pktbl); + p->pktbl = tbl; +} + +RBIMPL_ATTR_NONNULL((1, 2, 3)) +static int parser_yyerror(struct parser_params*, const YYLTYPE *yylloc, const char*); +RBIMPL_ATTR_NONNULL((1, 2)) +static int parser_yyerror0(struct parser_params*, const char*); +#define yyerror0(msg) parser_yyerror0(p, (msg)) +#define yyerror1(loc, msg) parser_yyerror(p, (loc), (msg)) +#define yyerror(yylloc, p, msg) parser_yyerror(p, yylloc, msg) +#define token_flush(ptr) ((ptr)->lex.ptok = (ptr)->lex.pcur) + +static void token_info_setup(token_info *ptinfo, const char *ptr, const rb_code_location_t *loc); +static void token_info_push(struct parser_params*, const char *token, const rb_code_location_t *loc); +static void token_info_pop(struct parser_params*, const char *token, const rb_code_location_t *loc); +static void token_info_warn(struct parser_params *p, const char *token, token_info *ptinfo_beg, int same, const rb_code_location_t *loc); +static void token_info_drop(struct parser_params *p, const char *token, rb_code_position_t beg_pos); + +#ifdef RIPPER +#define compile_for_eval (0) +#else +#define compile_for_eval (p->parent_iseq != 0) #endif + +#define token_column ((int)(p->lex.ptok - p->lex.pbeg)) + +#define CALL_Q_P(q) ((q) == TOKEN2VAL(tANDDOT)) +#define NODE_CALL_Q(q) (CALL_Q_P(q) ? NODE_QCALL : NODE_CALL) +#define NEW_QCALL(q,r,m,a,loc) NEW_NODE(NODE_CALL_Q(q),r,m,a,loc) + +#define lambda_beginning_p() (p->lex.lpar_beg == p->lex.paren_nest) + +#define ANON_BLOCK_ID '&' + +static enum yytokentype yylex(YYSTYPE*, YYLTYPE*, struct parser_params*); + +#ifndef RIPPER +static inline void +rb_discard_node(struct parser_params *p, NODE *n) +{ + rb_ast_delete_node(p->ast, n); +} +#endif + +#ifdef RIPPER +static inline VALUE +add_mark_object(struct parser_params *p, VALUE obj) +{ + if (!SPECIAL_CONST_P(obj) + && !RB_TYPE_P(obj, T_NODE) /* Ripper jumbles NODE objects and other objects... */ + ) { + rb_ast_add_mark_object(p->ast, obj); + } + return obj; +} +#else +static NODE* node_newnode_with_locals(struct parser_params *, enum node_type, VALUE, VALUE, const rb_code_location_t*); +#endif + +static NODE* node_newnode(struct parser_params *, enum node_type, VALUE, VALUE, VALUE, const rb_code_location_t*); +#define rb_node_newnode(type, a1, a2, a3, loc) node_newnode(p, (type), (a1), (a2), (a3), (loc)) + +static NODE *nd_set_loc(NODE *nd, const YYLTYPE *loc); + +static int +parser_get_node_id(struct parser_params *p) +{ + int node_id = p->node_id; + p->node_id++; + return node_id; +} + +#ifndef RIPPER +static inline void +set_line_body(NODE *body, int line) +{ + if (!body) return; + switch (nd_type(body)) { + case NODE_RESCUE: + case NODE_ENSURE: + nd_set_line(body, line); + } +} + +#define yyparse ruby_yyparse + +static NODE* cond(struct parser_params *p, NODE *node, const YYLTYPE *loc); +static NODE* method_cond(struct parser_params *p, NODE *node, const YYLTYPE *loc); +#define new_nil(loc) NEW_NIL(loc) +static NODE *new_nil_at(struct parser_params *p, const rb_code_position_t *pos); +static NODE *new_if(struct parser_params*,NODE*,NODE*,NODE*,const YYLTYPE*); +static NODE *new_unless(struct parser_params*,NODE*,NODE*,NODE*,const YYLTYPE*); +static NODE *logop(struct parser_params*,ID,NODE*,NODE*,const YYLTYPE*,const YYLTYPE*); + +static NODE *newline_node(NODE*); +static void fixpos(NODE*,NODE*); + +static int value_expr_gen(struct parser_params*,NODE*); +static void void_expr(struct parser_params*,NODE*); +static NODE *remove_begin(NODE*); +static NODE *remove_begin_all(NODE*); +#define value_expr(node) value_expr_gen(p, (node)) +static NODE *void_stmts(struct parser_params*,NODE*); +static void reduce_nodes(struct parser_params*,NODE**); +static void block_dup_check(struct parser_params*,NODE*,NODE*); + +static NODE *block_append(struct parser_params*,NODE*,NODE*); +static NODE *list_append(struct parser_params*,NODE*,NODE*); +static NODE *list_concat(NODE*,NODE*); +static NODE *arg_append(struct parser_params*,NODE*,NODE*,const YYLTYPE*); +static NODE *last_arg_append(struct parser_params *p, NODE *args, NODE *last_arg, const YYLTYPE *loc); +static NODE *rest_arg_append(struct parser_params *p, NODE *args, NODE *rest_arg, const YYLTYPE *loc); +static NODE *literal_concat(struct parser_params*,NODE*,NODE*,const YYLTYPE*); +static NODE *new_evstr(struct parser_params*,NODE*,const YYLTYPE*); +static NODE *new_dstr(struct parser_params*,NODE*,const YYLTYPE*); +static NODE *evstr2dstr(struct parser_params*,NODE*); +static NODE *splat_array(NODE*); +static void mark_lvar_used(struct parser_params *p, NODE *rhs); + +static NODE *call_bin_op(struct parser_params*,NODE*,ID,NODE*,const YYLTYPE*,const YYLTYPE*); +static NODE *call_uni_op(struct parser_params*,NODE*,ID,const YYLTYPE*,const YYLTYPE*); +static NODE *new_qcall(struct parser_params* p, ID atype, NODE *recv, ID mid, NODE *args, const YYLTYPE *op_loc, const YYLTYPE *loc); +static NODE *new_command_qcall(struct parser_params* p, ID atype, NODE *recv, ID mid, NODE *args, NODE *block, const YYLTYPE *op_loc, const YYLTYPE *loc); +static NODE *method_add_block(struct parser_params*p, NODE *m, NODE *b, const YYLTYPE *loc) {b->nd_iter = m; b->nd_loc = *loc; return b;} + +static bool args_info_empty_p(struct rb_args_info *args); +static NODE *new_args(struct parser_params*,NODE*,NODE*,ID,NODE*,NODE*,const YYLTYPE*); +static NODE *new_args_tail(struct parser_params*,NODE*,ID,ID,const YYLTYPE*); +static NODE *new_array_pattern(struct parser_params *p, NODE *constant, NODE *pre_arg, NODE *aryptn, const YYLTYPE *loc); +static NODE *new_array_pattern_tail(struct parser_params *p, NODE *pre_args, int has_rest, ID rest_arg, NODE *post_args, const YYLTYPE *loc); +static NODE *new_find_pattern(struct parser_params *p, NODE *constant, NODE *fndptn, const YYLTYPE *loc); +static NODE *new_find_pattern_tail(struct parser_params *p, ID pre_rest_arg, NODE *args, ID post_rest_arg, const YYLTYPE *loc); +static NODE *new_hash_pattern(struct parser_params *p, NODE *constant, NODE *hshptn, const YYLTYPE *loc); +static NODE *new_hash_pattern_tail(struct parser_params *p, NODE *kw_args, ID kw_rest_arg, const YYLTYPE *loc); + +static NODE *new_kw_arg(struct parser_params *p, NODE *k, const YYLTYPE *loc); +static NODE *args_with_numbered(struct parser_params*,NODE*,int); + +static VALUE negate_lit(struct parser_params*, VALUE); +static NODE *ret_args(struct parser_params*,NODE*); +static NODE *arg_blk_pass(NODE*,NODE*); +static NODE *new_yield(struct parser_params*,NODE*,const YYLTYPE*); +static NODE *dsym_node(struct parser_params*,NODE*,const YYLTYPE*); + +static NODE *gettable(struct parser_params*,ID,const YYLTYPE*); +static NODE *assignable(struct parser_params*,ID,NODE*,const YYLTYPE*); + +static NODE *aryset(struct parser_params*,NODE*,NODE*,const YYLTYPE*); +static NODE *attrset(struct parser_params*,NODE*,ID,ID,const YYLTYPE*); + +static void rb_backref_error(struct parser_params*,NODE*); +static NODE *node_assign(struct parser_params*,NODE*,NODE*,struct lex_context,const YYLTYPE*); + +static NODE *new_op_assign(struct parser_params *p, NODE *lhs, ID op, NODE *rhs, struct lex_context, const YYLTYPE *loc); +static NODE *new_ary_op_assign(struct parser_params *p, NODE *ary, NODE *args, ID op, NODE *rhs, const YYLTYPE *args_loc, const YYLTYPE *loc); +static NODE *new_attr_op_assign(struct parser_params *p, NODE *lhs, ID atype, ID attr, ID op, NODE *rhs, const YYLTYPE *loc); +static NODE *new_const_op_assign(struct parser_params *p, NODE *lhs, ID op, NODE *rhs, struct lex_context, const YYLTYPE *loc); +static NODE *new_bodystmt(struct parser_params *p, NODE *head, NODE *rescue, NODE *rescue_else, NODE *ensure, const YYLTYPE *loc); + +static NODE *const_decl(struct parser_params *p, NODE* path, const YYLTYPE *loc); + +static NODE *opt_arg_append(NODE*, NODE*); +static NODE *kwd_append(NODE*, NODE*); + +static NODE *new_hash(struct parser_params *p, NODE *hash, const YYLTYPE *loc); +static NODE *new_unique_key_hash(struct parser_params *p, NODE *hash, const YYLTYPE *loc); + +static NODE *new_defined(struct parser_params *p, NODE *expr, const YYLTYPE *loc); + +static NODE *new_regexp(struct parser_params *, NODE *, int, const YYLTYPE *); + +#define make_list(list, loc) ((list) ? (nd_set_loc(list, loc), list) : NEW_ZLIST(loc)) + +static NODE *new_xstring(struct parser_params *, NODE *, const YYLTYPE *loc); + +static NODE *symbol_append(struct parser_params *p, NODE *symbols, NODE *symbol); + +static NODE *match_op(struct parser_params*,NODE*,NODE*,const YYLTYPE*,const YYLTYPE*); + +static rb_ast_id_table_t *local_tbl(struct parser_params*); + +static VALUE reg_compile(struct parser_params*, VALUE, int); +static void reg_fragment_setenc(struct parser_params*, VALUE, int); +static int reg_fragment_check(struct parser_params*, VALUE, int); +static NODE *reg_named_capture_assign(struct parser_params* p, VALUE regexp, const YYLTYPE *loc); + +static int literal_concat0(struct parser_params *p, VALUE head, VALUE tail); +static NODE *heredoc_dedent(struct parser_params*,NODE*); + +static void check_literal_when(struct parser_params *p, NODE *args, const YYLTYPE *loc); + +#define get_id(id) (id) +#define get_value(val) (val) +#define get_num(num) (num) +#else /* RIPPER */ +#define NODE_RIPPER NODE_CDECL +#define NEW_RIPPER(a,b,c,loc) (VALUE)NEW_CDECL(a,b,c,loc) + +static inline int ripper_is_node_yylval(VALUE n); + +static inline VALUE +ripper_new_yylval(struct parser_params *p, ID a, VALUE b, VALUE c) +{ + if (ripper_is_node_yylval(c)) c = RNODE(c)->nd_cval; + add_mark_object(p, b); + add_mark_object(p, c); + return NEW_RIPPER(a, b, c, &NULL_LOC); +} + +static inline int +ripper_is_node_yylval(VALUE n) +{ + return RB_TYPE_P(n, T_NODE) && nd_type_p(RNODE(n), NODE_RIPPER); +} + +#define value_expr(node) ((void)(node)) +#define remove_begin(node) (node) +#define void_stmts(p,x) (x) +#define rb_dvar_defined(id, base) 0 +#define rb_local_defined(id, base) 0 +static ID ripper_get_id(VALUE); +#define get_id(id) ripper_get_id(id) +static VALUE ripper_get_value(VALUE); +#define get_value(val) ripper_get_value(val) +#define get_num(num) (int)get_id(num) +static VALUE assignable(struct parser_params*,VALUE); +static int id_is_var(struct parser_params *p, ID id); + +#define method_cond(p,node,loc) (node) +#define call_bin_op(p, recv,id,arg1,op_loc,loc) dispatch3(binary, (recv), STATIC_ID2SYM(id), (arg1)) +#define match_op(p,node1,node2,op_loc,loc) call_bin_op(0, (node1), idEqTilde, (node2), op_loc, loc) +#define call_uni_op(p, recv,id,op_loc,loc) dispatch2(unary, STATIC_ID2SYM(id), (recv)) +#define logop(p,id,node1,node2,op_loc,loc) call_bin_op(0, (node1), (id), (node2), op_loc, loc) + +#define new_nil(loc) Qnil + +static VALUE new_regexp(struct parser_params *, VALUE, VALUE, const YYLTYPE *); + +static VALUE const_decl(struct parser_params *p, VALUE path); + +static VALUE var_field(struct parser_params *p, VALUE a); +static VALUE assign_error(struct parser_params *p, const char *mesg, VALUE a); + +static VALUE parser_reg_compile(struct parser_params*, VALUE, int, VALUE *); + +static VALUE backref_error(struct parser_params*, NODE *, VALUE); +#endif /* !RIPPER */ + +/* forward declaration */ +typedef struct rb_strterm_heredoc_struct rb_strterm_heredoc_t; + +RUBY_SYMBOL_EXPORT_BEGIN +VALUE rb_parser_reg_compile(struct parser_params* p, VALUE str, int options); +int rb_reg_fragment_setenc(struct parser_params*, VALUE, int); +enum lex_state_e rb_parser_trace_lex_state(struct parser_params *, enum lex_state_e, enum lex_state_e, int); +VALUE rb_parser_lex_state_name(enum lex_state_e state); +void rb_parser_show_bitstack(struct parser_params *, stack_type, const char *, int); +PRINTF_ARGS(void rb_parser_fatal(struct parser_params *p, const char *fmt, ...), 2, 3); +YYLTYPE *rb_parser_set_location_from_strterm_heredoc(struct parser_params *p, rb_strterm_heredoc_t *here, YYLTYPE *yylloc); +YYLTYPE *rb_parser_set_location_of_none(struct parser_params *p, YYLTYPE *yylloc); +YYLTYPE *rb_parser_set_location(struct parser_params *p, YYLTYPE *yylloc); +RUBY_SYMBOL_EXPORT_END + +static void error_duplicate_pattern_variable(struct parser_params *p, ID id, const YYLTYPE *loc); +static void error_duplicate_pattern_key(struct parser_params *p, ID id, const YYLTYPE *loc); +#ifndef RIPPER +static ID formal_argument(struct parser_params*, ID); +#else +static ID formal_argument(struct parser_params*, VALUE); #endif +static ID shadowing_lvar(struct parser_params*,ID); +static void new_bv(struct parser_params*,ID); + +static void local_push(struct parser_params*,int); +static void local_pop(struct parser_params*); +static void local_var(struct parser_params*, ID); +static void arg_var(struct parser_params*, ID); +static int local_id(struct parser_params *p, ID id); +static int local_id_ref(struct parser_params*, ID, ID **); +#ifndef RIPPER +static ID internal_id(struct parser_params*); +static NODE *new_args_forward_call(struct parser_params*, NODE*, const YYLTYPE*, const YYLTYPE*); +#endif +static int check_forwarding_args(struct parser_params*); +static void add_forwarding_args(struct parser_params *p); + +static const struct vtable *dyna_push(struct parser_params *); +static void dyna_pop(struct parser_params*, const struct vtable *); +static int dyna_in_block(struct parser_params*); +#define dyna_var(p, id) local_var(p, id) +static int dvar_defined(struct parser_params*, ID); +static int dvar_defined_ref(struct parser_params*, ID, ID**); +static int dvar_curr(struct parser_params*,ID); + +static int lvar_defined(struct parser_params*, ID); + +static NODE *numparam_push(struct parser_params *p); +static void numparam_pop(struct parser_params *p, NODE *prev_inner); + +#ifdef RIPPER +# define METHOD_NOT idNOT +#else +# define METHOD_NOT '!' +#endif + +#define idFWD_REST '*' +#ifdef RUBY3_KEYWORDS +#define idFWD_KWREST idPow /* Use simple "**", as tDSTAR is "**arg" */ +#else +#define idFWD_KWREST 0 +#endif +#define idFWD_BLOCK '&' + +#define RE_OPTION_ONCE (1<<16) +#define RE_OPTION_ENCODING_SHIFT 8 +#define RE_OPTION_ENCODING(e) (((e)&0xff)<<RE_OPTION_ENCODING_SHIFT) +#define RE_OPTION_ENCODING_IDX(o) (((o)>>RE_OPTION_ENCODING_SHIFT)&0xff) +#define RE_OPTION_ENCODING_NONE(o) ((o)&RE_OPTION_ARG_ENCODING_NONE) +#define RE_OPTION_MASK 0xff +#define RE_OPTION_ARG_ENCODING_NONE 32 + +/* structs for managing terminator of string literal and heredocment */ +typedef struct rb_strterm_literal_struct { + union { + VALUE dummy; + long nest; + } u0; + union { + VALUE dummy; + long func; /* STR_FUNC_* (e.g., STR_FUNC_ESCAPE and STR_FUNC_EXPAND) */ + } u1; + union { + VALUE dummy; + long paren; /* '(' of `%q(...)` */ + } u2; + union { + VALUE dummy; + long term; /* ')' of `%q(...)` */ + } u3; +} rb_strterm_literal_t; + +#define HERETERM_LENGTH_BITS ((SIZEOF_VALUE - 1) * CHAR_BIT - 1) + +struct rb_strterm_heredoc_struct { + VALUE lastline; /* the string of line that contains `<<"END"` */ + long offset; /* the column of END in `<<"END"` */ + int sourceline; /* lineno of the line that contains `<<"END"` */ + unsigned length /* the length of END in `<<"END"` */ +#if HERETERM_LENGTH_BITS < SIZEOF_INT * CHAR_BIT + : HERETERM_LENGTH_BITS +# define HERETERM_LENGTH_MAX ((1U << HERETERM_LENGTH_BITS) - 1) +#else +# define HERETERM_LENGTH_MAX UINT_MAX +#endif + ; +#if HERETERM_LENGTH_BITS < SIZEOF_INT * CHAR_BIT + unsigned quote: 1; + unsigned func: 8; +#else + uint8_t quote; + uint8_t func; +#endif +}; +STATIC_ASSERT(rb_strterm_heredoc_t, sizeof(rb_strterm_heredoc_t) <= 4 * SIZEOF_VALUE); + +#define STRTERM_HEREDOC IMEMO_FL_USER0 + +struct rb_strterm_struct { + VALUE flags; + union { + rb_strterm_literal_t literal; + rb_strterm_heredoc_t heredoc; + } u; +}; + +#ifndef RIPPER +void +rb_strterm_mark(VALUE obj) +{ + rb_strterm_t *strterm = (rb_strterm_t*)obj; + if (RBASIC(obj)->flags & STRTERM_HEREDOC) { + rb_strterm_heredoc_t *heredoc = &strterm->u.heredoc; + rb_gc_mark(heredoc->lastline); + } +} +#endif + +#define yytnamerr(yyres, yystr) (YYSIZE_T)rb_yytnamerr(p, yyres, yystr) +size_t rb_yytnamerr(struct parser_params *p, char *yyres, const char *yystr); + +#define TOKEN2ID(tok) ( \ + tTOKEN_LOCAL_BEGIN<(tok)&&(tok)<tTOKEN_LOCAL_END ? TOKEN2LOCALID(tok) : \ + tTOKEN_INSTANCE_BEGIN<(tok)&&(tok)<tTOKEN_INSTANCE_END ? TOKEN2INSTANCEID(tok) : \ + tTOKEN_GLOBAL_BEGIN<(tok)&&(tok)<tTOKEN_GLOBAL_END ? TOKEN2GLOBALID(tok) : \ + tTOKEN_CONST_BEGIN<(tok)&&(tok)<tTOKEN_CONST_END ? TOKEN2CONSTID(tok) : \ + tTOKEN_CLASS_BEGIN<(tok)&&(tok)<tTOKEN_CLASS_END ? TOKEN2CLASSID(tok) : \ + tTOKEN_ATTRSET_BEGIN<(tok)&&(tok)<tTOKEN_ATTRSET_END ? TOKEN2ATTRSETID(tok) : \ + ((tok) / ((tok)<tPRESERVED_ID_END && ((tok)>=128 || rb_ispunct(tok))))) + +/****** Ripper *******/ + +#ifdef RIPPER +#define RIPPER_VERSION "0.1.0" + +static inline VALUE intern_sym(const char *name); + +#include "eventids1.c" +#include "eventids2.c" + +static VALUE ripper_dispatch0(struct parser_params*,ID); +static VALUE ripper_dispatch1(struct parser_params*,ID,VALUE); +static VALUE ripper_dispatch2(struct parser_params*,ID,VALUE,VALUE); +static VALUE ripper_dispatch3(struct parser_params*,ID,VALUE,VALUE,VALUE); +static VALUE ripper_dispatch4(struct parser_params*,ID,VALUE,VALUE,VALUE,VALUE); +static VALUE ripper_dispatch5(struct parser_params*,ID,VALUE,VALUE,VALUE,VALUE,VALUE); +static VALUE ripper_dispatch7(struct parser_params*,ID,VALUE,VALUE,VALUE,VALUE,VALUE,VALUE,VALUE); +static void ripper_error(struct parser_params *p); + +#define dispatch0(n) ripper_dispatch0(p, TOKEN_PASTE(ripper_id_, n)) +#define dispatch1(n,a) ripper_dispatch1(p, TOKEN_PASTE(ripper_id_, n), (a)) +#define dispatch2(n,a,b) ripper_dispatch2(p, TOKEN_PASTE(ripper_id_, n), (a), (b)) +#define dispatch3(n,a,b,c) ripper_dispatch3(p, TOKEN_PASTE(ripper_id_, n), (a), (b), (c)) +#define dispatch4(n,a,b,c,d) ripper_dispatch4(p, TOKEN_PASTE(ripper_id_, n), (a), (b), (c), (d)) +#define dispatch5(n,a,b,c,d,e) ripper_dispatch5(p, TOKEN_PASTE(ripper_id_, n), (a), (b), (c), (d), (e)) +#define dispatch7(n,a,b,c,d,e,f,g) ripper_dispatch7(p, TOKEN_PASTE(ripper_id_, n), (a), (b), (c), (d), (e), (f), (g)) +#define yyparse ripper_yyparse + +#define ID2VAL(id) STATIC_ID2SYM(id) +#define TOKEN2VAL(t) ID2VAL(TOKEN2ID(t)) +#define KWD2EID(t, v) ripper_new_yylval(p, keyword_##t, get_value(v), 0) + +#define params_new(pars, opts, rest, pars2, kws, kwrest, blk) \ + dispatch7(params, (pars), (opts), (rest), (pars2), (kws), (kwrest), (blk)) + +#define escape_Qundef(x) ((x)==Qundef ? Qnil : (x)) + +static inline VALUE +new_args(struct parser_params *p, VALUE pre_args, VALUE opt_args, VALUE rest_arg, VALUE post_args, VALUE tail, YYLTYPE *loc) +{ + NODE *t = (NODE *)tail; + VALUE kw_args = t->u1.value, kw_rest_arg = t->u2.value, block = t->u3.value; + return params_new(pre_args, opt_args, rest_arg, post_args, kw_args, kw_rest_arg, escape_Qundef(block)); +} + +static inline VALUE +new_args_tail(struct parser_params *p, VALUE kw_args, VALUE kw_rest_arg, VALUE block, YYLTYPE *loc) +{ + NODE *t = rb_node_newnode(NODE_ARGS_AUX, kw_args, kw_rest_arg, block, &NULL_LOC); + add_mark_object(p, kw_args); + add_mark_object(p, kw_rest_arg); + add_mark_object(p, block); + return (VALUE)t; +} + +static inline VALUE +args_with_numbered(struct parser_params *p, VALUE args, int max_numparam) +{ + return args; +} + +static VALUE +new_array_pattern(struct parser_params *p, VALUE constant, VALUE pre_arg, VALUE aryptn, const YYLTYPE *loc) +{ + NODE *t = (NODE *)aryptn; + VALUE pre_args = t->u1.value, rest_arg = t->u2.value, post_args = t->u3.value; + + if (!NIL_P(pre_arg)) { + if (!NIL_P(pre_args)) { + rb_ary_unshift(pre_args, pre_arg); + } + else { + pre_args = rb_ary_new_from_args(1, pre_arg); + } + } + return dispatch4(aryptn, constant, pre_args, rest_arg, post_args); +} + +static VALUE +new_array_pattern_tail(struct parser_params *p, VALUE pre_args, VALUE has_rest, VALUE rest_arg, VALUE post_args, const YYLTYPE *loc) +{ + NODE *t; + + if (has_rest) { + rest_arg = dispatch1(var_field, rest_arg ? rest_arg : Qnil); + } + else { + rest_arg = Qnil; + } + + t = rb_node_newnode(NODE_ARYPTN, pre_args, rest_arg, post_args, &NULL_LOC); + add_mark_object(p, pre_args); + add_mark_object(p, rest_arg); + add_mark_object(p, post_args); + return (VALUE)t; +} + +static VALUE +new_find_pattern(struct parser_params *p, VALUE constant, VALUE fndptn, const YYLTYPE *loc) +{ + NODE *t = (NODE *)fndptn; + VALUE pre_rest_arg = t->u1.value, args = t->u2.value, post_rest_arg = t->u3.value; + + return dispatch4(fndptn, constant, pre_rest_arg, args, post_rest_arg); +} + +static VALUE +new_find_pattern_tail(struct parser_params *p, VALUE pre_rest_arg, VALUE args, VALUE post_rest_arg, const YYLTYPE *loc) +{ + NODE *t; + + pre_rest_arg = dispatch1(var_field, pre_rest_arg ? pre_rest_arg : Qnil); + post_rest_arg = dispatch1(var_field, post_rest_arg ? post_rest_arg : Qnil); + + t = rb_node_newnode(NODE_FNDPTN, pre_rest_arg, args, post_rest_arg, &NULL_LOC); + add_mark_object(p, pre_rest_arg); + add_mark_object(p, args); + add_mark_object(p, post_rest_arg); + return (VALUE)t; +} + +#define new_hash(p,h,l) rb_ary_new_from_args(0) + +static VALUE +new_unique_key_hash(struct parser_params *p, VALUE ary, const YYLTYPE *loc) +{ + return ary; +} + +static VALUE +new_hash_pattern(struct parser_params *p, VALUE constant, VALUE hshptn, const YYLTYPE *loc) +{ + NODE *t = (NODE *)hshptn; + VALUE kw_args = t->u1.value, kw_rest_arg = t->u2.value; + return dispatch3(hshptn, constant, kw_args, kw_rest_arg); +} + +static VALUE +new_hash_pattern_tail(struct parser_params *p, VALUE kw_args, VALUE kw_rest_arg, const YYLTYPE *loc) +{ + NODE *t; + if (kw_rest_arg) { + kw_rest_arg = dispatch1(var_field, kw_rest_arg); + } + else { + kw_rest_arg = Qnil; + } + t = rb_node_newnode(NODE_HSHPTN, kw_args, kw_rest_arg, 0, &NULL_LOC); + + add_mark_object(p, kw_args); + add_mark_object(p, kw_rest_arg); + return (VALUE)t; +} + +#define new_defined(p,expr,loc) dispatch1(defined, (expr)) + +static VALUE heredoc_dedent(struct parser_params*,VALUE); + +#else +#define ID2VAL(id) (id) +#define TOKEN2VAL(t) ID2VAL(t) +#define KWD2EID(t, v) keyword_##t + +static NODE * +set_defun_body(struct parser_params *p, NODE *n, NODE *args, NODE *body, const YYLTYPE *loc) +{ + body = remove_begin(body); + reduce_nodes(p, &body); + n->nd_defn = NEW_SCOPE(args, body, loc); + n->nd_loc = *loc; + nd_set_line(n->nd_defn, loc->end_pos.lineno); + set_line_body(body, loc->beg_pos.lineno); + return n; +} + +static NODE * +rescued_expr(struct parser_params *p, NODE *arg, NODE *rescue, + const YYLTYPE *arg_loc, const YYLTYPE *mod_loc, const YYLTYPE *res_loc) +{ + YYLTYPE loc = code_loc_gen(mod_loc, res_loc); + rescue = NEW_RESBODY(0, remove_begin(rescue), 0, &loc); + loc.beg_pos = arg_loc->beg_pos; + return NEW_RESCUE(arg, rescue, 0, &loc); +} + +#endif /* RIPPER */ + +static void +restore_defun(struct parser_params *p, NODE *name) +{ + YYSTYPE c = {.val = name->nd_cval}; + p->cur_arg = name->nd_vid; + p->ctxt.in_def = c.ctxt.in_def; + p->ctxt.shareable_constant_value = c.ctxt.shareable_constant_value; +} + +static void +endless_method_name(struct parser_params *p, NODE *defn, const YYLTYPE *loc) +{ +#ifdef RIPPER + defn = defn->nd_defn; +#endif + ID mid = defn->nd_mid; + if (is_attrset_id(mid)) { + yyerror1(loc, "setter method cannot be defined in an endless method definition"); + } + token_info_drop(p, "def", loc->beg_pos); +} + +#ifndef RIPPER +# define Qnone 0 +# define Qnull 0 +# define ifndef_ripper(x) (x) +#else +# define Qnone Qnil +# define Qnull Qundef +# define ifndef_ripper(x) +#endif + +# define rb_warn0(fmt) WARN_CALL(WARN_ARGS(fmt, 1)) +# define rb_warn1(fmt,a) WARN_CALL(WARN_ARGS(fmt, 2), (a)) +# define rb_warn2(fmt,a,b) WARN_CALL(WARN_ARGS(fmt, 3), (a), (b)) +# define rb_warn3(fmt,a,b,c) WARN_CALL(WARN_ARGS(fmt, 4), (a), (b), (c)) +# define rb_warn4(fmt,a,b,c,d) WARN_CALL(WARN_ARGS(fmt, 5), (a), (b), (c), (d)) +# define rb_warning0(fmt) WARNING_CALL(WARNING_ARGS(fmt, 1)) +# define rb_warning1(fmt,a) WARNING_CALL(WARNING_ARGS(fmt, 2), (a)) +# define rb_warning2(fmt,a,b) WARNING_CALL(WARNING_ARGS(fmt, 3), (a), (b)) +# define rb_warning3(fmt,a,b,c) WARNING_CALL(WARNING_ARGS(fmt, 4), (a), (b), (c)) +# define rb_warning4(fmt,a,b,c,d) WARNING_CALL(WARNING_ARGS(fmt, 5), (a), (b), (c), (d)) +# define rb_warn0L(l,fmt) WARN_CALL(WARN_ARGS_L(l, fmt, 1)) +# define rb_warn1L(l,fmt,a) WARN_CALL(WARN_ARGS_L(l, fmt, 2), (a)) +# define rb_warn2L(l,fmt,a,b) WARN_CALL(WARN_ARGS_L(l, fmt, 3), (a), (b)) +# define rb_warn3L(l,fmt,a,b,c) WARN_CALL(WARN_ARGS_L(l, fmt, 4), (a), (b), (c)) +# define rb_warn4L(l,fmt,a,b,c,d) WARN_CALL(WARN_ARGS_L(l, fmt, 5), (a), (b), (c), (d)) +# define rb_warning0L(l,fmt) WARNING_CALL(WARNING_ARGS_L(l, fmt, 1)) +# define rb_warning1L(l,fmt,a) WARNING_CALL(WARNING_ARGS_L(l, fmt, 2), (a)) +# define rb_warning2L(l,fmt,a,b) WARNING_CALL(WARNING_ARGS_L(l, fmt, 3), (a), (b)) +# define rb_warning3L(l,fmt,a,b,c) WARNING_CALL(WARNING_ARGS_L(l, fmt, 4), (a), (b), (c)) +# define rb_warning4L(l,fmt,a,b,c,d) WARNING_CALL(WARNING_ARGS_L(l, fmt, 5), (a), (b), (c), (d)) +#ifdef RIPPER +static ID id_warn, id_warning, id_gets, id_assoc; +# define ERR_MESG() STR_NEW2(mesg) /* to bypass Ripper DSL */ +# define WARN_S_L(s,l) STR_NEW(s,l) +# define WARN_S(s) STR_NEW2(s) +# define WARN_I(i) INT2NUM(i) +# define WARN_ID(i) rb_id2str(i) +# define WARN_IVAL(i) i +# define PRIsWARN "s" +# define rb_warn0L_experimental(l,fmt) WARN_CALL(WARN_ARGS_L(l, fmt, 1)) +# define WARN_ARGS(fmt,n) p->value, id_warn, n, rb_usascii_str_new_lit(fmt) +# define WARN_ARGS_L(l,fmt,n) WARN_ARGS(fmt,n) +# ifdef HAVE_VA_ARGS_MACRO +# define WARN_CALL(...) rb_funcall(__VA_ARGS__) +# else +# define WARN_CALL rb_funcall +# endif +# define WARNING_ARGS(fmt,n) p->value, id_warning, n, rb_usascii_str_new_lit(fmt) +# define WARNING_ARGS_L(l, fmt,n) WARNING_ARGS(fmt,n) +# ifdef HAVE_VA_ARGS_MACRO +# define WARNING_CALL(...) rb_funcall(__VA_ARGS__) +# else +# define WARNING_CALL rb_funcall +# endif +PRINTF_ARGS(static void ripper_compile_error(struct parser_params*, const char *fmt, ...), 2, 3); +# define compile_error ripper_compile_error +#else +# define WARN_S_L(s,l) s +# define WARN_S(s) s +# define WARN_I(i) i +# define WARN_ID(i) rb_id2name(i) +# define WARN_IVAL(i) NUM2INT(i) +# define PRIsWARN PRIsVALUE +# define WARN_ARGS(fmt,n) WARN_ARGS_L(p->ruby_sourceline,fmt,n) +# define WARN_ARGS_L(l,fmt,n) p->ruby_sourcefile, (l), (fmt) +# define WARN_CALL rb_compile_warn +# define rb_warn0L_experimental(l,fmt) rb_category_compile_warn(RB_WARN_CATEGORY_EXPERIMENTAL, WARN_ARGS_L(l, fmt, 1)) +# define WARNING_ARGS(fmt,n) WARN_ARGS(fmt,n) +# define WARNING_ARGS_L(l,fmt,n) WARN_ARGS_L(l,fmt,n) +# define WARNING_CALL rb_compile_warning +PRINTF_ARGS(static void parser_compile_error(struct parser_params*, const char *fmt, ...), 2, 3); +# define compile_error parser_compile_error +#endif + +#define WARN_EOL(tok) \ + (looking_at_eol_p(p) ? \ + (void)rb_warning0("`" tok "' at the end of line without an expression") : \ + (void)0) +static int looking_at_eol_p(struct parser_params *p); %} +%expect 0 +%define api.pure +%define parse.error verbose +%printer { +#ifndef RIPPER + rb_parser_printf(p, "%"PRIsVALUE, rb_id2str($$)); +#else + rb_parser_printf(p, "%"PRIsVALUE, RNODE($$)->nd_rval); +#endif +} tIDENTIFIER tFID tGVAR tIVAR tCONSTANT tCVAR tLABEL tOP_ASGN +%printer { +#ifndef RIPPER + rb_parser_printf(p, "%+"PRIsVALUE, $$->nd_lit); +#else + rb_parser_printf(p, "%+"PRIsVALUE, get_value($$)); +#endif +} tINTEGER tFLOAT tRATIONAL tIMAGINARY tSTRING_CONTENT tCHAR +%printer { +#ifndef RIPPER + rb_parser_printf(p, "$%ld", $$->nd_nth); +#else + rb_parser_printf(p, "%"PRIsVALUE, $$); +#endif +} tNTH_REF +%printer { +#ifndef RIPPER + rb_parser_printf(p, "$%c", (int)$$->nd_nth); +#else + rb_parser_printf(p, "%"PRIsVALUE, $$); +#endif +} tBACK_REF + +%lex-param {struct parser_params *p} +%parse-param {struct parser_params *p} +%initial-action +{ + RUBY_SET_YYLLOC_OF_NONE(@$); +}; + %union { + VALUE val; NODE *node; ID id; int num; - struct RVarmap *vars; -} - -%token kCLASS - kMODULE - kDEF - kUNDEF - kBEGIN - kRESCUE - kENSURE - kEND - kIF - kUNLESS - kTHEN - kELSIF - kELSE - kCASE - kWHEN - kWHILE - kUNTIL - kFOR - kBREAK - kNEXT - kREDO - kRETRY - kIN - kDO - kDO_COND - kDO_BLOCK - kRETURN - kYIELD - kSUPER - kSELF - kNIL - kTRUE - kFALSE - kAND - kOR - kNOT - kIF_MOD - kUNLESS_MOD - kWHILE_MOD - kUNTIL_MOD - kRESCUE_MOD - kALIAS - kDEFINED - klBEGIN - klEND - k__LINE__ - k__FILE__ - -%token <id> tIDENTIFIER tFID tGVAR tIVAR tCONSTANT tCVAR -%token <node> tINTEGER tFLOAT tSTRING_CONTENT -%token <node> tNTH_REF tBACK_REF + st_table *tbl; + const struct vtable *vars; + struct rb_strterm_struct *strterm; + struct lex_context ctxt; +} + +%token <id> + keyword_class "`class'" + keyword_module "`module'" + keyword_def "`def'" + keyword_undef "`undef'" + keyword_begin "`begin'" + keyword_rescue "`rescue'" + keyword_ensure "`ensure'" + keyword_end "`end'" + keyword_if "`if'" + keyword_unless "`unless'" + keyword_then "`then'" + keyword_elsif "`elsif'" + keyword_else "`else'" + keyword_case "`case'" + keyword_when "`when'" + keyword_while "`while'" + keyword_until "`until'" + keyword_for "`for'" + keyword_break "`break'" + keyword_next "`next'" + keyword_redo "`redo'" + keyword_retry "`retry'" + keyword_in "`in'" + keyword_do "`do'" + keyword_do_cond "`do' for condition" + keyword_do_block "`do' for block" + keyword_do_LAMBDA "`do' for lambda" + keyword_return "`return'" + keyword_yield "`yield'" + keyword_super "`super'" + keyword_self "`self'" + keyword_nil "`nil'" + keyword_true "`true'" + keyword_false "`false'" + keyword_and "`and'" + keyword_or "`or'" + keyword_not "`not'" + modifier_if "`if' modifier" + modifier_unless "`unless' modifier" + modifier_while "`while' modifier" + modifier_until "`until' modifier" + modifier_rescue "`rescue' modifier" + keyword_alias "`alias'" + keyword_defined "`defined?'" + keyword_BEGIN "`BEGIN'" + keyword_END "`END'" + keyword__LINE__ "`__LINE__'" + keyword__FILE__ "`__FILE__'" + keyword__ENCODING__ "`__ENCODING__'" + +%token <id> tIDENTIFIER "local variable or method" +%token <id> tFID "method" +%token <id> tGVAR "global variable" +%token <id> tIVAR "instance variable" +%token <id> tCONSTANT "constant" +%token <id> tCVAR "class variable" +%token <id> tLABEL "label" +%token <node> tINTEGER "integer literal" +%token <node> tFLOAT "float literal" +%token <node> tRATIONAL "rational literal" +%token <node> tIMAGINARY "imaginary literal" +%token <node> tCHAR "char literal" +%token <node> tNTH_REF "numbered reference" +%token <node> tBACK_REF "back reference" +%token <node> tSTRING_CONTENT "literal content" %token <num> tREGEXP_END %type <node> singleton strings string string1 xstring regexp -%type <node> string_contents xstring_contents string_content -%type <node> words qwords word_list qword_list word -%type <node> literal numeric dsym cpath -%type <node> bodystmt compstmt stmts stmt expr arg primary command command_call method_call -%type <node> expr_value arg_value primary_value -%type <node> if_tail opt_else case_body cases opt_rescue exc_list exc_var opt_ensure -%type <node> args when_args call_args call_args2 open_args paren_args opt_paren_args +%type <node> string_contents xstring_contents regexp_contents string_content +%type <node> words symbols symbol_list qwords qsymbols word_list qword_list qsym_list word +%type <node> literal numeric simple_numeric ssym dsym symbol cpath def_name defn_head defs_head +%type <node> top_compstmt top_stmts top_stmt begin_block +%type <node> bodystmt compstmt stmts stmt_or_begin stmt expr arg primary command command_call method_call +%type <node> expr_value expr_value_do arg_value primary_value fcall rel_expr +%type <node> if_tail opt_else case_body case_args cases opt_rescue exc_list exc_var opt_ensure +%type <node> args call_args opt_call_args +%type <node> paren_args opt_paren_args args_tail opt_args_tail block_args_tail opt_block_args_tail %type <node> command_args aref_args opt_block_arg block_arg var_ref var_lhs -%type <node> mrhs superclass block_call block_command -%type <node> f_arglist f_args f_optarg f_opt f_rest_arg f_block_arg opt_f_block_arg -%type <node> assoc_list assocs assoc undef_list backref string_dvar -%type <node> block_var opt_block_var brace_block cmd_brace_block do_block lhs none fitem -%type <node> mlhs mlhs_head mlhs_basic mlhs_entry mlhs_item mlhs_node -%type <id> fsym variable sym symbol operation operation2 operation3 -%type <id> cname fname op -%type <num> f_norm_arg f_arg -%token tUPLUS /* unary+ */ -%token tUMINUS /* unary- */ -%token tPOW /* ** */ -%token tCMP /* <=> */ -%token tEQ /* == */ -%token tEQQ /* === */ -%token tNEQ /* != */ -%token tGEQ /* >= */ -%token tLEQ /* <= */ -%token tANDOP tOROP /* && and || */ -%token tMATCH tNMATCH /* =~ and !~ */ -%token tDOT2 tDOT3 /* .. and ... */ -%token tAREF tASET /* [] and []= */ -%token tLSHFT tRSHFT /* << and >> */ -%token tCOLON2 /* :: */ -%token tCOLON3 /* :: at EXPR_BEG */ -%token <id> tOP_ASGN /* +=, -= etc. */ -%token tASSOC /* => */ -%token tLPAREN /* ( */ -%token tLPAREN_ARG /* ( */ -%token tRPAREN /* ) */ -%token tLBRACK /* [ */ -%token tLBRACE /* { */ -%token tLBRACE_ARG /* { */ -%token tSTAR /* * */ -%token tAMPER /* & */ -%token tSYMBEG tSTRING_BEG tXSTRING_BEG tREGEXP_BEG tWORDS_BEG tQWORDS_BEG -%token tSTRING_DBEG tSTRING_DVAR tSTRING_END +%type <node> command_rhs arg_rhs +%type <node> command_asgn mrhs mrhs_arg superclass block_call block_command +%type <node> f_block_optarg f_block_opt +%type <node> f_arglist f_opt_paren_args f_paren_args f_args f_arg f_arg_item +%type <node> f_optarg f_marg f_marg_list f_margs f_rest_marg +%type <node> assoc_list assocs assoc undef_list backref string_dvar for_var +%type <node> block_param opt_block_param block_param_def f_opt +%type <node> f_kwarg f_kw f_block_kwarg f_block_kw +%type <node> bv_decls opt_bv_decl bvar +%type <node> lambda f_larglist lambda_body brace_body do_body +%type <node> brace_block cmd_brace_block do_block lhs none fitem +%type <node> mlhs mlhs_head mlhs_basic mlhs_item mlhs_node mlhs_post mlhs_inner +%type <node> p_case_body p_cases p_top_expr p_top_expr_body +%type <node> p_expr p_as p_alt p_expr_basic p_find +%type <node> p_args p_args_head p_args_tail p_args_post p_arg +%type <node> p_value p_primitive p_variable p_var_ref p_expr_ref p_const +%type <node> p_kwargs p_kwarg p_kw +%type <id> keyword_variable user_variable sym operation operation2 operation3 +%type <id> cname fname op f_rest_arg f_block_arg opt_f_block_arg f_norm_arg f_bad_arg +%type <id> f_kwrest f_label f_arg_asgn call_op call_op2 reswords relop dot_or_colon +%type <id> p_rest p_kwrest p_kwnorest p_any_kwrest p_kw_label +%type <id> f_no_kwarg f_any_kwrest args_forward excessed_comma nonlocal_var + %type <ctxt> lex_ctxt /* keep <ctxt> in ripper */ +%token END_OF_INPUT 0 "end-of-input" +%token <id> '.' +/* escaped chars, should be ignored otherwise */ +%token <id> '\\' "backslash" +%token tSP "escaped space" +%token <id> '\t' "escaped horizontal tab" +%token <id> '\f' "escaped form feed" +%token <id> '\r' "escaped carriage return" +%token <id> '\13' "escaped vertical tab" +%token tUPLUS RUBY_TOKEN(UPLUS) "unary+" +%token tUMINUS RUBY_TOKEN(UMINUS) "unary-" +%token tPOW RUBY_TOKEN(POW) "**" +%token tCMP RUBY_TOKEN(CMP) "<=>" +%token tEQ RUBY_TOKEN(EQ) "==" +%token tEQQ RUBY_TOKEN(EQQ) "===" +%token tNEQ RUBY_TOKEN(NEQ) "!=" +%token tGEQ RUBY_TOKEN(GEQ) ">=" +%token tLEQ RUBY_TOKEN(LEQ) "<=" +%token tANDOP RUBY_TOKEN(ANDOP) "&&" +%token tOROP RUBY_TOKEN(OROP) "||" +%token tMATCH RUBY_TOKEN(MATCH) "=~" +%token tNMATCH RUBY_TOKEN(NMATCH) "!~" +%token tDOT2 RUBY_TOKEN(DOT2) ".." +%token tDOT3 RUBY_TOKEN(DOT3) "..." +%token tBDOT2 RUBY_TOKEN(BDOT2) "(.." +%token tBDOT3 RUBY_TOKEN(BDOT3) "(..." +%token tAREF RUBY_TOKEN(AREF) "[]" +%token tASET RUBY_TOKEN(ASET) "[]=" +%token tLSHFT RUBY_TOKEN(LSHFT) "<<" +%token tRSHFT RUBY_TOKEN(RSHFT) ">>" +%token <id> tANDDOT RUBY_TOKEN(ANDDOT) "&." +%token <id> tCOLON2 RUBY_TOKEN(COLON2) "::" +%token tCOLON3 ":: at EXPR_BEG" +%token <id> tOP_ASGN "operator-assignment" /* +=, -= etc. */ +%token tASSOC "=>" +%token tLPAREN "(" +%token tLPAREN_ARG "( arg" +%token tRPAREN ")" +%token tLBRACK "[" +%token tLBRACE "{" +%token tLBRACE_ARG "{ arg" +%token tSTAR "*" +%token tDSTAR "**arg" +%token tAMPER "&" +%token tLAMBDA "->" +%token tSYMBEG "symbol literal" +%token tSTRING_BEG "string literal" +%token tXSTRING_BEG "backtick literal" +%token tREGEXP_BEG "regexp literal" +%token tWORDS_BEG "word list" +%token tQWORDS_BEG "verbatim word list" +%token tSYMBOLS_BEG "symbol list" +%token tQSYMBOLS_BEG "verbatim symbol list" +%token tSTRING_END "terminator" +%token tSTRING_DEND "'}'" +%token tSTRING_DBEG tSTRING_DVAR tLAMBEG tLABEL_END /* * precedence table @@ -319,14 +1300,14 @@ static void top_local_setup(); %nonassoc tLOWEST %nonassoc tLBRACE_ARG -%nonassoc kIF_MOD kUNLESS_MOD kWHILE_MOD kUNTIL_MOD -%left kOR kAND -%right kNOT -%nonassoc kDEFINED +%nonassoc modifier_if modifier_unless modifier_while modifier_until keyword_in +%left keyword_or keyword_and +%right keyword_not +%nonassoc keyword_defined %right '=' tOP_ASGN -%left kRESCUE_MOD +%left modifier_rescue %right '?' ':' -%nonassoc tDOT2 tDOT3 +%nonassoc tDOT2 tDOT3 tBDOT2 tBDOT3 %left tOROP %left tANDOP %nonassoc tCMP tEQ tEQQ tNEQ tMATCH tNMATCH @@ -344,558 +1325,963 @@ static void top_local_setup(); %% program : { - lex_state = EXPR_BEG; - top_local_init(); - if (ruby_class == rb_cObject) class_nest = 0; - else class_nest = 1; + SET_LEX_STATE(EXPR_BEG); + local_push(p, ifndef_ripper(1)+0); } - compstmt + top_compstmt { + /*%%%*/ if ($2 && !compile_for_eval) { - /* last expression should not be void */ - if (nd_type($2) != NODE_BLOCK) void_expr($2); - else { - NODE *node = $2; + NODE *node = $2; + /* last expression should not be void */ + if (nd_type_p(node, NODE_BLOCK)) { while (node->nd_next) { node = node->nd_next; } - void_expr(node->nd_head); + node = node->nd_head; } + node = remove_begin(node); + void_expr(p, node); } - ruby_eval_tree = block_append(ruby_eval_tree, $2); - top_local_setup(); - class_nest = 0; + p->eval_tree = NEW_SCOPE(0, block_append(p, p->eval_tree, $2), &@$); + /*% %*/ + /*% ripper[final]: program!($2) %*/ + local_pop(p); + } + ; + +top_compstmt : top_stmts opt_terms + { + $$ = void_stmts(p, $1); + } + ; + +top_stmts : none + { + /*%%%*/ + $$ = NEW_BEGIN(0, &@$); + /*% %*/ + /*% ripper: stmts_add!(stmts_new!, void_stmt!) %*/ + } + | top_stmt + { + /*%%%*/ + $$ = newline_node($1); + /*% %*/ + /*% ripper: stmts_add!(stmts_new!, $1) %*/ + } + | top_stmts terms top_stmt + { + /*%%%*/ + $$ = block_append(p, $1, newline_node($3)); + /*% %*/ + /*% ripper: stmts_add!($1, $3) %*/ + } + | error top_stmt + { + $$ = remove_begin($2); + } + ; + +top_stmt : stmt + | keyword_BEGIN begin_block + { + $$ = $2; + } + ; + +begin_block : '{' top_compstmt '}' + { + /*%%%*/ + p->eval_tree_begin = block_append(p, p->eval_tree_begin, + NEW_BEGIN($2, &@$)); + $$ = NEW_BEGIN(0, &@$); + /*% %*/ + /*% ripper: BEGIN!($2) %*/ } ; bodystmt : compstmt opt_rescue - opt_else + k_else {if (!$2) {yyerror1(&@3, "else without rescue is useless");}} + compstmt opt_ensure { - $$ = $1; - if ($2) { - $$ = NEW_RESCUE($1, $2, $3); - } - else if ($3) { - rb_warn("else without rescue is useless"); - $$ = block_append($$, $3); - } - if ($4) { - $$ = NEW_ENSURE($$, $4); - } - fixpos($$, $1); + /*%%%*/ + $$ = new_bodystmt(p, $1, $2, $5, $6, &@$); + /*% %*/ + /*% ripper: bodystmt!(escape_Qundef($1), escape_Qundef($2), escape_Qundef($5), escape_Qundef($6)) %*/ + } + | compstmt + opt_rescue + opt_ensure + { + /*%%%*/ + $$ = new_bodystmt(p, $1, $2, 0, $3, &@$); + /*% %*/ + /*% ripper: bodystmt!(escape_Qundef($1), escape_Qundef($2), Qnil, escape_Qundef($3)) %*/ } ; compstmt : stmts opt_terms { - void_stmts($1); - $$ = $1; + $$ = void_stmts(p, $1); } ; stmts : none - | stmt + { + /*%%%*/ + $$ = NEW_BEGIN(0, &@$); + /*% %*/ + /*% ripper: stmts_add!(stmts_new!, void_stmt!) %*/ + } + | stmt_or_begin { + /*%%%*/ $$ = newline_node($1); + /*% %*/ + /*% ripper: stmts_add!(stmts_new!, $1) %*/ } - | stmts terms stmt + | stmts terms stmt_or_begin { - $$ = block_append($1, newline_node($3)); + /*%%%*/ + $$ = block_append(p, $1, newline_node($3)); + /*% %*/ + /*% ripper: stmts_add!($1, $3) %*/ } | error stmt { - $$ = $2; + $$ = remove_begin($2); } ; -stmt : kALIAS fitem {lex_state = EXPR_FNAME;} fitem - { - $$ = NEW_ALIAS($2, $4); +stmt_or_begin : stmt + { + $$ = $1; } - | kALIAS tGVAR tGVAR + | keyword_BEGIN { - $$ = NEW_VALIAS($2, $3); + yyerror1(&@1, "BEGIN is permitted only at toplevel"); } - | kALIAS tGVAR tBACK_REF + begin_block { - char buf[3]; + $$ = $3; + } + ; - sprintf(buf, "$%c", (char)$3->nd_nth); - $$ = NEW_VALIAS($2, rb_intern(buf)); +stmt : keyword_alias fitem {SET_LEX_STATE(EXPR_FNAME|EXPR_FITEM);} fitem + { + /*%%%*/ + $$ = NEW_ALIAS($2, $4, &@$); + /*% %*/ + /*% ripper: alias!($2, $4) %*/ } - | kALIAS tGVAR tNTH_REF + | keyword_alias tGVAR tGVAR { - yyerror("can't make alias for the number variables"); - $$ = 0; + /*%%%*/ + $$ = NEW_VALIAS($2, $3, &@$); + /*% %*/ + /*% ripper: var_alias!($2, $3) %*/ + } + | keyword_alias tGVAR tBACK_REF + { + /*%%%*/ + char buf[2]; + buf[0] = '$'; + buf[1] = (char)$3->nd_nth; + $$ = NEW_VALIAS($2, rb_intern2(buf, 2), &@$); + /*% %*/ + /*% ripper: var_alias!($2, $3) %*/ + } + | keyword_alias tGVAR tNTH_REF + { + static const char mesg[] = "can't make alias for the number variables"; + /*%%%*/ + yyerror1(&@3, mesg); + $$ = NEW_BEGIN(0, &@$); + /*% %*/ + /*% ripper[error]: alias_error!(ERR_MESG(), $3) %*/ } - | kUNDEF undef_list + | keyword_undef undef_list { + /*%%%*/ $$ = $2; + /*% %*/ + /*% ripper: undef!($2) %*/ } - | stmt kIF_MOD expr_value + | stmt modifier_if expr_value { - $$ = NEW_IF(cond($3), $1, 0); - fixpos($$, $3); - if (cond_negative(&$$->nd_cond)) { - $$->nd_else = $$->nd_body; - $$->nd_body = 0; - } + /*%%%*/ + $$ = new_if(p, $3, remove_begin($1), 0, &@$); + fixpos($$, $3); + /*% %*/ + /*% ripper: if_mod!($3, $1) %*/ } - | stmt kUNLESS_MOD expr_value + | stmt modifier_unless expr_value { - $$ = NEW_UNLESS(cond($3), $1, 0); - fixpos($$, $3); - if (cond_negative(&$$->nd_cond)) { - $$->nd_body = $$->nd_else; - $$->nd_else = 0; - } + /*%%%*/ + $$ = new_unless(p, $3, remove_begin($1), 0, &@$); + fixpos($$, $3); + /*% %*/ + /*% ripper: unless_mod!($3, $1) %*/ } - | stmt kWHILE_MOD expr_value + | stmt modifier_while expr_value { - if ($1 && nd_type($1) == NODE_BEGIN) { - $$ = NEW_WHILE(cond($3), $1->nd_body, 0); + /*%%%*/ + if ($1 && nd_type_p($1, NODE_BEGIN)) { + $$ = NEW_WHILE(cond(p, $3, &@3), $1->nd_body, 0, &@$); } else { - $$ = NEW_WHILE(cond($3), $1, 1); - } - if (cond_negative(&$$->nd_cond)) { - nd_set_type($$, NODE_UNTIL); + $$ = NEW_WHILE(cond(p, $3, &@3), $1, 1, &@$); } + /*% %*/ + /*% ripper: while_mod!($3, $1) %*/ } - | stmt kUNTIL_MOD expr_value + | stmt modifier_until expr_value { - if ($1 && nd_type($1) == NODE_BEGIN) { - $$ = NEW_UNTIL(cond($3), $1->nd_body, 0); + /*%%%*/ + if ($1 && nd_type_p($1, NODE_BEGIN)) { + $$ = NEW_UNTIL(cond(p, $3, &@3), $1->nd_body, 0, &@$); } else { - $$ = NEW_UNTIL(cond($3), $1, 1); - } - if (cond_negative(&$$->nd_cond)) { - nd_set_type($$, NODE_WHILE); + $$ = NEW_UNTIL(cond(p, $3, &@3), $1, 1, &@$); } + /*% %*/ + /*% ripper: until_mod!($3, $1) %*/ } - | stmt kRESCUE_MOD stmt + | stmt modifier_rescue stmt { - $$ = NEW_RESCUE($1, NEW_RESBODY(0,$3,0), 0); + /*%%%*/ + NODE *resq; + YYLTYPE loc = code_loc_gen(&@2, &@3); + resq = NEW_RESBODY(0, remove_begin($3), 0, &loc); + $$ = NEW_RESCUE(remove_begin($1), resq, 0, &@$); + /*% %*/ + /*% ripper: rescue_mod!($1, $3) %*/ } - | klBEGIN + | keyword_END '{' compstmt '}' { - if (in_def || in_single) { - yyerror("BEGIN in method"); + if (p->ctxt.in_def) { + rb_warn0("END in method; use at_exit"); } - local_push(0); + /*%%%*/ + { + NODE *scope = NEW_NODE( + NODE_SCOPE, 0 /* tbl */, $3 /* body */, 0 /* args */, &@$); + $$ = NEW_POSTEXE(scope, &@$); + } + /*% %*/ + /*% ripper: END!($3) %*/ } - '{' compstmt '}' + | command_asgn + | mlhs '=' lex_ctxt command_call { - ruby_eval_tree_begin = block_append(ruby_eval_tree_begin, - NEW_PREEXE($4)); - local_pop(); - $$ = 0; - } - | klEND '{' compstmt '}' + /*%%%*/ + value_expr($4); + $$ = node_assign(p, $1, $4, $3, &@$); + /*% %*/ + /*% ripper: massign!($1, $4) %*/ + } + | lhs '=' lex_ctxt mrhs + { + /*%%%*/ + $$ = node_assign(p, $1, $4, $3, &@$); + /*% %*/ + /*% ripper: assign!($1, $4) %*/ + } + | mlhs '=' lex_ctxt mrhs_arg modifier_rescue stmt + { + /*%%%*/ + YYLTYPE loc = code_loc_gen(&@5, &@6); + $$ = node_assign(p, $1, NEW_RESCUE($4, NEW_RESBODY(0, remove_begin($6), 0, &loc), 0, &@$), $3, &@$); + /*% %*/ + /*% ripper: massign!($1, rescue_mod!($4, $6)) %*/ + } + | mlhs '=' lex_ctxt mrhs_arg { - if (in_def || in_single) { - rb_warn("END in method; use at_exit"); - } + /*%%%*/ + $$ = node_assign(p, $1, $4, $3, &@$); + /*% %*/ + /*% ripper: massign!($1, $4) %*/ + } + | expr + ; - $$ = NEW_ITER(0, NEW_POSTEXE(), $3); +command_asgn : lhs '=' lex_ctxt command_rhs + { + /*%%%*/ + $$ = node_assign(p, $1, $4, $3, &@$); + /*% %*/ + /*% ripper: assign!($1, $4) %*/ + } + | var_lhs tOP_ASGN lex_ctxt command_rhs + { + /*%%%*/ + $$ = new_op_assign(p, $1, $2, $4, $3, &@$); + /*% %*/ + /*% ripper: opassign!($1, $2, $4) %*/ + } + | primary_value '[' opt_call_args rbracket tOP_ASGN lex_ctxt command_rhs + { + /*%%%*/ + $$ = new_ary_op_assign(p, $1, $3, $5, $7, &@3, &@$); + /*% %*/ + /*% ripper: opassign!(aref_field!($1, escape_Qundef($3)), $5, $7) %*/ + + } + | primary_value call_op tIDENTIFIER tOP_ASGN lex_ctxt command_rhs + { + /*%%%*/ + $$ = new_attr_op_assign(p, $1, $2, $3, $4, $6, &@$); + /*% %*/ + /*% ripper: opassign!(field!($1, $2, $3), $4, $6) %*/ + } + | primary_value call_op tCONSTANT tOP_ASGN lex_ctxt command_rhs + { + /*%%%*/ + $$ = new_attr_op_assign(p, $1, $2, $3, $4, $6, &@$); + /*% %*/ + /*% ripper: opassign!(field!($1, $2, $3), $4, $6) %*/ + } + | primary_value tCOLON2 tCONSTANT tOP_ASGN lex_ctxt command_rhs + { + /*%%%*/ + YYLTYPE loc = code_loc_gen(&@1, &@3); + $$ = new_const_op_assign(p, NEW_COLON2($1, $3, &loc), $4, $6, $5, &@$); + /*% %*/ + /*% ripper: opassign!(const_path_field!($1, $3), $4, $6) %*/ + } + | primary_value tCOLON2 tIDENTIFIER tOP_ASGN lex_ctxt command_rhs + { + /*%%%*/ + $$ = new_attr_op_assign(p, $1, ID2VAL(idCOLON2), $3, $4, $6, &@$); + /*% %*/ + /*% ripper: opassign!(field!($1, $2, $3), $4, $6) %*/ + } + | defn_head f_opt_paren_args '=' command + { + endless_method_name(p, $<node>1, &@1); + restore_defun(p, $<node>1->nd_defn); + /*%%%*/ + $$ = set_defun_body(p, $1, $2, $4, &@$); + /*% %*/ + /*% ripper[$4]: bodystmt!($4, Qnil, Qnil, Qnil) %*/ + /*% ripper: def!(get_value($1), $2, $4) %*/ + local_pop(p); + } + | defn_head f_opt_paren_args '=' command modifier_rescue arg + { + endless_method_name(p, $<node>1, &@1); + restore_defun(p, $<node>1->nd_defn); + /*%%%*/ + $4 = rescued_expr(p, $4, $6, &@4, &@5, &@6); + $$ = set_defun_body(p, $1, $2, $4, &@$); + /*% %*/ + /*% ripper[$4]: bodystmt!(rescue_mod!($4, $6), Qnil, Qnil, Qnil) %*/ + /*% ripper: def!(get_value($1), $2, $4) %*/ + local_pop(p); + } + | defs_head f_opt_paren_args '=' command + { + endless_method_name(p, $<node>1, &@1); + restore_defun(p, $<node>1->nd_defn); + /*%%%*/ + $$ = set_defun_body(p, $1, $2, $4, &@$); + /*% + $1 = get_value($1); + %*/ + /*% ripper[$4]: bodystmt!($4, Qnil, Qnil, Qnil) %*/ + /*% ripper: defs!(AREF($1, 0), AREF($1, 1), AREF($1, 2), $2, $4) %*/ + local_pop(p); + } + | defs_head f_opt_paren_args '=' command modifier_rescue arg + { + endless_method_name(p, $<node>1, &@1); + restore_defun(p, $<node>1->nd_defn); + /*%%%*/ + $4 = rescued_expr(p, $4, $6, &@4, &@5, &@6); + $$ = set_defun_body(p, $1, $2, $4, &@$); + /*% + $1 = get_value($1); + %*/ + /*% ripper[$4]: bodystmt!(rescue_mod!($4, $6), Qnil, Qnil, Qnil) %*/ + /*% ripper: defs!(AREF($1, 0), AREF($1, 1), AREF($1, 2), $2, $4) %*/ + local_pop(p); + } + | backref tOP_ASGN lex_ctxt command_rhs + { + /*%%%*/ + rb_backref_error(p, $1); + $$ = NEW_BEGIN(0, &@$); + /*% %*/ + /*% ripper[error]: backref_error(p, RNODE($1), assign!(var_field(p, $1), $4)) %*/ } - | lhs '=' command_call + ; + +command_rhs : command_call %prec tOP_ASGN { - $$ = node_assign($1, $3); + value_expr($1); + $$ = $1; } - | mlhs '=' command_call + | command_call modifier_rescue stmt { - value_expr($3); - $1->nd_value = ($1->nd_head) ? NEW_TO_ARY($3) : NEW_ARRAY($3); - $$ = $1; + /*%%%*/ + YYLTYPE loc = code_loc_gen(&@2, &@3); + value_expr($1); + $$ = NEW_RESCUE($1, NEW_RESBODY(0, remove_begin($3), 0, &loc), 0, &@$); + /*% %*/ + /*% ripper: rescue_mod!($1, $3) %*/ } - | var_lhs tOP_ASGN command_call + | command_asgn + ; + +expr : command_call + | expr keyword_and expr { - value_expr($3); - if ($1) { - ID vid = $1->nd_vid; - if ($2 == tOROP) { - $1->nd_value = $3; - $$ = NEW_OP_ASGN_OR(gettable(vid), $1); - if (is_asgn_or_id(vid)) { - $$->nd_aid = vid; - } - } - else if ($2 == tANDOP) { - $1->nd_value = $3; - $$ = NEW_OP_ASGN_AND(gettable(vid), $1); - } - else { - $$ = $1; - $$->nd_value = call_op(gettable(vid),$2,1,$3); - } - } - else { - $$ = 0; - } + $$ = logop(p, idAND, $1, $3, &@2, &@$); } - | primary_value '[' aref_args ']' tOP_ASGN command_call + | expr keyword_or expr { - NODE *args; - - value_expr($6); - if (!$3) $3 = NEW_ZARRAY(); - args = arg_concat($6, $3); - if ($5 == tOROP) { - $5 = 0; - } - else if ($5 == tANDOP) { - $5 = 1; - } - $$ = NEW_OP_ASGN1($1, $5, args); - fixpos($$, $1); + $$ = logop(p, idOR, $1, $3, &@2, &@$); } - | primary_value '.' tIDENTIFIER tOP_ASGN command_call + | keyword_not opt_nl expr { - value_expr($5); - if ($4 == tOROP) { - $4 = 0; - } - else if ($4 == tANDOP) { - $4 = 1; - } - $$ = NEW_OP_ASGN2($1, $3, $4, $5); - fixpos($$, $1); + $$ = call_uni_op(p, method_cond(p, $3, &@3), METHOD_NOT, &@1, &@$); } - | primary_value '.' tCONSTANT tOP_ASGN command_call + | '!' command_call { - value_expr($5); - if ($4 == tOROP) { - $4 = 0; - } - else if ($4 == tANDOP) { - $4 = 1; - } - $$ = NEW_OP_ASGN2($1, $3, $4, $5); - fixpos($$, $1); + $$ = call_uni_op(p, method_cond(p, $2, &@2), '!', &@1, &@$); } - | primary_value tCOLON2 tIDENTIFIER tOP_ASGN command_call + | arg tASSOC { - value_expr($5); - if ($4 == tOROP) { - $4 = 0; - } - else if ($4 == tANDOP) { - $4 = 1; - } - $$ = NEW_OP_ASGN2($1, $3, $4, $5); - fixpos($$, $1); + value_expr($1); + SET_LEX_STATE(EXPR_BEG|EXPR_LABEL); + p->command_start = FALSE; + $<ctxt>2 = p->ctxt; + p->ctxt.in_kwarg = 1; + $<tbl>$ = push_pvtbl(p); } - | backref tOP_ASGN command_call { - rb_backref_error($1); - $$ = 0; + $<tbl>$ = push_pktbl(p); } - | lhs '=' mrhs + p_top_expr_body { - $$ = node_assign($1, NEW_SVALUE($3)); + pop_pktbl(p, $<tbl>4); + pop_pvtbl(p, $<tbl>3); + p->ctxt.in_kwarg = $<ctxt>2.in_kwarg; + /*%%%*/ + $$ = NEW_CASE3($1, NEW_IN($5, 0, 0, &@5), &@$); + /*% %*/ + /*% ripper: case!($1, in!($5, Qnil, Qnil)) %*/ } - | mlhs '=' arg_value + | arg keyword_in { - $1->nd_value = ($1->nd_head) ? NEW_TO_ARY($3) : NEW_ARRAY($3); - $$ = $1; + value_expr($1); + SET_LEX_STATE(EXPR_BEG|EXPR_LABEL); + p->command_start = FALSE; + $<ctxt>2 = p->ctxt; + p->ctxt.in_kwarg = 1; + $<tbl>$ = push_pvtbl(p); } - | mlhs '=' mrhs { - $1->nd_value = $3; - $$ = $1; + $<tbl>$ = push_pktbl(p); } - | expr + p_top_expr_body + { + pop_pktbl(p, $<tbl>4); + pop_pvtbl(p, $<tbl>3); + p->ctxt.in_kwarg = $<ctxt>2.in_kwarg; + /*%%%*/ + $$ = NEW_CASE3($1, NEW_IN($5, NEW_TRUE(&@5), NEW_FALSE(&@5), &@5), &@$); + /*% %*/ + /*% ripper: case!($1, in!($5, Qnil, Qnil)) %*/ + } + | arg %prec tLBRACE_ARG ; -expr : command_call - | expr kAND expr - { - $$ = logop(NODE_AND, $1, $3); +def_name : fname + { + ID fname = get_id($1); + ID cur_arg = p->cur_arg; + YYSTYPE c = {.ctxt = p->ctxt}; + numparam_name(p, fname); + local_push(p, 0); + p->cur_arg = 0; + p->ctxt.in_def = 1; + $<node>$ = NEW_NODE(NODE_SELF, /*vid*/cur_arg, /*mid*/fname, /*cval*/c.val, &@$); + /*%%%*/ + /*% + $$ = NEW_RIPPER(fname, get_value($1), $$, &NULL_LOC); + %*/ } - | expr kOR expr + ; + +defn_head : k_def def_name { - $$ = logop(NODE_OR, $1, $3); + $$ = $2; + /*%%%*/ + $$ = NEW_NODE(NODE_DEFN, 0, $$->nd_mid, $$, &@$); + /*% %*/ } - | kNOT expr + ; + +defs_head : k_def singleton dot_or_colon { - $$ = NEW_NOT(cond($2)); + SET_LEX_STATE(EXPR_FNAME); + p->ctxt.in_argdef = 1; } - | '!' command_call + def_name { - $$ = NEW_NOT(cond($2)); + SET_LEX_STATE(EXPR_ENDFN|EXPR_LABEL); /* force for args */ + $$ = $5; + /*%%%*/ + $$ = NEW_NODE(NODE_DEFS, $2, $$->nd_mid, $$, &@$); + /*% + VALUE ary = rb_ary_new_from_args(3, $2, $3, get_value($$)); + add_mark_object(p, ary); + $<node>$->nd_rval = ary; + %*/ } - | arg ; expr_value : expr { - value_expr($$); + value_expr($1); $$ = $1; } ; -command_call : command - | block_command - | kRETURN call_args - { - $$ = NEW_RETURN(ret_args($2)); - } - | kBREAK call_args - { - $$ = NEW_BREAK(ret_args($2)); - } - | kNEXT call_args +expr_value_do : {COND_PUSH(1);} expr_value do {COND_POP();} { - $$ = NEW_NEXT(ret_args($2)); + $$ = $2; } ; +command_call : command + | block_command + ; + block_command : block_call - | block_call '.' operation2 command_args - { - $$ = new_call($1, $3, $4); - } - | block_call tCOLON2 operation2 command_args + | block_call call_op2 operation2 command_args { - $$ = new_call($1, $3, $4); + /*%%%*/ + $$ = new_qcall(p, $2, $1, $3, $4, &@3, &@$); + /*% %*/ + /*% ripper: method_add_arg!(call!($1, $2, $3), $4) %*/ } ; -cmd_brace_block : tLBRACE_ARG +cmd_brace_block : tLBRACE_ARG brace_body '}' { - $<vars>$ = dyna_push(); - $<num>1 = ruby_sourceline; + $$ = $2; + /*%%%*/ + $$->nd_body->nd_loc = code_loc_gen(&@1, &@3); + nd_set_line($$, @1.end_pos.lineno); + /*% %*/ } - opt_block_var {$<vars>$ = ruby_dyna_vars;} - compstmt - '}' + ; + +fcall : operation { - $$ = NEW_ITER($3, 0, dyna_init($5, $<vars>4)); - nd_set_line($$, $<num>1); - dyna_pop($<vars>2); + /*%%%*/ + $$ = NEW_FCALL($1, 0, &@$); + nd_set_line($$, p->tokline); + /*% %*/ + /*% ripper: $1 %*/ } ; -command : operation command_args %prec tLOWEST +command : fcall command_args %prec tLOWEST { - $$ = new_fcall($1, $2); - fixpos($$, $2); - } - | operation command_args cmd_brace_block + /*%%%*/ + $1->nd_args = $2; + nd_set_last_loc($1, @2.end_pos); + $$ = $1; + /*% %*/ + /*% ripper: command!($1, $2) %*/ + } + | fcall command_args cmd_brace_block { - $$ = new_fcall($1, $2); - if ($3) { - if (nd_type($$) == NODE_BLOCK_PASS) { - rb_compile_error("both block arg and actual block given"); - } - $3->nd_iter = $$; - $$ = $3; - } - fixpos($$, $2); - } - | primary_value '.' operation2 command_args %prec tLOWEST + /*%%%*/ + block_dup_check(p, $2, $3); + $1->nd_args = $2; + $$ = method_add_block(p, $1, $3, &@$); + fixpos($$, $1); + nd_set_last_loc($1, @2.end_pos); + /*% %*/ + /*% ripper: method_add_block!(command!($1, $2), $3) %*/ + } + | primary_value call_op operation2 command_args %prec tLOWEST { - $$ = new_call($1, $3, $4); - fixpos($$, $1); + /*%%%*/ + $$ = new_command_qcall(p, $2, $1, $3, $4, Qnull, &@3, &@$); + /*% %*/ + /*% ripper: command_call!($1, $2, $3, $4) %*/ } - | primary_value '.' operation2 command_args cmd_brace_block + | primary_value call_op operation2 command_args cmd_brace_block { - $$ = new_call($1, $3, $4); - if ($5) { - if (nd_type($$) == NODE_BLOCK_PASS) { - rb_compile_error("both block arg and actual block given"); - } - $5->nd_iter = $$; - $$ = $5; - } - fixpos($$, $1); - } + /*%%%*/ + $$ = new_command_qcall(p, $2, $1, $3, $4, $5, &@3, &@$); + /*% %*/ + /*% ripper: method_add_block!(command_call!($1, $2, $3, $4), $5) %*/ + } | primary_value tCOLON2 operation2 command_args %prec tLOWEST { - $$ = new_call($1, $3, $4); - fixpos($$, $1); + /*%%%*/ + $$ = new_command_qcall(p, ID2VAL(idCOLON2), $1, $3, $4, Qnull, &@3, &@$); + /*% %*/ + /*% ripper: command_call!($1, $2, $3, $4) %*/ } | primary_value tCOLON2 operation2 command_args cmd_brace_block { - $$ = new_call($1, $3, $4); - if ($5) { - if (nd_type($$) == NODE_BLOCK_PASS) { - rb_compile_error("both block arg and actual block given"); - } - $5->nd_iter = $$; - $$ = $5; - } - fixpos($$, $1); + /*%%%*/ + $$ = new_command_qcall(p, ID2VAL(idCOLON2), $1, $3, $4, $5, &@3, &@$); + /*% %*/ + /*% ripper: method_add_block!(command_call!($1, $2, $3, $4), $5) %*/ } - | kSUPER command_args + | keyword_super command_args + { + /*%%%*/ + $$ = NEW_SUPER($2, &@$); + fixpos($$, $2); + /*% %*/ + /*% ripper: super!($2) %*/ + } + | keyword_yield command_args + { + /*%%%*/ + $$ = new_yield(p, $2, &@$); + fixpos($$, $2); + /*% %*/ + /*% ripper: yield!($2) %*/ + } + | k_return call_args + { + /*%%%*/ + $$ = NEW_RETURN(ret_args(p, $2), &@$); + /*% %*/ + /*% ripper: return!($2) %*/ + } + | keyword_break call_args { - $$ = new_super($2); - fixpos($$, $2); + /*%%%*/ + $$ = NEW_BREAK(ret_args(p, $2), &@$); + /*% %*/ + /*% ripper: break!($2) %*/ } - | kYIELD command_args + | keyword_next call_args { - $$ = new_yield($2); - fixpos($$, $2); + /*%%%*/ + $$ = NEW_NEXT(ret_args(p, $2), &@$); + /*% %*/ + /*% ripper: next!($2) %*/ } ; mlhs : mlhs_basic - | tLPAREN mlhs_entry ')' + | tLPAREN mlhs_inner rparen { + /*%%%*/ $$ = $2; + /*% %*/ + /*% ripper: mlhs_paren!($2) %*/ } ; -mlhs_entry : mlhs_basic - | tLPAREN mlhs_entry ')' +mlhs_inner : mlhs_basic + | tLPAREN mlhs_inner rparen { - $$ = NEW_MASGN(NEW_LIST($2), 0); + /*%%%*/ + $$ = NEW_MASGN(NEW_LIST($2, &@$), 0, &@$); + /*% %*/ + /*% ripper: mlhs_paren!($2) %*/ } ; mlhs_basic : mlhs_head { - $$ = NEW_MASGN($1, 0); + /*%%%*/ + $$ = NEW_MASGN($1, 0, &@$); + /*% %*/ + /*% ripper: $1 %*/ } | mlhs_head mlhs_item { - $$ = NEW_MASGN(list_append($1,$2), 0); + /*%%%*/ + $$ = NEW_MASGN(list_append(p, $1,$2), 0, &@$); + /*% %*/ + /*% ripper: mlhs_add!($1, $2) %*/ } | mlhs_head tSTAR mlhs_node { - $$ = NEW_MASGN($1, $3); + /*%%%*/ + $$ = NEW_MASGN($1, $3, &@$); + /*% %*/ + /*% ripper: mlhs_add_star!($1, $3) %*/ + } + | mlhs_head tSTAR mlhs_node ',' mlhs_post + { + /*%%%*/ + $$ = NEW_MASGN($1, NEW_POSTARG($3,$5,&@$), &@$); + /*% %*/ + /*% ripper: mlhs_add_post!(mlhs_add_star!($1, $3), $5) %*/ } | mlhs_head tSTAR { - $$ = NEW_MASGN($1, -1); + /*%%%*/ + $$ = NEW_MASGN($1, NODE_SPECIAL_NO_NAME_REST, &@$); + /*% %*/ + /*% ripper: mlhs_add_star!($1, Qnil) %*/ + } + | mlhs_head tSTAR ',' mlhs_post + { + /*%%%*/ + $$ = NEW_MASGN($1, NEW_POSTARG(NODE_SPECIAL_NO_NAME_REST, $4, &@$), &@$); + /*% %*/ + /*% ripper: mlhs_add_post!(mlhs_add_star!($1, Qnil), $4) %*/ } | tSTAR mlhs_node { - $$ = NEW_MASGN(0, $2); + /*%%%*/ + $$ = NEW_MASGN(0, $2, &@$); + /*% %*/ + /*% ripper: mlhs_add_star!(mlhs_new!, $2) %*/ + } + | tSTAR mlhs_node ',' mlhs_post + { + /*%%%*/ + $$ = NEW_MASGN(0, NEW_POSTARG($2,$4,&@$), &@$); + /*% %*/ + /*% ripper: mlhs_add_post!(mlhs_add_star!(mlhs_new!, $2), $4) %*/ } | tSTAR { - $$ = NEW_MASGN(0, -1); + /*%%%*/ + $$ = NEW_MASGN(0, NODE_SPECIAL_NO_NAME_REST, &@$); + /*% %*/ + /*% ripper: mlhs_add_star!(mlhs_new!, Qnil) %*/ + } + | tSTAR ',' mlhs_post + { + /*%%%*/ + $$ = NEW_MASGN(0, NEW_POSTARG(NODE_SPECIAL_NO_NAME_REST, $3, &@$), &@$); + /*% %*/ + /*% ripper: mlhs_add_post!(mlhs_add_star!(mlhs_new!, Qnil), $3) %*/ } ; mlhs_item : mlhs_node - | tLPAREN mlhs_entry ')' + | tLPAREN mlhs_inner rparen { + /*%%%*/ $$ = $2; + /*% %*/ + /*% ripper: mlhs_paren!($2) %*/ } ; mlhs_head : mlhs_item ',' { - $$ = NEW_LIST($1); + /*%%%*/ + $$ = NEW_LIST($1, &@1); + /*% %*/ + /*% ripper: mlhs_add!(mlhs_new!, $1) %*/ } | mlhs_head mlhs_item ',' { - $$ = list_append($1, $2); + /*%%%*/ + $$ = list_append(p, $1, $2); + /*% %*/ + /*% ripper: mlhs_add!($1, $2) %*/ } ; -mlhs_node : variable +mlhs_post : mlhs_item { - $$ = assignable($1, 0); + /*%%%*/ + $$ = NEW_LIST($1, &@$); + /*% %*/ + /*% ripper: mlhs_add!(mlhs_new!, $1) %*/ } - | primary_value '[' aref_args ']' + | mlhs_post ',' mlhs_item { - $$ = aryset($1, $3); + /*%%%*/ + $$ = list_append(p, $1, $3); + /*% %*/ + /*% ripper: mlhs_add!($1, $3) %*/ } - | primary_value '.' tIDENTIFIER + ; + +mlhs_node : user_variable + { + /*%%%*/ + $$ = assignable(p, $1, 0, &@$); + /*% %*/ + /*% ripper: assignable(p, var_field(p, $1)) %*/ + } + | keyword_variable + { + /*%%%*/ + $$ = assignable(p, $1, 0, &@$); + /*% %*/ + /*% ripper: assignable(p, var_field(p, $1)) %*/ + } + | primary_value '[' opt_call_args rbracket + { + /*%%%*/ + $$ = aryset(p, $1, $3, &@$); + /*% %*/ + /*% ripper: aref_field!($1, escape_Qundef($3)) %*/ + } + | primary_value call_op tIDENTIFIER { - $$ = attrset($1, $3); + if ($2 == tANDDOT) { + yyerror1(&@2, "&. inside multiple assignment destination"); + } + /*%%%*/ + $$ = attrset(p, $1, $2, $3, &@$); + /*% %*/ + /*% ripper: field!($1, $2, $3) %*/ } | primary_value tCOLON2 tIDENTIFIER { - $$ = attrset($1, $3); + /*%%%*/ + $$ = attrset(p, $1, idCOLON2, $3, &@$); + /*% %*/ + /*% ripper: const_path_field!($1, $3) %*/ } - | primary_value '.' tCONSTANT + | primary_value call_op tCONSTANT { - $$ = attrset($1, $3); + if ($2 == tANDDOT) { + yyerror1(&@2, "&. inside multiple assignment destination"); + } + /*%%%*/ + $$ = attrset(p, $1, $2, $3, &@$); + /*% %*/ + /*% ripper: field!($1, $2, $3) %*/ } | primary_value tCOLON2 tCONSTANT { - if (in_def || in_single) - yyerror("dynamic constant assignment"); - $$ = NEW_CDECL(0, 0, NEW_COLON2($1, $3)); + /*%%%*/ + $$ = const_decl(p, NEW_COLON2($1, $3, &@$), &@$); + /*% %*/ + /*% ripper: const_decl(p, const_path_field!($1, $3)) %*/ } | tCOLON3 tCONSTANT { - if (in_def || in_single) - yyerror("dynamic constant assignment"); - $$ = NEW_CDECL(0, 0, NEW_COLON3($2)); + /*%%%*/ + $$ = const_decl(p, NEW_COLON3($2, &@$), &@$); + /*% %*/ + /*% ripper: const_decl(p, top_const_field!($2)) %*/ } | backref { - rb_backref_error($1); - $$ = 0; + /*%%%*/ + rb_backref_error(p, $1); + $$ = NEW_BEGIN(0, &@$); + /*% %*/ + /*% ripper[error]: backref_error(p, RNODE($1), var_field(p, $1)) %*/ } ; -lhs : variable +lhs : user_variable + { + /*%%%*/ + $$ = assignable(p, $1, 0, &@$); + /*% %*/ + /*% ripper: assignable(p, var_field(p, $1)) %*/ + } + | keyword_variable { - $$ = assignable($1, 0); + /*%%%*/ + $$ = assignable(p, $1, 0, &@$); + /*% %*/ + /*% ripper: assignable(p, var_field(p, $1)) %*/ } - | primary_value '[' aref_args ']' + | primary_value '[' opt_call_args rbracket { - $$ = aryset($1, $3); + /*%%%*/ + $$ = aryset(p, $1, $3, &@$); + /*% %*/ + /*% ripper: aref_field!($1, escape_Qundef($3)) %*/ } - | primary_value '.' tIDENTIFIER + | primary_value call_op tIDENTIFIER { - $$ = attrset($1, $3); + /*%%%*/ + $$ = attrset(p, $1, $2, $3, &@$); + /*% %*/ + /*% ripper: field!($1, $2, $3) %*/ } | primary_value tCOLON2 tIDENTIFIER { - $$ = attrset($1, $3); + /*%%%*/ + $$ = attrset(p, $1, idCOLON2, $3, &@$); + /*% %*/ + /*% ripper: field!($1, idCOLON2, $3) %*/ } - | primary_value '.' tCONSTANT + | primary_value call_op tCONSTANT { - $$ = attrset($1, $3); + /*%%%*/ + $$ = attrset(p, $1, $2, $3, &@$); + /*% %*/ + /*% ripper: field!($1, $2, $3) %*/ } | primary_value tCOLON2 tCONSTANT { - if (in_def || in_single) - yyerror("dynamic constant assignment"); - $$ = NEW_CDECL(0, 0, NEW_COLON2($1, $3)); + /*%%%*/ + $$ = const_decl(p, NEW_COLON2($1, $3, &@$), &@$); + /*% %*/ + /*% ripper: const_decl(p, const_path_field!($1, $3)) %*/ } | tCOLON3 tCONSTANT { - if (in_def || in_single) - yyerror("dynamic constant assignment"); - $$ = NEW_CDECL(0, 0, NEW_COLON3($2)); + /*%%%*/ + $$ = const_decl(p, NEW_COLON3($2, &@$), &@$); + /*% %*/ + /*% ripper: const_decl(p, top_const_field!($2)) %*/ } | backref { - rb_backref_error($1); - $$ = 0; + /*%%%*/ + rb_backref_error(p, $1); + $$ = NEW_BEGIN(0, &@$); + /*% %*/ + /*% ripper[error]: backref_error(p, RNODE($1), var_field(p, $1)) %*/ } ; cname : tIDENTIFIER { - yyerror("class/module name must be CONSTANT"); + static const char mesg[] = "class/module name must be CONSTANT"; + /*%%%*/ + yyerror1(&@1, mesg); + /*% %*/ + /*% ripper[error]: class_name_error!(ERR_MESG(), $1) %*/ } | tCONSTANT ; cpath : tCOLON3 cname { - $$ = NEW_COLON3($2); + /*%%%*/ + $$ = NEW_COLON3($2, &@$); + /*% %*/ + /*% ripper: top_const_ref!($2) %*/ } | cname { - $$ = NEW_COLON2(0, $$); + /*%%%*/ + $$ = NEW_COLON2(0, $$, &@$); + /*% %*/ + /*% ripper: const_ref!($1) %*/ } | primary_value tCOLON2 cname { - $$ = NEW_COLON2($1, $3); + /*%%%*/ + $$ = NEW_COLON2($1, $3, &@$); + /*% %*/ + /*% ripper: const_path_ref!($1, $3) %*/ } ; @@ -904,330 +2290,361 @@ fname : tIDENTIFIER | tFID | op { - lex_state = EXPR_END; + SET_LEX_STATE(EXPR_ENDFN); $$ = $1; } | reswords - { - lex_state = EXPR_END; - $$ = $<id>1; - } ; -fsym : fname - | symbol - ; - -fitem : fsym +fitem : fname { - $$ = NEW_LIT(ID2SYM($1)); + /*%%%*/ + $$ = NEW_LIT(ID2SYM($1), &@$); + /*% %*/ + /*% ripper: symbol_literal!($1) %*/ } - | dsym + | symbol ; undef_list : fitem { - $$ = NEW_UNDEF($1); + /*%%%*/ + $$ = NEW_UNDEF($1, &@$); + /*% %*/ + /*% ripper: rb_ary_new3(1, get_value($1)) %*/ } - | undef_list ',' {lex_state = EXPR_FNAME;} fitem + | undef_list ',' {SET_LEX_STATE(EXPR_FNAME|EXPR_FITEM);} fitem { - $$ = block_append($1, NEW_UNDEF($4)); + /*%%%*/ + NODE *undef = NEW_UNDEF($4, &@4); + $$ = block_append(p, $1, undef); + /*% %*/ + /*% ripper: rb_ary_push($1, get_value($4)) %*/ } ; -op : '|' { $$ = '|'; } - | '^' { $$ = '^'; } - | '&' { $$ = '&'; } - | tCMP { $$ = tCMP; } - | tEQ { $$ = tEQ; } - | tEQQ { $$ = tEQQ; } - | tMATCH { $$ = tMATCH; } - | '>' { $$ = '>'; } - | tGEQ { $$ = tGEQ; } - | '<' { $$ = '<'; } - | tLEQ { $$ = tLEQ; } - | tLSHFT { $$ = tLSHFT; } - | tRSHFT { $$ = tRSHFT; } - | '+' { $$ = '+'; } - | '-' { $$ = '-'; } - | '*' { $$ = '*'; } - | tSTAR { $$ = '*'; } - | '/' { $$ = '/'; } - | '%' { $$ = '%'; } - | tPOW { $$ = tPOW; } - | '~' { $$ = '~'; } - | tUPLUS { $$ = tUPLUS; } - | tUMINUS { $$ = tUMINUS; } - | tAREF { $$ = tAREF; } - | tASET { $$ = tASET; } - | '`' { $$ = '`'; } +op : '|' { ifndef_ripper($$ = '|'); } + | '^' { ifndef_ripper($$ = '^'); } + | '&' { ifndef_ripper($$ = '&'); } + | tCMP { ifndef_ripper($$ = tCMP); } + | tEQ { ifndef_ripper($$ = tEQ); } + | tEQQ { ifndef_ripper($$ = tEQQ); } + | tMATCH { ifndef_ripper($$ = tMATCH); } + | tNMATCH { ifndef_ripper($$ = tNMATCH); } + | '>' { ifndef_ripper($$ = '>'); } + | tGEQ { ifndef_ripper($$ = tGEQ); } + | '<' { ifndef_ripper($$ = '<'); } + | tLEQ { ifndef_ripper($$ = tLEQ); } + | tNEQ { ifndef_ripper($$ = tNEQ); } + | tLSHFT { ifndef_ripper($$ = tLSHFT); } + | tRSHFT { ifndef_ripper($$ = tRSHFT); } + | '+' { ifndef_ripper($$ = '+'); } + | '-' { ifndef_ripper($$ = '-'); } + | '*' { ifndef_ripper($$ = '*'); } + | tSTAR { ifndef_ripper($$ = '*'); } + | '/' { ifndef_ripper($$ = '/'); } + | '%' { ifndef_ripper($$ = '%'); } + | tPOW { ifndef_ripper($$ = tPOW); } + | tDSTAR { ifndef_ripper($$ = tDSTAR); } + | '!' { ifndef_ripper($$ = '!'); } + | '~' { ifndef_ripper($$ = '~'); } + | tUPLUS { ifndef_ripper($$ = tUPLUS); } + | tUMINUS { ifndef_ripper($$ = tUMINUS); } + | tAREF { ifndef_ripper($$ = tAREF); } + | tASET { ifndef_ripper($$ = tASET); } + | '`' { ifndef_ripper($$ = '`'); } ; -reswords : k__LINE__ | k__FILE__ | klBEGIN | klEND - | kALIAS | kAND | kBEGIN | kBREAK | kCASE | kCLASS | kDEF - | kDEFINED | kDO | kELSE | kELSIF | kEND | kENSURE | kFALSE - | kFOR | kIN | kMODULE | kNEXT | kNIL | kNOT - | kOR | kREDO | kRESCUE | kRETRY | kRETURN | kSELF | kSUPER - | kTHEN | kTRUE | kUNDEF | kWHEN | kYIELD - | kIF | kUNLESS | kWHILE | kUNTIL +reswords : keyword__LINE__ | keyword__FILE__ | keyword__ENCODING__ + | keyword_BEGIN | keyword_END + | keyword_alias | keyword_and | keyword_begin + | keyword_break | keyword_case | keyword_class | keyword_def + | keyword_defined | keyword_do | keyword_else | keyword_elsif + | keyword_end | keyword_ensure | keyword_false + | keyword_for | keyword_in | keyword_module | keyword_next + | keyword_nil | keyword_not | keyword_or | keyword_redo + | keyword_rescue | keyword_retry | keyword_return | keyword_self + | keyword_super | keyword_then | keyword_true | keyword_undef + | keyword_when | keyword_yield | keyword_if | keyword_unless + | keyword_while | keyword_until ; -arg : lhs '=' arg - { - $$ = node_assign($1, $3); - } - | lhs '=' arg kRESCUE_MOD arg +arg : lhs '=' lex_ctxt arg_rhs { - $$ = node_assign($1, NEW_RESCUE($3, NEW_RESBODY(0,$5,0), 0)); + /*%%%*/ + $$ = node_assign(p, $1, $4, $3, &@$); + /*% %*/ + /*% ripper: assign!($1, $4) %*/ } - | var_lhs tOP_ASGN arg + | var_lhs tOP_ASGN lex_ctxt arg_rhs { - value_expr($3); - if ($1) { - ID vid = $1->nd_vid; - if ($2 == tOROP) { - $1->nd_value = $3; - $$ = NEW_OP_ASGN_OR(gettable(vid), $1); - if (is_asgn_or_id(vid)) { - $$->nd_aid = vid; - } - } - else if ($2 == tANDOP) { - $1->nd_value = $3; - $$ = NEW_OP_ASGN_AND(gettable(vid), $1); - } - else { - $$ = $1; - $$->nd_value = call_op(gettable(vid),$2,1,$3); - } - } - else { - $$ = 0; - } + /*%%%*/ + $$ = new_op_assign(p, $1, $2, $4, $3, &@$); + /*% %*/ + /*% ripper: opassign!($1, $2, $4) %*/ } - | primary_value '[' aref_args ']' tOP_ASGN arg + | primary_value '[' opt_call_args rbracket tOP_ASGN lex_ctxt arg_rhs { - NODE *args; - - value_expr($6); - if (!$3) $3 = NEW_ZARRAY(); - args = arg_concat($6, $3); - if ($5 == tOROP) { - $5 = 0; - } - else if ($5 == tANDOP) { - $5 = 1; - } - $$ = NEW_OP_ASGN1($1, $5, args); - fixpos($$, $1); + /*%%%*/ + $$ = new_ary_op_assign(p, $1, $3, $5, $7, &@3, &@$); + /*% %*/ + /*% ripper: opassign!(aref_field!($1, escape_Qundef($3)), $5, $7) %*/ } - | primary_value '.' tIDENTIFIER tOP_ASGN arg + | primary_value call_op tIDENTIFIER tOP_ASGN lex_ctxt arg_rhs { - value_expr($5); - if ($4 == tOROP) { - $4 = 0; - } - else if ($4 == tANDOP) { - $4 = 1; - } - $$ = NEW_OP_ASGN2($1, $3, $4, $5); - fixpos($$, $1); + /*%%%*/ + $$ = new_attr_op_assign(p, $1, $2, $3, $4, $6, &@$); + /*% %*/ + /*% ripper: opassign!(field!($1, $2, $3), $4, $6) %*/ } - | primary_value '.' tCONSTANT tOP_ASGN arg + | primary_value call_op tCONSTANT tOP_ASGN lex_ctxt arg_rhs { - value_expr($5); - if ($4 == tOROP) { - $4 = 0; - } - else if ($4 == tANDOP) { - $4 = 1; - } - $$ = NEW_OP_ASGN2($1, $3, $4, $5); - fixpos($$, $1); + /*%%%*/ + $$ = new_attr_op_assign(p, $1, $2, $3, $4, $6, &@$); + /*% %*/ + /*% ripper: opassign!(field!($1, $2, $3), $4, $6) %*/ } - | primary_value tCOLON2 tIDENTIFIER tOP_ASGN arg + | primary_value tCOLON2 tIDENTIFIER tOP_ASGN lex_ctxt arg_rhs { - value_expr($5); - if ($4 == tOROP) { - $4 = 0; - } - else if ($4 == tANDOP) { - $4 = 1; - } - $$ = NEW_OP_ASGN2($1, $3, $4, $5); - fixpos($$, $1); + /*%%%*/ + $$ = new_attr_op_assign(p, $1, ID2VAL(idCOLON2), $3, $4, $6, &@$); + /*% %*/ + /*% ripper: opassign!(field!($1, $2, $3), $4, $6) %*/ } - | primary_value tCOLON2 tCONSTANT tOP_ASGN arg + | primary_value tCOLON2 tCONSTANT tOP_ASGN lex_ctxt arg_rhs { - yyerror("constant re-assignment"); - $$ = 0; + /*%%%*/ + YYLTYPE loc = code_loc_gen(&@1, &@3); + $$ = new_const_op_assign(p, NEW_COLON2($1, $3, &loc), $4, $6, $5, &@$); + /*% %*/ + /*% ripper: opassign!(const_path_field!($1, $3), $4, $6) %*/ } - | tCOLON3 tCONSTANT tOP_ASGN arg + | tCOLON3 tCONSTANT tOP_ASGN lex_ctxt arg_rhs { - yyerror("constant re-assignment"); - $$ = 0; + /*%%%*/ + YYLTYPE loc = code_loc_gen(&@1, &@2); + $$ = new_const_op_assign(p, NEW_COLON3($2, &loc), $3, $5, $4, &@$); + /*% %*/ + /*% ripper: opassign!(top_const_field!($2), $3, $5) %*/ } - | backref tOP_ASGN arg + | backref tOP_ASGN lex_ctxt arg_rhs { - rb_backref_error($1); - $$ = 0; + /*%%%*/ + rb_backref_error(p, $1); + $$ = NEW_BEGIN(0, &@$); + /*% %*/ + /*% ripper[error]: backref_error(p, RNODE($1), opassign!(var_field(p, $1), $2, $4)) %*/ } | arg tDOT2 arg { + /*%%%*/ value_expr($1); value_expr($3); - if (nd_type($1) == NODE_LIT && FIXNUM_P($1->nd_lit) && - nd_type($3) == NODE_LIT && FIXNUM_P($3->nd_lit)) { - $1->nd_lit = rb_range_new($1->nd_lit, $3->nd_lit, Qfalse); - $$ = $1; - } - else { - $$ = NEW_DOT2($1, $3); - } + $$ = NEW_DOT2($1, $3, &@$); + /*% %*/ + /*% ripper: dot2!($1, $3) %*/ } | arg tDOT3 arg { + /*%%%*/ value_expr($1); value_expr($3); - if (nd_type($1) == NODE_LIT && FIXNUM_P($1->nd_lit) && - nd_type($3) == NODE_LIT && FIXNUM_P($3->nd_lit)) { - $1->nd_lit = rb_range_new($1->nd_lit, $3->nd_lit, Qtrue); - $$ = $1; - } - else { - $$ = NEW_DOT3($1, $3); - } + $$ = NEW_DOT3($1, $3, &@$); + /*% %*/ + /*% ripper: dot3!($1, $3) %*/ + } + | arg tDOT2 + { + /*%%%*/ + value_expr($1); + $$ = NEW_DOT2($1, new_nil_at(p, &@2.end_pos), &@$); + /*% %*/ + /*% ripper: dot2!($1, Qnil) %*/ + } + | arg tDOT3 + { + /*%%%*/ + value_expr($1); + $$ = NEW_DOT3($1, new_nil_at(p, &@2.end_pos), &@$); + /*% %*/ + /*% ripper: dot3!($1, Qnil) %*/ + } + | tBDOT2 arg + { + /*%%%*/ + value_expr($2); + $$ = NEW_DOT2(new_nil_at(p, &@1.beg_pos), $2, &@$); + /*% %*/ + /*% ripper: dot2!(Qnil, $2) %*/ + } + | tBDOT3 arg + { + /*%%%*/ + value_expr($2); + $$ = NEW_DOT3(new_nil_at(p, &@1.beg_pos), $2, &@$); + /*% %*/ + /*% ripper: dot3!(Qnil, $2) %*/ } | arg '+' arg { - $$ = call_op($1, '+', 1, $3); + $$ = call_bin_op(p, $1, '+', $3, &@2, &@$); } | arg '-' arg { - $$ = call_op($1, '-', 1, $3); + $$ = call_bin_op(p, $1, '-', $3, &@2, &@$); } | arg '*' arg { - $$ = call_op($1, '*', 1, $3); + $$ = call_bin_op(p, $1, '*', $3, &@2, &@$); } | arg '/' arg { - $$ = call_op($1, '/', 1, $3); + $$ = call_bin_op(p, $1, '/', $3, &@2, &@$); } | arg '%' arg { - $$ = call_op($1, '%', 1, $3); + $$ = call_bin_op(p, $1, '%', $3, &@2, &@$); } | arg tPOW arg { - $$ = call_op($1, tPOW, 1, $3); + $$ = call_bin_op(p, $1, idPow, $3, &@2, &@$); } - | tUMINUS_NUM tINTEGER tPOW arg + | tUMINUS_NUM simple_numeric tPOW arg { - $$ = call_op(call_op($2, tPOW, 1, $4), tUMINUS, 0, 0); - } - | tUMINUS_NUM tFLOAT tPOW arg - { - $$ = call_op(call_op($2, tPOW, 1, $4), tUMINUS, 0, 0); + $$ = call_uni_op(p, call_bin_op(p, $2, idPow, $4, &@2, &@$), idUMinus, &@1, &@$); } | tUPLUS arg { - if ($2 && nd_type($2) == NODE_LIT) { - $$ = $2; - } - else { - $$ = call_op($2, tUPLUS, 0, 0); - } + $$ = call_uni_op(p, $2, idUPlus, &@1, &@$); } | tUMINUS arg { - $$ = call_op($2, tUMINUS, 0, 0); + $$ = call_uni_op(p, $2, idUMinus, &@1, &@$); } | arg '|' arg { - $$ = call_op($1, '|', 1, $3); + $$ = call_bin_op(p, $1, '|', $3, &@2, &@$); } | arg '^' arg { - $$ = call_op($1, '^', 1, $3); + $$ = call_bin_op(p, $1, '^', $3, &@2, &@$); } | arg '&' arg { - $$ = call_op($1, '&', 1, $3); + $$ = call_bin_op(p, $1, '&', $3, &@2, &@$); } | arg tCMP arg { - $$ = call_op($1, tCMP, 1, $3); - } - | arg '>' arg - { - $$ = call_op($1, '>', 1, $3); - } - | arg tGEQ arg - { - $$ = call_op($1, tGEQ, 1, $3); - } - | arg '<' arg - { - $$ = call_op($1, '<', 1, $3); - } - | arg tLEQ arg - { - $$ = call_op($1, tLEQ, 1, $3); + $$ = call_bin_op(p, $1, idCmp, $3, &@2, &@$); } + | rel_expr %prec tCMP | arg tEQ arg { - $$ = call_op($1, tEQ, 1, $3); + $$ = call_bin_op(p, $1, idEq, $3, &@2, &@$); } | arg tEQQ arg { - $$ = call_op($1, tEQQ, 1, $3); + $$ = call_bin_op(p, $1, idEqq, $3, &@2, &@$); } | arg tNEQ arg { - $$ = NEW_NOT(call_op($1, tEQ, 1, $3)); + $$ = call_bin_op(p, $1, idNeq, $3, &@2, &@$); } | arg tMATCH arg { - $$ = match_gen($1, $3); + $$ = match_op(p, $1, $3, &@2, &@$); } | arg tNMATCH arg { - $$ = NEW_NOT(match_gen($1, $3)); + $$ = call_bin_op(p, $1, idNeqTilde, $3, &@2, &@$); } | '!' arg { - $$ = NEW_NOT(cond($2)); + $$ = call_uni_op(p, method_cond(p, $2, &@2), '!', &@1, &@$); } | '~' arg { - $$ = call_op($2, '~', 0, 0); + $$ = call_uni_op(p, $2, '~', &@1, &@$); } | arg tLSHFT arg { - $$ = call_op($1, tLSHFT, 1, $3); + $$ = call_bin_op(p, $1, idLTLT, $3, &@2, &@$); } | arg tRSHFT arg { - $$ = call_op($1, tRSHFT, 1, $3); + $$ = call_bin_op(p, $1, idGTGT, $3, &@2, &@$); } | arg tANDOP arg { - $$ = logop(NODE_AND, $1, $3); + $$ = logop(p, idANDOP, $1, $3, &@2, &@$); } | arg tOROP arg { - $$ = logop(NODE_OR, $1, $3); + $$ = logop(p, idOROP, $1, $3, &@2, &@$); } - | kDEFINED opt_nl {in_defined = 1;} arg + | keyword_defined opt_nl {p->ctxt.in_defined = 1;} arg { - in_defined = 0; - $$ = NEW_DEFINED($4); + p->ctxt.in_defined = 0; + $$ = new_defined(p, $4, &@$); } - | arg '?' arg ':' arg + | arg '?' arg opt_nl ':' arg { - $$ = NEW_IF(cond($1), $3, $5); - fixpos($$, $1); + /*%%%*/ + value_expr($1); + $$ = new_if(p, $1, $3, $6, &@$); + fixpos($$, $1); + /*% %*/ + /*% ripper: ifop!($1, $3, $6) %*/ + } + | defn_head f_opt_paren_args '=' arg + { + endless_method_name(p, $<node>1, &@1); + restore_defun(p, $<node>1->nd_defn); + /*%%%*/ + $$ = set_defun_body(p, $1, $2, $4, &@$); + /*% %*/ + /*% ripper[$4]: bodystmt!($4, Qnil, Qnil, Qnil) %*/ + /*% ripper: def!(get_value($1), $2, $4) %*/ + local_pop(p); + } + | defn_head f_opt_paren_args '=' arg modifier_rescue arg + { + endless_method_name(p, $<node>1, &@1); + restore_defun(p, $<node>1->nd_defn); + /*%%%*/ + $4 = rescued_expr(p, $4, $6, &@4, &@5, &@6); + $$ = set_defun_body(p, $1, $2, $4, &@$); + /*% %*/ + /*% ripper[$4]: bodystmt!(rescue_mod!($4, $6), Qnil, Qnil, Qnil) %*/ + /*% ripper: def!(get_value($1), $2, $4) %*/ + local_pop(p); + } + | defs_head f_opt_paren_args '=' arg + { + endless_method_name(p, $<node>1, &@1); + restore_defun(p, $<node>1->nd_defn); + /*%%%*/ + $$ = set_defun_body(p, $1, $2, $4, &@$); + /*% + $1 = get_value($1); + %*/ + /*% ripper[$4]: bodystmt!($4, Qnil, Qnil, Qnil) %*/ + /*% ripper: defs!(AREF($1, 0), AREF($1, 1), AREF($1, 2), $2, $4) %*/ + local_pop(p); + } + | defs_head f_opt_paren_args '=' arg modifier_rescue arg + { + endless_method_name(p, $<node>1, &@1); + restore_defun(p, $<node>1->nd_defn); + /*%%%*/ + $4 = rescued_expr(p, $4, $6, &@4, &@5, &@6); + $$ = set_defun_body(p, $1, $2, $4, &@$); + /*% + $1 = get_value($1); + %*/ + /*% ripper[$4]: bodystmt!(rescue_mod!($4, $6), Qnil, Qnil, Qnil) %*/ + /*% ripper: defs!(AREF($1, 0), AREF($1, 1), AREF($1, 2), $2, $4) %*/ + local_pop(p); } | primary { @@ -1235,6 +2652,33 @@ arg : lhs '=' arg } ; +relop : '>' {$$ = '>';} + | '<' {$$ = '<';} + | tGEQ {$$ = idGE;} + | tLEQ {$$ = idLE;} + ; + +rel_expr : arg relop arg %prec '>' + { + $$ = call_bin_op(p, $1, $2, $3, &@2, &@$); + } + | rel_expr relop arg %prec '>' + { + rb_warning1("comparison '%s' after comparison", WARN_ID($2)); + $$ = call_bin_op(p, $1, $2, $3, &@2, &@$); + } + ; + +lex_ctxt : tSP + { + $$ = p->ctxt; + } + | none + { + $$ = p->ctxt; + } + ; + arg_value : arg { value_expr($1); @@ -1243,48 +2687,71 @@ arg_value : arg ; aref_args : none - | command opt_nl - { - rb_warn("parenthesize argument(s) for future version"); - $$ = NEW_LIST($1); - } | args trailer { $$ = $1; } - | args ',' tSTAR arg opt_nl + | args ',' assocs trailer { - value_expr($4); - $$ = arg_concat($1, $4); + /*%%%*/ + $$ = $3 ? arg_append(p, $1, new_hash(p, $3, &@3), &@$) : $1; + /*% %*/ + /*% ripper: args_add!($1, bare_assoc_hash!($3)) %*/ } | assocs trailer { - $$ = NEW_LIST(NEW_HASH($1)); - } - | tSTAR arg opt_nl - { - value_expr($2); - $$ = NEW_NEWLINE(NEW_SPLAT($2)); + /*%%%*/ + $$ = $1 ? NEW_LIST(new_hash(p, $1, &@1), &@$) : 0; + /*% %*/ + /*% ripper: args_add!(args_new!, bare_assoc_hash!($1)) %*/ } ; -paren_args : '(' none ')' +arg_rhs : arg %prec tOP_ASGN { - $$ = $2; + value_expr($1); + $$ = $1; + } + | arg modifier_rescue arg + { + /*%%%*/ + value_expr($1); + $$ = rescued_expr(p, $1, $3, &@1, &@2, &@3); + /*% %*/ + /*% ripper: rescue_mod!($1, $3) %*/ } - | '(' call_args opt_nl ')' + ; + +paren_args : '(' opt_call_args rparen { + /*%%%*/ $$ = $2; + /*% %*/ + /*% ripper: arg_paren!(escape_Qundef($2)) %*/ } - | '(' block_call opt_nl ')' + | '(' args ',' args_forward rparen { - rb_warn("parenthesize argument for future version"); - $$ = NEW_LIST($2); + if (!check_forwarding_args(p)) { + $$ = Qnone; + } + else { + /*%%%*/ + $$ = new_args_forward_call(p, $2, &@4, &@$); + /*% %*/ + /*% ripper: arg_paren!(args_add!($2, $4)) %*/ + } } - | '(' args ',' block_call opt_nl ')' + | '(' args_forward rparen { - rb_warn("parenthesize argument for future version"); - $$ = list_append($2, $4); + if (!check_forwarding_args(p)) { + $$ = Qnone; + } + else { + /*%%%*/ + $$ = new_args_forward_call(p, 0, &@2, &@$); + /*% %*/ + /*% ripper: arg_paren!($2) %*/ + } } ; @@ -1292,132 +2759,118 @@ opt_paren_args : none | paren_args ; -call_args : command - { - rb_warn("parenthesize argument(s) for future version"); - $$ = NEW_LIST($1); - } - | args opt_block_arg - { - $$ = arg_blk_pass($1, $2); - } - | args ',' tSTAR arg_value opt_block_arg - { - $$ = arg_concat($1, $4); - $$ = arg_blk_pass($$, $5); - } - | assocs opt_block_arg - { - $$ = NEW_LIST(NEW_HASH($1)); - $$ = arg_blk_pass($$, $2); - } - | assocs ',' tSTAR arg_value opt_block_arg - { - $$ = arg_concat(NEW_LIST(NEW_HASH($1)), $4); - $$ = arg_blk_pass($$, $5); - } - | args ',' assocs opt_block_arg +opt_call_args : none + | call_args + | args ',' { - $$ = list_append($1, NEW_HASH($3)); - $$ = arg_blk_pass($$, $4); + $$ = $1; } - | args ',' assocs ',' tSTAR arg opt_block_arg + | args ',' assocs ',' { - value_expr($6); - $$ = arg_concat(list_append($1, NEW_HASH($3)), $6); - $$ = arg_blk_pass($$, $7); + /*%%%*/ + $$ = $3 ? arg_append(p, $1, new_hash(p, $3, &@3), &@$) : $1; + /*% %*/ + /*% ripper: args_add!($1, bare_assoc_hash!($3)) %*/ } - | tSTAR arg_value opt_block_arg + | assocs ',' { - $$ = arg_blk_pass(NEW_SPLAT($2), $3); + /*%%%*/ + $$ = $1 ? NEW_LIST(new_hash(p, $1, &@1), &@1) : 0; + /*% %*/ + /*% ripper: args_add!(args_new!, bare_assoc_hash!($1)) %*/ } - | block_arg ; -call_args2 : arg_value ',' args opt_block_arg - { - $$ = arg_blk_pass(list_concat(NEW_LIST($1),$3), $4); - } - | arg_value ',' block_arg - { - $$ = arg_blk_pass($1, $3); - } - | arg_value ',' tSTAR arg_value opt_block_arg +call_args : command { - $$ = arg_concat(NEW_LIST($1), $4); - $$ = arg_blk_pass($$, $5); + /*%%%*/ + value_expr($1); + $$ = NEW_LIST($1, &@$); + /*% %*/ + /*% ripper: args_add!(args_new!, $1) %*/ } - | arg_value ',' args ',' tSTAR arg_value opt_block_arg + | args opt_block_arg { - $$ = arg_concat(list_concat(NEW_LIST($1),$3), $6); - $$ = arg_blk_pass($$, $7); + /*%%%*/ + $$ = arg_blk_pass($1, $2); + /*% %*/ + /*% ripper: args_add_block!($1, $2) %*/ } | assocs opt_block_arg { - $$ = NEW_LIST(NEW_HASH($1)); + /*%%%*/ + $$ = $1 ? NEW_LIST(new_hash(p, $1, &@1), &@1) : 0; $$ = arg_blk_pass($$, $2); + /*% %*/ + /*% ripper: args_add_block!(args_add!(args_new!, bare_assoc_hash!($1)), $2) %*/ } - | assocs ',' tSTAR arg_value opt_block_arg - { - $$ = arg_concat(NEW_LIST(NEW_HASH($1)), $4); - $$ = arg_blk_pass($$, $5); - } - | arg_value ',' assocs opt_block_arg + | args ',' assocs opt_block_arg { - $$ = list_append(NEW_LIST($1), NEW_HASH($3)); + /*%%%*/ + $$ = $3 ? arg_append(p, $1, new_hash(p, $3, &@3), &@$) : $1; $$ = arg_blk_pass($$, $4); - } - | arg_value ',' args ',' assocs opt_block_arg - { - $$ = list_append(list_concat(NEW_LIST($1),$3), NEW_HASH($5)); - $$ = arg_blk_pass($$, $6); - } - | arg_value ',' assocs ',' tSTAR arg_value opt_block_arg - { - $$ = arg_concat(list_append(NEW_LIST($1), NEW_HASH($3)), $6); - $$ = arg_blk_pass($$, $7); - } - | arg_value ',' args ',' assocs ',' tSTAR arg_value opt_block_arg - { - $$ = arg_concat(list_append(list_concat(NEW_LIST($1), $3), NEW_HASH($5)), $8); - $$ = arg_blk_pass($$, $9); - } - | tSTAR arg_value opt_block_arg - { - $$ = arg_blk_pass(NEW_SPLAT($2), $3); + /*% %*/ + /*% ripper: args_add_block!(args_add!($1, bare_assoc_hash!($3)), $4) %*/ } | block_arg + /*% ripper[brace]: args_add_block!(args_new!, $1) %*/ ; -command_args : { - $<num>$ = cmdarg_stack; +command_args : { + /* If call_args starts with a open paren '(' or '[', + * look-ahead reading of the letters calls CMDARG_PUSH(0), + * but the push must be done after CMDARG_PUSH(1). + * So this code makes them consistent by first cancelling + * the premature CMDARG_PUSH(0), doing CMDARG_PUSH(1), + * and finally redoing CMDARG_PUSH(0). + */ + int lookahead = 0; + switch (yychar) { + case '(': case tLPAREN: case tLPAREN_ARG: case '[': case tLBRACK: + lookahead = 1; + } + if (lookahead) CMDARG_POP(); CMDARG_PUSH(1); - } - open_args - { - /* CMDARG_POP() */ - cmdarg_stack = $<num>1; - $$ = $2; - } - ; - -open_args : call_args - | tLPAREN_ARG {lex_state = EXPR_ENDARG;} ')' - { - rb_warn("don't put space before argument parentheses"); - $$ = 0; - } - | tLPAREN_ARG call_args2 {lex_state = EXPR_ENDARG;} ')' - { - rb_warn("don't put space before argument parentheses"); + if (lookahead) CMDARG_PUSH(0); + } + call_args + { + /* call_args can be followed by tLBRACE_ARG (that does CMDARG_PUSH(0) in the lexer) + * but the push must be done after CMDARG_POP() in the parser. + * So this code does CMDARG_POP() to pop 0 pushed by tLBRACE_ARG, + * CMDARG_POP() to pop 1 pushed by command_args, + * and CMDARG_PUSH(0) to restore back the flag set by tLBRACE_ARG. + */ + int lookahead = 0; + switch (yychar) { + case tLBRACE_ARG: + lookahead = 1; + } + if (lookahead) CMDARG_POP(); + CMDARG_POP(); + if (lookahead) CMDARG_PUSH(0); $$ = $2; } ; block_arg : tAMPER arg_value { - $$ = NEW_BLOCK_PASS($2); - } + /*%%%*/ + $$ = NEW_BLOCK_PASS($2, &@$); + /*% %*/ + /*% ripper: $2 %*/ + } + | tAMPER + { + /*%%%*/ + if (!local_id(p, ANON_BLOCK_ID)) { + compile_error(p, "no anonymous block parameter"); + } + $$ = NEW_BLOCK_PASS(NEW_LVAR(ANON_BLOCK_ID, &@1), &@$); + /*% + $$ = Qnil; + %*/ + } ; opt_block_arg : ',' block_arg @@ -1425,29 +2878,68 @@ opt_block_arg : ',' block_arg $$ = $2; } | none + { + $$ = 0; + } ; -args : arg_value +/* value */ +args : arg_value + { + /*%%%*/ + $$ = NEW_LIST($1, &@$); + /*% %*/ + /*% ripper: args_add!(args_new!, $1) %*/ + } + | tSTAR arg_value { - $$ = NEW_LIST($1); + /*%%%*/ + $$ = NEW_SPLAT($2, &@$); + /*% %*/ + /*% ripper: args_add_star!(args_new!, $2) %*/ } | args ',' arg_value { - $$ = list_append($1, $3); + /*%%%*/ + $$ = last_arg_append(p, $1, $3, &@$); + /*% %*/ + /*% ripper: args_add!($1, $3) %*/ + } + | args ',' tSTAR arg_value + { + /*%%%*/ + $$ = rest_arg_append(p, $1, $4, &@$); + /*% %*/ + /*% ripper: args_add_star!($1, $4) %*/ } ; +/* value */ +mrhs_arg : mrhs + | arg_value + ; + +/* value */ mrhs : args ',' arg_value { - $$ = list_append($1, $3); + /*%%%*/ + $$ = last_arg_append(p, $1, $3, &@$); + /*% %*/ + /*% ripper: mrhs_add!(mrhs_new_from_args!($1), $3) %*/ } | args ',' tSTAR arg_value { - $$ = arg_concat($1, $4); + /*%%%*/ + $$ = rest_arg_append(p, $1, $4, &@$); + /*% %*/ + /*% ripper: mrhs_add_star!(mrhs_new_from_args!($1), $4) %*/ } | tSTAR arg_value { - $$ = NEW_SPLAT($2); + /*%%%*/ + $$ = NEW_SPLAT($2, &@$); + /*% %*/ + /*% ripper: mrhs_add_star!(mrhs_new!, $2) %*/ } ; @@ -1457,432 +2949,1124 @@ primary : literal | regexp | words | qwords + | symbols + | qsymbols | var_ref | backref | tFID { - $$ = NEW_FCALL($1, 0); + /*%%%*/ + $$ = NEW_FCALL($1, 0, &@$); + /*% %*/ + /*% ripper: method_add_arg!(fcall!($1), args_new!) %*/ } - | kBEGIN + | k_begin { - $<num>1 = ruby_sourceline; + CMDARG_PUSH(0); } bodystmt - kEND + k_end { - if ($3 == NULL) - $$ = NEW_NIL(); - else - $$ = NEW_BEGIN($3); - nd_set_line($$, $<num>1); + CMDARG_POP(); + /*%%%*/ + set_line_body($3, @1.end_pos.lineno); + $$ = NEW_BEGIN($3, &@$); + nd_set_line($$, @1.end_pos.lineno); + /*% %*/ + /*% ripper: begin!($3) %*/ + } + | tLPAREN_ARG {SET_LEX_STATE(EXPR_ENDARG);} rparen + { + /*%%%*/ + $$ = NEW_BEGIN(0, &@$); + /*% %*/ + /*% ripper: paren!(0) %*/ } - | tLPAREN_ARG expr {lex_state = EXPR_ENDARG;} opt_nl ')' + | tLPAREN_ARG stmt {SET_LEX_STATE(EXPR_ENDARG);} rparen { - rb_warning("(...) interpreted as grouped expression"); + /*%%%*/ + if (nd_type_p($2, NODE_SELF)) $2->nd_state = 0; $$ = $2; + /*% %*/ + /*% ripper: paren!($2) %*/ } | tLPAREN compstmt ')' { - if (!$2) $$ = NEW_NIL(); - else $$ = $2; + /*%%%*/ + if (nd_type_p($2, NODE_SELF)) $2->nd_state = 0; + $$ = $2; + /*% %*/ + /*% ripper: paren!($2) %*/ } | primary_value tCOLON2 tCONSTANT { - $$ = NEW_COLON2($1, $3); + /*%%%*/ + $$ = NEW_COLON2($1, $3, &@$); + /*% %*/ + /*% ripper: const_path_ref!($1, $3) %*/ } | tCOLON3 tCONSTANT { - $$ = NEW_COLON3($2); - } - | primary_value '[' aref_args ']' - { - if ($1 && nd_type($1) == NODE_SELF) - $$ = NEW_FCALL(tAREF, $3); - else - $$ = NEW_CALL($1, tAREF, $3); - fixpos($$, $1); + /*%%%*/ + $$ = NEW_COLON3($2, &@$); + /*% %*/ + /*% ripper: top_const_ref!($2) %*/ } | tLBRACK aref_args ']' { - if ($2 == 0) { - $$ = NEW_ZARRAY(); /* zero length array*/ - } - else { - $$ = $2; - } + /*%%%*/ + $$ = make_list($2, &@$); + /*% %*/ + /*% ripper: array!(escape_Qundef($2)) %*/ } | tLBRACE assoc_list '}' { - $$ = NEW_HASH($2); + /*%%%*/ + $$ = new_hash(p, $2, &@$); + $$->nd_brace = TRUE; + /*% %*/ + /*% ripper: hash!(escape_Qundef($2)) %*/ } - | kRETURN + | k_return { - $$ = NEW_RETURN(0); + /*%%%*/ + $$ = NEW_RETURN(0, &@$); + /*% %*/ + /*% ripper: return0! %*/ } - | kYIELD '(' call_args ')' + | keyword_yield '(' call_args rparen { - $$ = new_yield($3); + /*%%%*/ + $$ = new_yield(p, $3, &@$); + /*% %*/ + /*% ripper: yield!(paren!($3)) %*/ } - | kYIELD '(' ')' + | keyword_yield '(' rparen { - $$ = NEW_YIELD(0, Qfalse); + /*%%%*/ + $$ = NEW_YIELD(0, &@$); + /*% %*/ + /*% ripper: yield!(paren!(args_new!)) %*/ } - | kYIELD + | keyword_yield { - $$ = NEW_YIELD(0, Qfalse); + /*%%%*/ + $$ = NEW_YIELD(0, &@$); + /*% %*/ + /*% ripper: yield0! %*/ } - | kDEFINED opt_nl '(' {in_defined = 1;} expr ')' + | keyword_defined opt_nl '(' {p->ctxt.in_defined = 1;} expr rparen { - in_defined = 0; - $$ = NEW_DEFINED($5); + p->ctxt.in_defined = 0; + $$ = new_defined(p, $5, &@$); } - | operation brace_block + | keyword_not '(' expr rparen { - $2->nd_iter = NEW_FCALL($1, 0); - $$ = $2; - fixpos($2->nd_iter, $2); + $$ = call_uni_op(p, method_cond(p, $3, &@3), METHOD_NOT, &@1, &@$); + } + | keyword_not '(' rparen + { + $$ = call_uni_op(p, method_cond(p, new_nil(&@2), &@2), METHOD_NOT, &@1, &@$); + } + | fcall brace_block + { + /*%%%*/ + $$ = method_add_block(p, $1, $2, &@$); + /*% %*/ + /*% ripper: method_add_block!(method_add_arg!(fcall!($1), args_new!), $2) %*/ } | method_call | method_call brace_block { - if ($1 && nd_type($1) == NODE_BLOCK_PASS) { - rb_compile_error("both block arg and actual block given"); - } - $2->nd_iter = $1; - $$ = $2; - fixpos($$, $1); + /*%%%*/ + block_dup_check(p, $1->nd_args, $2); + $$ = method_add_block(p, $1, $2, &@$); + /*% %*/ + /*% ripper: method_add_block!($1, $2) %*/ } - | kIF expr_value then + | lambda + | k_if expr_value then compstmt if_tail - kEND - { - $$ = NEW_IF(cond($2), $4, $5); - fixpos($$, $2); - if (cond_negative(&$$->nd_cond)) { - NODE *tmp = $$->nd_body; - $$->nd_body = $$->nd_else; - $$->nd_else = tmp; - } + k_end + { + /*%%%*/ + $$ = new_if(p, $2, $4, $5, &@$); + fixpos($$, $2); + /*% %*/ + /*% ripper: if!($2, $4, escape_Qundef($5)) %*/ } - | kUNLESS expr_value then + | k_unless expr_value then compstmt opt_else - kEND - { - $$ = NEW_UNLESS(cond($2), $4, $5); - fixpos($$, $2); - if (cond_negative(&$$->nd_cond)) { - NODE *tmp = $$->nd_body; - $$->nd_body = $$->nd_else; - $$->nd_else = tmp; - } + k_end + { + /*%%%*/ + $$ = new_unless(p, $2, $4, $5, &@$); + fixpos($$, $2); + /*% %*/ + /*% ripper: unless!($2, $4, escape_Qundef($5)) %*/ } - | kWHILE {COND_PUSH(1);} expr_value do {COND_POP();} + | k_while expr_value_do compstmt - kEND + k_end { - $$ = NEW_WHILE(cond($3), $6, 1); - fixpos($$, $3); - if (cond_negative(&$$->nd_cond)) { - nd_set_type($$, NODE_UNTIL); - } + /*%%%*/ + $$ = NEW_WHILE(cond(p, $2, &@2), $3, 1, &@$); + fixpos($$, $2); + /*% %*/ + /*% ripper: while!($2, $3) %*/ } - | kUNTIL {COND_PUSH(1);} expr_value do {COND_POP();} + | k_until expr_value_do compstmt - kEND + k_end { - $$ = NEW_UNTIL(cond($3), $6, 1); - fixpos($$, $3); - if (cond_negative(&$$->nd_cond)) { - nd_set_type($$, NODE_WHILE); - } + /*%%%*/ + $$ = NEW_UNTIL(cond(p, $2, &@2), $3, 1, &@$); + fixpos($$, $2); + /*% %*/ + /*% ripper: until!($2, $3) %*/ } - | kCASE expr_value opt_terms - case_body - kEND + | k_case expr_value opt_terms { - $$ = NEW_CASE($2, $4); - fixpos($$, $2); + $<val>$ = p->case_labels; + p->case_labels = Qnil; } - | kCASE opt_terms case_body kEND + case_body + k_end { - $$ = $3; + if (RTEST(p->case_labels)) rb_hash_clear(p->case_labels); + p->case_labels = $<val>4; + /*%%%*/ + $$ = NEW_CASE($2, $5, &@$); + fixpos($$, $2); + /*% %*/ + /*% ripper: case!($2, $5) %*/ } - | kCASE opt_terms kELSE compstmt kEND + | k_case opt_terms { - $$ = $4; + $<val>$ = p->case_labels; + p->case_labels = 0; } - | kFOR block_var kIN {COND_PUSH(1);} expr_value do {COND_POP();} - compstmt - kEND + case_body + k_end { - $$ = NEW_FOR($2, $5, $8); - fixpos($$, $2); + if (RTEST(p->case_labels)) rb_hash_clear(p->case_labels); + p->case_labels = $<val>3; + /*%%%*/ + $$ = NEW_CASE2($4, &@$); + /*% %*/ + /*% ripper: case!(Qnil, $4) %*/ } - | kCLASS cpath superclass + | k_case expr_value opt_terms + p_case_body + k_end { - if (in_def || in_single) - yyerror("class definition in method body"); - class_nest++; - local_push(0); - $<num>$ = ruby_sourceline; + /*%%%*/ + $$ = NEW_CASE3($2, $4, &@$); + /*% %*/ + /*% ripper: case!($2, $4) %*/ + } + | k_for for_var keyword_in expr_value_do + compstmt + k_end + { + /*%%%*/ + /* + * for a, b, c in e + * #=> + * e.each{|*x| a, b, c = x} + * + * for a in e + * #=> + * e.each{|x| a, = x} + */ + ID id = internal_id(p); + NODE *m = NEW_ARGS_AUX(0, 0, &NULL_LOC); + NODE *args, *scope, *internal_var = NEW_DVAR(id, &@2); + rb_ast_id_table_t *tbl = rb_ast_new_local_table(p->ast, 1); + tbl->ids[0] = id; /* internal id */ + + switch (nd_type($2)) { + case NODE_LASGN: + case NODE_DASGN: /* e.each {|internal_var| a = internal_var; ... } */ + $2->nd_value = internal_var; + id = 0; + m->nd_plen = 1; + m->nd_next = $2; + break; + case NODE_MASGN: /* e.each {|*internal_var| a, b, c = (internal_var.length == 1 && Array === (tmp = internal_var[0]) ? tmp : internal_var); ... } */ + m->nd_next = node_assign(p, $2, NEW_FOR_MASGN(internal_var, &@2), NO_LEX_CTXT, &@2); + break; + default: /* e.each {|*internal_var| @a, B, c[1], d.attr = internal_val; ... } */ + m->nd_next = node_assign(p, NEW_MASGN(NEW_LIST($2, &@2), 0, &@2), internal_var, NO_LEX_CTXT, &@2); + } + /* {|*internal_id| <m> = internal_id; ... } */ + args = new_args(p, m, 0, id, 0, new_args_tail(p, 0, 0, 0, &@2), &@2); + scope = NEW_NODE(NODE_SCOPE, tbl, $5, args, &@$); + $$ = NEW_FOR($4, scope, &@$); + fixpos($$, $2); + /*% %*/ + /*% ripper: for!($2, $4, $5) %*/ + } + | k_class cpath superclass + { + if (p->ctxt.in_def) { + YYLTYPE loc = code_loc_gen(&@1, &@2); + yyerror1(&loc, "class definition in method body"); + } + p->ctxt.in_class = 1; + local_push(p, 0); } bodystmt - kEND + k_end { - $$ = NEW_CLASS($2, $5, $3); - nd_set_line($$, $<num>4); - local_pop(); - class_nest--; + /*%%%*/ + $$ = NEW_CLASS($2, $5, $3, &@$); + nd_set_line($$->nd_body, @6.end_pos.lineno); + set_line_body($5, @3.end_pos.lineno); + nd_set_line($$, @3.end_pos.lineno); + /*% %*/ + /*% ripper: class!($2, $3, $5) %*/ + local_pop(p); + p->ctxt.in_class = $<ctxt>1.in_class; + p->ctxt.shareable_constant_value = $<ctxt>1.shareable_constant_value; } - | kCLASS tLSHFT expr + | k_class tLSHFT expr { - $<num>$ = in_def; - in_def = 0; + p->ctxt.in_def = 0; + p->ctxt.in_class = 0; + local_push(p, 0); } term + bodystmt + k_end + { + /*%%%*/ + $$ = NEW_SCLASS($3, $6, &@$); + nd_set_line($$->nd_body, @7.end_pos.lineno); + set_line_body($6, nd_line($3)); + fixpos($$, $3); + /*% %*/ + /*% ripper: sclass!($3, $6) %*/ + local_pop(p); + p->ctxt.in_def = $<ctxt>1.in_def; + p->ctxt.in_class = $<ctxt>1.in_class; + p->ctxt.shareable_constant_value = $<ctxt>1.shareable_constant_value; + } + | k_module cpath + { + if (p->ctxt.in_def) { + YYLTYPE loc = code_loc_gen(&@1, &@2); + yyerror1(&loc, "module definition in method body"); + } + p->ctxt.in_class = 1; + local_push(p, 0); + } + bodystmt + k_end + { + /*%%%*/ + $$ = NEW_MODULE($2, $4, &@$); + nd_set_line($$->nd_body, @5.end_pos.lineno); + set_line_body($4, @2.end_pos.lineno); + nd_set_line($$, @2.end_pos.lineno); + /*% %*/ + /*% ripper: module!($2, $4) %*/ + local_pop(p); + p->ctxt.in_class = $<ctxt>1.in_class; + p->ctxt.shareable_constant_value = $<ctxt>1.shareable_constant_value; + } + | defn_head + f_arglist + bodystmt + k_end { - $<num>$ = in_single; - in_single = 0; - class_nest++; - local_push(0); + restore_defun(p, $<node>1->nd_defn); + /*%%%*/ + $$ = set_defun_body(p, $1, $2, $3, &@$); + /*% %*/ + /*% ripper: def!(get_value($1), $2, $3) %*/ + local_pop(p); } + | defs_head + f_arglist bodystmt - kEND + k_end { - $$ = NEW_SCLASS($3, $7); - fixpos($$, $3); - local_pop(); - class_nest--; - in_def = $<num>4; - in_single = $<num>6; + restore_defun(p, $<node>1->nd_defn); + /*%%%*/ + $$ = set_defun_body(p, $1, $2, $3, &@$); + /*% + $1 = get_value($1); + %*/ + /*% ripper: defs!(AREF($1, 0), AREF($1, 1), AREF($1, 2), $2, $3) %*/ + local_pop(p); } - | kMODULE cpath + | keyword_break { - if (in_def || in_single) - yyerror("module definition in method body"); - class_nest++; - local_push(0); - $<num>$ = ruby_sourceline; + /*%%%*/ + $$ = NEW_BREAK(0, &@$); + /*% %*/ + /*% ripper: break!(args_new!) %*/ } - bodystmt - kEND + | keyword_next { - $$ = NEW_MODULE($2, $4); - nd_set_line($$, $<num>3); - local_pop(); - class_nest--; + /*%%%*/ + $$ = NEW_NEXT(0, &@$); + /*% %*/ + /*% ripper: next!(args_new!) %*/ } - | kDEF fname + | keyword_redo { - $<id>$ = cur_mid; - cur_mid = $2; - in_def++; - local_push(0); + /*%%%*/ + $$ = NEW_REDO(&@$); + /*% %*/ + /*% ripper: redo! %*/ } - f_arglist - bodystmt - kEND + | keyword_retry + { + /*%%%*/ + $$ = NEW_RETRY(&@$); + /*% %*/ + /*% ripper: retry! %*/ + } + ; + +primary_value : primary { - if (!$5) $5 = NEW_NIL(); - $$ = NEW_DEFN($2, $4, $5, NOEX_PRIVATE); - fixpos($$, $4); - local_pop(); - in_def--; - cur_mid = $<id>3; + value_expr($1); + $$ = $1; } - | kDEF singleton dot_or_colon {lex_state = EXPR_FNAME;} fname + ; + +k_begin : keyword_begin { - in_single++; - local_push(0); - lex_state = EXPR_END; /* force for args */ + token_info_push(p, "begin", &@$); } - f_arglist - bodystmt - kEND + ; + +k_if : keyword_if + { + WARN_EOL("if"); + token_info_push(p, "if", &@$); + if (p->token_info && p->token_info->nonspc && + p->token_info->next && !strcmp(p->token_info->next->token, "else")) { + const char *tok = p->lex.ptok; + const char *beg = p->lex.pbeg + p->token_info->next->beg.column; + beg += rb_strlen_lit("else"); + while (beg < tok && ISSPACE(*beg)) beg++; + if (beg == tok) { + p->token_info->nonspc = 0; + } + } + } + ; + +k_unless : keyword_unless + { + token_info_push(p, "unless", &@$); + } + ; + +k_while : keyword_while + { + token_info_push(p, "while", &@$); + } + ; + +k_until : keyword_until { - $$ = NEW_DEFS($2, $5, $7, $8); - fixpos($$, $2); - local_pop(); - in_single--; + token_info_push(p, "until", &@$); } - | kBREAK + ; + +k_case : keyword_case { - $$ = NEW_BREAK(0); + token_info_push(p, "case", &@$); } - | kNEXT + ; + +k_for : keyword_for { - $$ = NEW_NEXT(0); + token_info_push(p, "for", &@$); } - | kREDO + ; + +k_class : keyword_class { - $$ = NEW_REDO(); + token_info_push(p, "class", &@$); + $<ctxt>$ = p->ctxt; } - | kRETRY + ; + +k_module : keyword_module { - $$ = NEW_RETRY(); + token_info_push(p, "module", &@$); + $<ctxt>$ = p->ctxt; } ; -primary_value : primary +k_def : keyword_def { - value_expr($1); - $$ = $1; + token_info_push(p, "def", &@$); + p->ctxt.in_argdef = 1; + } + ; + +k_do : keyword_do + { + token_info_push(p, "do", &@$); + } + ; + +k_do_block : keyword_do_block + { + token_info_push(p, "do", &@$); + } + ; + +k_rescue : keyword_rescue + { + token_info_warn(p, "rescue", p->token_info, 1, &@$); + } + ; + +k_ensure : keyword_ensure + { + token_info_warn(p, "ensure", p->token_info, 1, &@$); + } + ; + +k_when : keyword_when + { + token_info_warn(p, "when", p->token_info, 0, &@$); + } + ; + +k_else : keyword_else + { + token_info *ptinfo_beg = p->token_info; + int same = ptinfo_beg && strcmp(ptinfo_beg->token, "case") != 0; + token_info_warn(p, "else", p->token_info, same, &@$); + if (same) { + token_info e; + e.next = ptinfo_beg->next; + e.token = "else"; + token_info_setup(&e, p->lex.pbeg, &@$); + if (!e.nonspc) *ptinfo_beg = e; + } + } + ; + +k_elsif : keyword_elsif + { + WARN_EOL("elsif"); + token_info_warn(p, "elsif", p->token_info, 1, &@$); + } + ; + +k_end : keyword_end + { + token_info_pop(p, "end", &@$); + } + ; + +k_return : keyword_return + { + if (p->ctxt.in_class && !p->ctxt.in_def && !dyna_in_block(p)) + yyerror1(&@1, "Invalid return in class/module body"); } ; then : term - | ':' - | kTHEN - | term kTHEN + | keyword_then + | term keyword_then ; do : term - | ':' - | kDO_COND + | keyword_do_cond ; if_tail : opt_else - | kELSIF expr_value then + | k_elsif expr_value then compstmt if_tail { - $$ = NEW_IF(cond($2), $4, $5); - fixpos($$, $2); + /*%%%*/ + $$ = new_if(p, $2, $4, $5, &@$); + fixpos($$, $2); + /*% %*/ + /*% ripper: elsif!($2, $4, escape_Qundef($5)) %*/ } ; opt_else : none - | kELSE compstmt + | k_else compstmt { + /*%%%*/ $$ = $2; + /*% %*/ + /*% ripper: else!($2) %*/ } ; -block_var : lhs +for_var : lhs | mlhs ; -opt_block_var : none - | '|' /* none */ '|' +f_marg : f_norm_arg + { + /*%%%*/ + $$ = assignable(p, $1, 0, &@$); + mark_lvar_used(p, $$); + /*% %*/ + /*% ripper: assignable(p, $1) %*/ + } + | tLPAREN f_margs rparen + { + /*%%%*/ + $$ = $2; + /*% %*/ + /*% ripper: mlhs_paren!($2) %*/ + } + ; + +f_marg_list : f_marg + { + /*%%%*/ + $$ = NEW_LIST($1, &@$); + /*% %*/ + /*% ripper: mlhs_add!(mlhs_new!, $1) %*/ + } + | f_marg_list ',' f_marg + { + /*%%%*/ + $$ = list_append(p, $1, $3); + /*% %*/ + /*% ripper: mlhs_add!($1, $3) %*/ + } + ; + +f_margs : f_marg_list + { + /*%%%*/ + $$ = NEW_MASGN($1, 0, &@$); + /*% %*/ + /*% ripper: $1 %*/ + } + | f_marg_list ',' f_rest_marg + { + /*%%%*/ + $$ = NEW_MASGN($1, $3, &@$); + /*% %*/ + /*% ripper: mlhs_add_star!($1, $3) %*/ + } + | f_marg_list ',' f_rest_marg ',' f_marg_list + { + /*%%%*/ + $$ = NEW_MASGN($1, NEW_POSTARG($3, $5, &@$), &@$); + /*% %*/ + /*% ripper: mlhs_add_post!(mlhs_add_star!($1, $3), $5) %*/ + } + | f_rest_marg + { + /*%%%*/ + $$ = NEW_MASGN(0, $1, &@$); + /*% %*/ + /*% ripper: mlhs_add_star!(mlhs_new!, $1) %*/ + } + | f_rest_marg ',' f_marg_list + { + /*%%%*/ + $$ = NEW_MASGN(0, NEW_POSTARG($1, $3, &@$), &@$); + /*% %*/ + /*% ripper: mlhs_add_post!(mlhs_add_star!(mlhs_new!, $1), $3) %*/ + } + ; + +f_rest_marg : tSTAR f_norm_arg + { + /*%%%*/ + $$ = assignable(p, $2, 0, &@$); + mark_lvar_used(p, $$); + /*% %*/ + /*% ripper: assignable(p, $2) %*/ + } + | tSTAR + { + /*%%%*/ + $$ = NODE_SPECIAL_NO_NAME_REST; + /*% %*/ + /*% ripper: Qnil %*/ + } + ; + +f_any_kwrest : f_kwrest + | f_no_kwarg {$$ = ID2VAL(idNil);} + ; + +f_eq : {p->ctxt.in_argdef = 0;} '='; + +block_args_tail : f_block_kwarg ',' f_kwrest opt_f_block_arg + { + $$ = new_args_tail(p, $1, $3, $4, &@3); + } + | f_block_kwarg opt_f_block_arg + { + $$ = new_args_tail(p, $1, Qnone, $2, &@1); + } + | f_any_kwrest opt_f_block_arg { - $$ = (NODE*)1; + $$ = new_args_tail(p, Qnone, $1, $2, &@1); } - | tOROP + | f_block_arg { - $$ = (NODE*)1; + $$ = new_args_tail(p, Qnone, Qnone, $1, &@1); } - | '|' block_var '|' + ; + +opt_block_args_tail : ',' block_args_tail { $$ = $2; } + | /* none */ + { + $$ = new_args_tail(p, Qnone, Qnone, Qnone, &@0); + } ; -do_block : kDO_BLOCK +excessed_comma : ',' { - $<vars>$ = dyna_push(); - $<num>1 = ruby_sourceline; + /* magic number for rest_id in iseq_set_arguments() */ + /*%%%*/ + $$ = NODE_SPECIAL_EXCESSIVE_COMMA; + /*% %*/ + /*% ripper: excessed_comma! %*/ } - opt_block_var {$<vars>$ = ruby_dyna_vars;} - compstmt - kEND + ; + +block_param : f_arg ',' f_block_optarg ',' f_rest_arg opt_block_args_tail + { + $$ = new_args(p, $1, $3, $5, Qnone, $6, &@$); + } + | f_arg ',' f_block_optarg ',' f_rest_arg ',' f_arg opt_block_args_tail + { + $$ = new_args(p, $1, $3, $5, $7, $8, &@$); + } + | f_arg ',' f_block_optarg opt_block_args_tail + { + $$ = new_args(p, $1, $3, Qnone, Qnone, $4, &@$); + } + | f_arg ',' f_block_optarg ',' f_arg opt_block_args_tail + { + $$ = new_args(p, $1, $3, Qnone, $5, $6, &@$); + } + | f_arg ',' f_rest_arg opt_block_args_tail + { + $$ = new_args(p, $1, Qnone, $3, Qnone, $4, &@$); + } + | f_arg excessed_comma + { + $$ = new_args_tail(p, Qnone, Qnone, Qnone, &@2); + $$ = new_args(p, $1, Qnone, $2, Qnone, $$, &@$); + } + | f_arg ',' f_rest_arg ',' f_arg opt_block_args_tail { - $$ = NEW_ITER($3, 0, dyna_init($5, $<vars>4)); - nd_set_line($$, $<num>1); - dyna_pop($<vars>2); + $$ = new_args(p, $1, Qnone, $3, $5, $6, &@$); + } + | f_arg opt_block_args_tail + { + $$ = new_args(p, $1, Qnone, Qnone, Qnone, $2, &@$); + } + | f_block_optarg ',' f_rest_arg opt_block_args_tail + { + $$ = new_args(p, Qnone, $1, $3, Qnone, $4, &@$); + } + | f_block_optarg ',' f_rest_arg ',' f_arg opt_block_args_tail + { + $$ = new_args(p, Qnone, $1, $3, $5, $6, &@$); + } + | f_block_optarg opt_block_args_tail + { + $$ = new_args(p, Qnone, $1, Qnone, Qnone, $2, &@$); + } + | f_block_optarg ',' f_arg opt_block_args_tail + { + $$ = new_args(p, Qnone, $1, Qnone, $3, $4, &@$); + } + | f_rest_arg opt_block_args_tail + { + $$ = new_args(p, Qnone, Qnone, $1, Qnone, $2, &@$); + } + | f_rest_arg ',' f_arg opt_block_args_tail + { + $$ = new_args(p, Qnone, Qnone, $1, $3, $4, &@$); + } + | block_args_tail + { + $$ = new_args(p, Qnone, Qnone, Qnone, Qnone, $1, &@$); + } + ; + +opt_block_param : none + | block_param_def + { + p->command_start = TRUE; + } + ; + +block_param_def : '|' opt_bv_decl '|' + { + p->cur_arg = 0; + p->max_numparam = ORDINAL_PARAM; + p->ctxt.in_argdef = 0; + /*%%%*/ + $$ = 0; + /*% %*/ + /*% ripper: block_var!(params!(Qnil,Qnil,Qnil,Qnil,Qnil,Qnil,Qnil), escape_Qundef($2)) %*/ + } + | '|' block_param opt_bv_decl '|' + { + p->cur_arg = 0; + p->max_numparam = ORDINAL_PARAM; + p->ctxt.in_argdef = 0; + /*%%%*/ + $$ = $2; + /*% %*/ + /*% ripper: block_var!(escape_Qundef($2), escape_Qundef($3)) %*/ + } + ; + + +opt_bv_decl : opt_nl + { + $$ = 0; + } + | opt_nl ';' bv_decls opt_nl + { + /*%%%*/ + $$ = 0; + /*% %*/ + /*% ripper: $3 %*/ + } + ; + +bv_decls : bvar + /*% ripper[brace]: rb_ary_new3(1, get_value($1)) %*/ + | bv_decls ',' bvar + /*% ripper[brace]: rb_ary_push($1, get_value($3)) %*/ + ; + +bvar : tIDENTIFIER + { + new_bv(p, get_id($1)); + /*% ripper: get_value($1) %*/ + } + | f_bad_arg + { + $$ = 0; + } + ; + +lambda : tLAMBDA + { + token_info_push(p, "->", &@1); + $<vars>1 = dyna_push(p); + $<num>$ = p->lex.lpar_beg; + p->lex.lpar_beg = p->lex.paren_nest; + } + { + $<num>$ = p->max_numparam; + p->max_numparam = 0; + } + { + $<node>$ = numparam_push(p); + } + f_larglist + { + CMDARG_PUSH(0); + } + lambda_body + { + int max_numparam = p->max_numparam; + p->lex.lpar_beg = $<num>2; + p->max_numparam = $<num>3; + CMDARG_POP(); + $5 = args_with_numbered(p, $5, max_numparam); + /*%%%*/ + { + YYLTYPE loc = code_loc_gen(&@5, &@7); + $$ = NEW_LAMBDA($5, $7, &loc); + nd_set_line($$->nd_body, @7.end_pos.lineno); + nd_set_line($$, @5.end_pos.lineno); + nd_set_first_loc($$, @1.beg_pos); + } + /*% %*/ + /*% ripper: lambda!($5, $7) %*/ + numparam_pop(p, $<node>4); + dyna_pop(p, $<vars>1); + } + ; + +f_larglist : '(' f_args opt_bv_decl ')' + { + p->ctxt.in_argdef = 0; + /*%%%*/ + $$ = $2; + p->max_numparam = ORDINAL_PARAM; + /*% %*/ + /*% ripper: paren!($2) %*/ + } + | f_args + { + p->ctxt.in_argdef = 0; + /*%%%*/ + if (!args_info_empty_p($1->nd_ainfo)) + p->max_numparam = ORDINAL_PARAM; + /*% %*/ + $$ = $1; + } + ; + +lambda_body : tLAMBEG compstmt '}' + { + token_info_pop(p, "}", &@3); + $$ = $2; + } + | keyword_do_LAMBDA bodystmt k_end + { + $$ = $2; + } + ; + +do_block : k_do_block do_body k_end + { + $$ = $2; + /*%%%*/ + $$->nd_body->nd_loc = code_loc_gen(&@1, &@3); + nd_set_line($$, @1.end_pos.lineno); + /*% %*/ } ; block_call : command do_block { - if ($1 && nd_type($1) == NODE_BLOCK_PASS) { - rb_compile_error("both block arg and actual block given"); + /*%%%*/ + if (nd_type_p($1, NODE_YIELD)) { + compile_error(p, "block given to yield"); } - $2->nd_iter = $1; - $$ = $2; - fixpos($$, $1); + else { + block_dup_check(p, $1->nd_args, $2); + } + $$ = method_add_block(p, $1, $2, &@$); + fixpos($$, $1); + /*% %*/ + /*% ripper: method_add_block!($1, $2) %*/ + } + | block_call call_op2 operation2 opt_paren_args + { + /*%%%*/ + $$ = new_qcall(p, $2, $1, $3, $4, &@3, &@$); + /*% %*/ + /*% ripper: opt_event(:method_add_arg!, call!($1, $2, $3), $4) %*/ } - | block_call '.' operation2 opt_paren_args + | block_call call_op2 operation2 opt_paren_args brace_block { - $$ = new_call($1, $3, $4); + /*%%%*/ + $$ = new_command_qcall(p, $2, $1, $3, $4, $5, &@3, &@$); + /*% %*/ + /*% ripper: opt_event(:method_add_block!, command_call!($1, $2, $3, $4), $5) %*/ } - | block_call tCOLON2 operation2 opt_paren_args + | block_call call_op2 operation2 command_args do_block { - $$ = new_call($1, $3, $4); + /*%%%*/ + $$ = new_command_qcall(p, $2, $1, $3, $4, $5, &@3, &@$); + /*% %*/ + /*% ripper: method_add_block!(command_call!($1, $2, $3, $4), $5) %*/ } ; -method_call : operation paren_args +method_call : fcall paren_args { - $$ = new_fcall($1, $2); - fixpos($$, $2); + /*%%%*/ + $$ = $1; + $$->nd_args = $2; + nd_set_last_loc($1, @2.end_pos); + /*% %*/ + /*% ripper: method_add_arg!(fcall!($1), $2) %*/ } - | primary_value '.' operation2 opt_paren_args + | primary_value call_op operation2 opt_paren_args { - $$ = new_call($1, $3, $4); - fixpos($$, $1); + /*%%%*/ + $$ = new_qcall(p, $2, $1, $3, $4, &@3, &@$); + nd_set_line($$, @3.end_pos.lineno); + /*% %*/ + /*% ripper: opt_event(:method_add_arg!, call!($1, $2, $3), $4) %*/ } | primary_value tCOLON2 operation2 paren_args { - $$ = new_call($1, $3, $4); - fixpos($$, $1); + /*%%%*/ + $$ = new_qcall(p, ID2VAL(idCOLON2), $1, $3, $4, &@3, &@$); + nd_set_line($$, @3.end_pos.lineno); + /*% %*/ + /*% ripper: method_add_arg!(call!($1, $2, $3), $4) %*/ } | primary_value tCOLON2 operation3 { - $$ = new_call($1, $3, 0); + /*%%%*/ + $$ = new_qcall(p, ID2VAL(idCOLON2), $1, $3, Qnull, &@3, &@$); + /*% %*/ + /*% ripper: call!($1, $2, $3) %*/ + } + | primary_value call_op paren_args + { + /*%%%*/ + $$ = new_qcall(p, $2, $1, ID2VAL(idCall), $3, &@2, &@$); + nd_set_line($$, @2.end_pos.lineno); + /*% %*/ + /*% ripper: method_add_arg!(call!($1, $2, ID2VAL(idCall)), $3) %*/ } - | kSUPER paren_args + | primary_value tCOLON2 paren_args { - $$ = new_super($2); + /*%%%*/ + $$ = new_qcall(p, ID2VAL(idCOLON2), $1, ID2VAL(idCall), $3, &@2, &@$); + nd_set_line($$, @2.end_pos.lineno); + /*% %*/ + /*% ripper: method_add_arg!(call!($1, $2, ID2VAL(idCall)), $3) %*/ } - | kSUPER + | keyword_super paren_args { - $$ = NEW_ZSUPER(); + /*%%%*/ + $$ = NEW_SUPER($2, &@$); + /*% %*/ + /*% ripper: super!($2) %*/ + } + | keyword_super + { + /*%%%*/ + $$ = NEW_ZSUPER(&@$); + /*% %*/ + /*% ripper: zsuper! %*/ + } + | primary_value '[' opt_call_args rbracket + { + /*%%%*/ + if ($1 && nd_type_p($1, NODE_SELF)) + $$ = NEW_FCALL(tAREF, $3, &@$); + else + $$ = NEW_CALL($1, tAREF, $3, &@$); + fixpos($$, $1); + /*% %*/ + /*% ripper: aref!($1, escape_Qundef($3)) %*/ } ; -brace_block : '{' +brace_block : '{' brace_body '}' + { + $$ = $2; + /*%%%*/ + $$->nd_body->nd_loc = code_loc_gen(&@1, &@3); + nd_set_line($$, @1.end_pos.lineno); + /*% %*/ + } + | k_do do_body k_end { - $<vars>$ = dyna_push(); - $<num>1 = ruby_sourceline; + $$ = $2; + /*%%%*/ + $$->nd_body->nd_loc = code_loc_gen(&@1, &@3); + nd_set_line($$, @1.end_pos.lineno); + /*% %*/ } - opt_block_var {$<vars>$ = ruby_dyna_vars;} - compstmt '}' + ; + +brace_body : {$<vars>$ = dyna_push(p);} { - $$ = NEW_ITER($3, 0, dyna_init($5, $<vars>4)); - nd_set_line($$, $<num>1); - dyna_pop($<vars>2); + $<num>$ = p->max_numparam; + p->max_numparam = 0; } - | kDO { - $<vars>$ = dyna_push(); - $<num>1 = ruby_sourceline; + $<node>$ = numparam_push(p); } - opt_block_var {$<vars>$ = ruby_dyna_vars;} - compstmt kEND + opt_block_param compstmt { - $$ = NEW_ITER($3, 0, dyna_init($5, $<vars>4)); - nd_set_line($$, $<num>1); - dyna_pop($<vars>2); + int max_numparam = p->max_numparam; + p->max_numparam = $<num>2; + $4 = args_with_numbered(p, $4, max_numparam); + /*%%%*/ + $$ = NEW_ITER($4, $5, &@$); + /*% %*/ + /*% ripper: brace_block!(escape_Qundef($4), $5) %*/ + numparam_pop(p, $<node>3); + dyna_pop(p, $<vars>1); } ; -case_body : kWHEN when_args then - compstmt - cases +do_body : {$<vars>$ = dyna_push(p);} + { + $<num>$ = p->max_numparam; + p->max_numparam = 0; + } { - $$ = NEW_WHEN($2, $4, $5); + $<node>$ = numparam_push(p); + CMDARG_PUSH(0); + } + opt_block_param bodystmt + { + int max_numparam = p->max_numparam; + p->max_numparam = $<num>2; + $4 = args_with_numbered(p, $4, max_numparam); + /*%%%*/ + $$ = NEW_ITER($4, $5, &@$); + /*% %*/ + /*% ripper: do_block!(escape_Qundef($4), $5) %*/ + CMDARG_POP(); + numparam_pop(p, $<node>3); + dyna_pop(p, $<vars>1); } ; -when_args : args - | args ',' tSTAR arg_value + +case_args : arg_value { - $$ = list_append($1, NEW_WHEN($4, 0, 0)); + /*%%%*/ + check_literal_when(p, $1, &@1); + $$ = NEW_LIST($1, &@$); + /*% %*/ + /*% ripper: args_add!(args_new!, $1) %*/ } | tSTAR arg_value { - $$ = NEW_LIST(NEW_WHEN($2, 0, 0)); + /*%%%*/ + $$ = NEW_SPLAT($2, &@$); + /*% %*/ + /*% ripper: args_add_star!(args_new!, $2) %*/ + } + | case_args ',' arg_value + { + /*%%%*/ + check_literal_when(p, $3, &@3); + $$ = last_arg_append(p, $1, $3, &@$); + /*% %*/ + /*% ripper: args_add!($1, $3) %*/ + } + | case_args ',' tSTAR arg_value + { + /*%%%*/ + $$ = rest_arg_append(p, $1, $4, &@$); + /*% %*/ + /*% ripper: args_add_star!($1, $4) %*/ + } + ; + +case_body : k_when case_args then + compstmt + cases + { + /*%%%*/ + $$ = NEW_WHEN($2, $4, $5, &@$); + fixpos($$, $2); + /*% %*/ + /*% ripper: when!($2, $4, escape_Qundef($5)) %*/ } ; @@ -1890,25 +4074,578 @@ cases : opt_else | case_body ; -opt_rescue : kRESCUE exc_list exc_var then +p_case_body : keyword_in + { + SET_LEX_STATE(EXPR_BEG|EXPR_LABEL); + p->command_start = FALSE; + $<ctxt>1 = p->ctxt; + p->ctxt.in_kwarg = 1; + $<tbl>$ = push_pvtbl(p); + } + { + $<tbl>$ = push_pktbl(p); + } + p_top_expr then + { + pop_pktbl(p, $<tbl>3); + pop_pvtbl(p, $<tbl>2); + p->ctxt.in_kwarg = $<ctxt>1.in_kwarg; + } compstmt - opt_rescue + p_cases + { + /*%%%*/ + $$ = NEW_IN($4, $7, $8, &@$); + /*% %*/ + /*% ripper: in!($4, $7, escape_Qundef($8)) %*/ + } + ; + +p_cases : opt_else + | p_case_body + ; + +p_top_expr : p_top_expr_body + | p_top_expr_body modifier_if expr_value + { + /*%%%*/ + $$ = new_if(p, $3, $1, 0, &@$); + fixpos($$, $3); + /*% %*/ + /*% ripper: if_mod!($3, $1) %*/ + } + | p_top_expr_body modifier_unless expr_value + { + /*%%%*/ + $$ = new_unless(p, $3, $1, 0, &@$); + fixpos($$, $3); + /*% %*/ + /*% ripper: unless_mod!($3, $1) %*/ + } + ; + +p_top_expr_body : p_expr + | p_expr ',' + { + $$ = new_array_pattern_tail(p, Qnone, 1, 0, Qnone, &@$); + $$ = new_array_pattern(p, Qnone, get_value($1), $$, &@$); + } + | p_expr ',' p_args + { + $$ = new_array_pattern(p, Qnone, get_value($1), $3, &@$); + /*%%%*/ + nd_set_first_loc($$, @1.beg_pos); + /*% + %*/ + } + | p_find + { + $$ = new_find_pattern(p, Qnone, $1, &@$); + } + | p_args_tail + { + $$ = new_array_pattern(p, Qnone, Qnone, $1, &@$); + } + | p_kwargs + { + $$ = new_hash_pattern(p, Qnone, $1, &@$); + } + ; + +p_expr : p_as + ; + +p_as : p_expr tASSOC p_variable + { + /*%%%*/ + NODE *n = NEW_LIST($1, &@$); + n = list_append(p, n, $3); + $$ = new_hash(p, n, &@$); + /*% %*/ + /*% ripper: binary!($1, STATIC_ID2SYM((id_assoc)), $3) %*/ + } + | p_alt + ; + +p_alt : p_alt '|' p_expr_basic + { + /*%%%*/ + $$ = NEW_NODE(NODE_OR, $1, $3, 0, &@$); + /*% %*/ + /*% ripper: binary!($1, STATIC_ID2SYM(idOr), $3) %*/ + } + | p_expr_basic + ; + +p_lparen : '(' {$<tbl>$ = push_pktbl(p);}; +p_lbracket : '[' {$<tbl>$ = push_pktbl(p);}; + +p_expr_basic : p_value + | p_variable + | p_const p_lparen p_args rparen + { + pop_pktbl(p, $<tbl>2); + $$ = new_array_pattern(p, $1, Qnone, $3, &@$); + /*%%%*/ + nd_set_first_loc($$, @1.beg_pos); + /*% + %*/ + } + | p_const p_lparen p_find rparen + { + pop_pktbl(p, $<tbl>2); + $$ = new_find_pattern(p, $1, $3, &@$); + /*%%%*/ + nd_set_first_loc($$, @1.beg_pos); + /*% + %*/ + } + | p_const p_lparen p_kwargs rparen + { + pop_pktbl(p, $<tbl>2); + $$ = new_hash_pattern(p, $1, $3, &@$); + /*%%%*/ + nd_set_first_loc($$, @1.beg_pos); + /*% + %*/ + } + | p_const '(' rparen + { + $$ = new_array_pattern_tail(p, Qnone, 0, 0, Qnone, &@$); + $$ = new_array_pattern(p, $1, Qnone, $$, &@$); + } + | p_const p_lbracket p_args rbracket + { + pop_pktbl(p, $<tbl>2); + $$ = new_array_pattern(p, $1, Qnone, $3, &@$); + /*%%%*/ + nd_set_first_loc($$, @1.beg_pos); + /*% + %*/ + } + | p_const p_lbracket p_find rbracket + { + pop_pktbl(p, $<tbl>2); + $$ = new_find_pattern(p, $1, $3, &@$); + /*%%%*/ + nd_set_first_loc($$, @1.beg_pos); + /*% + %*/ + } + | p_const p_lbracket p_kwargs rbracket + { + pop_pktbl(p, $<tbl>2); + $$ = new_hash_pattern(p, $1, $3, &@$); + /*%%%*/ + nd_set_first_loc($$, @1.beg_pos); + /*% + %*/ + } + | p_const '[' rbracket + { + $$ = new_array_pattern_tail(p, Qnone, 0, 0, Qnone, &@$); + $$ = new_array_pattern(p, $1, Qnone, $$, &@$); + } + | tLBRACK p_args rbracket + { + $$ = new_array_pattern(p, Qnone, Qnone, $2, &@$); + } + | tLBRACK p_find rbracket + { + $$ = new_find_pattern(p, Qnone, $2, &@$); + } + | tLBRACK rbracket + { + $$ = new_array_pattern_tail(p, Qnone, 0, 0, Qnone, &@$); + $$ = new_array_pattern(p, Qnone, Qnone, $$, &@$); + } + | tLBRACE + { + $<tbl>$ = push_pktbl(p); + $<ctxt>1 = p->ctxt; + p->ctxt.in_kwarg = 0; + } + p_kwargs rbrace + { + pop_pktbl(p, $<tbl>2); + p->ctxt.in_kwarg = $<ctxt>1.in_kwarg; + $$ = new_hash_pattern(p, Qnone, $3, &@$); + } + | tLBRACE rbrace + { + $$ = new_hash_pattern_tail(p, Qnone, 0, &@$); + $$ = new_hash_pattern(p, Qnone, $$, &@$); + } + | tLPAREN {$<tbl>$ = push_pktbl(p);} p_expr rparen + { + pop_pktbl(p, $<tbl>2); + $$ = $3; + } + ; + +p_args : p_expr + { + /*%%%*/ + NODE *pre_args = NEW_LIST($1, &@$); + $$ = new_array_pattern_tail(p, pre_args, 0, 0, Qnone, &@$); + /*% + $$ = new_array_pattern_tail(p, rb_ary_new_from_args(1, get_value($1)), 0, 0, Qnone, &@$); + %*/ + } + | p_args_head + { + $$ = new_array_pattern_tail(p, $1, 1, 0, Qnone, &@$); + } + | p_args_head p_arg + { + /*%%%*/ + $$ = new_array_pattern_tail(p, list_concat($1, $2), 0, 0, Qnone, &@$); + /*% + VALUE pre_args = rb_ary_concat($1, get_value($2)); + $$ = new_array_pattern_tail(p, pre_args, 0, 0, Qnone, &@$); + %*/ + } + | p_args_head tSTAR tIDENTIFIER + { + $$ = new_array_pattern_tail(p, $1, 1, $3, Qnone, &@$); + } + | p_args_head tSTAR tIDENTIFIER ',' p_args_post + { + $$ = new_array_pattern_tail(p, $1, 1, $3, $5, &@$); + } + | p_args_head tSTAR + { + $$ = new_array_pattern_tail(p, $1, 1, 0, Qnone, &@$); + } + | p_args_head tSTAR ',' p_args_post + { + $$ = new_array_pattern_tail(p, $1, 1, 0, $4, &@$); + } + | p_args_tail + ; + +p_args_head : p_arg ',' + { + $$ = $1; + } + | p_args_head p_arg ',' + { + /*%%%*/ + $$ = list_concat($1, $2); + /*% %*/ + /*% ripper: rb_ary_concat($1, get_value($2)) %*/ + } + ; + +p_args_tail : p_rest + { + $$ = new_array_pattern_tail(p, Qnone, 1, $1, Qnone, &@$); + } + | p_rest ',' p_args_post + { + $$ = new_array_pattern_tail(p, Qnone, 1, $1, $3, &@$); + } + ; + +p_find : p_rest ',' p_args_post ',' p_rest + { + $$ = new_find_pattern_tail(p, $1, $3, $5, &@$); + + if (rb_warning_category_enabled_p(RB_WARN_CATEGORY_EXPERIMENTAL)) + rb_warn0L_experimental(nd_line($$), "Find pattern is experimental, and the behavior may change in future versions of Ruby!"); + } + ; + + +p_rest : tSTAR tIDENTIFIER + { + $$ = $2; + } + | tSTAR + { + $$ = 0; + } + ; + +p_args_post : p_arg + | p_args_post ',' p_arg + { + /*%%%*/ + $$ = list_concat($1, $3); + /*% %*/ + /*% ripper: rb_ary_concat($1, get_value($3)) %*/ + } + ; + +p_arg : p_expr + { + /*%%%*/ + $$ = NEW_LIST($1, &@$); + /*% %*/ + /*% ripper: rb_ary_new_from_args(1, get_value($1)) %*/ + } + ; + +p_kwargs : p_kwarg ',' p_any_kwrest + { + $$ = new_hash_pattern_tail(p, new_unique_key_hash(p, $1, &@$), $3, &@$); + } + | p_kwarg + { + $$ = new_hash_pattern_tail(p, new_unique_key_hash(p, $1, &@$), 0, &@$); + } + | p_kwarg ',' + { + $$ = new_hash_pattern_tail(p, new_unique_key_hash(p, $1, &@$), 0, &@$); + } + | p_any_kwrest + { + $$ = new_hash_pattern_tail(p, new_hash(p, Qnone, &@$), $1, &@$); + } + ; + +p_kwarg : p_kw + /*% ripper[brace]: rb_ary_new_from_args(1, $1) %*/ + | p_kwarg ',' p_kw + { + /*%%%*/ + $$ = list_concat($1, $3); + /*% %*/ + /*% ripper: rb_ary_push($1, $3) %*/ + } + ; + +p_kw : p_kw_label p_expr + { + error_duplicate_pattern_key(p, get_id($1), &@1); + /*%%%*/ + $$ = list_append(p, NEW_LIST(NEW_LIT(ID2SYM($1), &@$), &@$), $2); + /*% %*/ + /*% ripper: rb_ary_new_from_args(2, get_value($1), get_value($2)) %*/ + } + | p_kw_label { - if ($3) { - $3 = node_assign($3, NEW_GVAR(rb_intern("$!"))); - $5 = block_append($3, $5); + error_duplicate_pattern_key(p, get_id($1), &@1); + if ($1 && !is_local_id(get_id($1))) { + yyerror1(&@1, "key must be valid as local variables"); + } + error_duplicate_pattern_variable(p, get_id($1), &@1); + /*%%%*/ + $$ = list_append(p, NEW_LIST(NEW_LIT(ID2SYM($1), &@$), &@$), assignable(p, $1, 0, &@$)); + /*% %*/ + /*% ripper: rb_ary_new_from_args(2, get_value($1), Qnil) %*/ + } + ; + +p_kw_label : tLABEL + | tSTRING_BEG string_contents tLABEL_END + { + YYLTYPE loc = code_loc_gen(&@1, &@3); + /*%%%*/ + if (!$2 || nd_type_p($2, NODE_STR)) { + NODE *node = dsym_node(p, $2, &loc); + $$ = SYM2ID(node->nd_lit); + } + /*% + if (ripper_is_node_yylval($2) && RNODE($2)->nd_cval) { + VALUE label = RNODE($2)->nd_cval; + VALUE rval = RNODE($2)->nd_rval; + $$ = ripper_new_yylval(p, rb_intern_str(label), rval, label); + RNODE($$)->nd_loc = loc; + } + %*/ + else { + yyerror1(&loc, "symbol literal with interpolation is not allowed"); + $$ = 0; } - $$ = NEW_RESBODY($2, $5, $6); - fixpos($$, $2?$2:$5); + } + ; + +p_kwrest : kwrest_mark tIDENTIFIER + { + $$ = $2; + } + | kwrest_mark + { + $$ = 0; + } + ; + +p_kwnorest : kwrest_mark keyword_nil + { + $$ = 0; + } + ; + +p_any_kwrest : p_kwrest + | p_kwnorest {$$ = ID2VAL(idNil);} + ; + +p_value : p_primitive + | p_primitive tDOT2 p_primitive + { + /*%%%*/ + value_expr($1); + value_expr($3); + $$ = NEW_DOT2($1, $3, &@$); + /*% %*/ + /*% ripper: dot2!($1, $3) %*/ + } + | p_primitive tDOT3 p_primitive + { + /*%%%*/ + value_expr($1); + value_expr($3); + $$ = NEW_DOT3($1, $3, &@$); + /*% %*/ + /*% ripper: dot3!($1, $3) %*/ + } + | p_primitive tDOT2 + { + /*%%%*/ + value_expr($1); + $$ = NEW_DOT2($1, new_nil_at(p, &@2.end_pos), &@$); + /*% %*/ + /*% ripper: dot2!($1, Qnil) %*/ + } + | p_primitive tDOT3 + { + /*%%%*/ + value_expr($1); + $$ = NEW_DOT3($1, new_nil_at(p, &@2.end_pos), &@$); + /*% %*/ + /*% ripper: dot3!($1, Qnil) %*/ + } + | p_var_ref + | p_expr_ref + | p_const + | tBDOT2 p_primitive + { + /*%%%*/ + value_expr($2); + $$ = NEW_DOT2(new_nil_at(p, &@1.beg_pos), $2, &@$); + /*% %*/ + /*% ripper: dot2!(Qnil, $2) %*/ + } + | tBDOT3 p_primitive + { + /*%%%*/ + value_expr($2); + $$ = NEW_DOT3(new_nil_at(p, &@1.beg_pos), $2, &@$); + /*% %*/ + /*% ripper: dot3!(Qnil, $2) %*/ + } + ; + +p_primitive : literal + | strings + | xstring + | regexp + | words + | qwords + | symbols + | qsymbols + | keyword_variable + { + /*%%%*/ + if (!($$ = gettable(p, $1, &@$))) $$ = NEW_BEGIN(0, &@$); + /*% %*/ + /*% ripper: var_ref!($1) %*/ + } + | lambda + ; + +p_variable : tIDENTIFIER + { + /*%%%*/ + error_duplicate_pattern_variable(p, $1, &@1); + $$ = assignable(p, $1, 0, &@$); + /*% %*/ + /*% ripper: assignable(p, var_field(p, $1)) %*/ + } + ; + +p_var_ref : '^' tIDENTIFIER + { + /*%%%*/ + NODE *n = gettable(p, $2, &@$); + if (!(nd_type_p(n, NODE_LVAR) || nd_type_p(n, NODE_DVAR))) { + compile_error(p, "%"PRIsVALUE": no such local variable", rb_id2str($2)); + } + $$ = n; + /*% %*/ + /*% ripper: var_ref!($2) %*/ + } + | '^' nonlocal_var + { + /*%%%*/ + if (!($$ = gettable(p, $2, &@$))) $$ = NEW_BEGIN(0, &@$); + /*% %*/ + /*% ripper: var_ref!($2) %*/ + } + ; + +p_expr_ref : '^' tLPAREN expr_value ')' + { + /*%%%*/ + $$ = NEW_BEGIN($3, &@$); + /*% %*/ + /*% ripper: begin!($3) %*/ + } + ; + +p_const : tCOLON3 cname + { + /*%%%*/ + $$ = NEW_COLON3($2, &@$); + /*% %*/ + /*% ripper: top_const_ref!($2) %*/ + } + | p_const tCOLON2 cname + { + /*%%%*/ + $$ = NEW_COLON2($1, $3, &@$); + /*% %*/ + /*% ripper: const_path_ref!($1, $3) %*/ + } + | tCONSTANT + { + /*%%%*/ + $$ = gettable(p, $1, &@$); + /*% %*/ + /*% ripper: var_ref!($1) %*/ + } + ; + +opt_rescue : k_rescue exc_list exc_var then + compstmt + opt_rescue + { + /*%%%*/ + $$ = NEW_RESBODY($2, + $3 ? block_append(p, node_assign(p, $3, NEW_ERRINFO(&@3), NO_LEX_CTXT, &@3), $5) : $5, + $6, &@$); + fixpos($$, $2?$2:$5); + /*% %*/ + /*% ripper: rescue!(escape_Qundef($2), escape_Qundef($3), escape_Qundef($5), escape_Qundef($6)) %*/ } | none ; exc_list : arg_value { - $$ = NEW_LIST($1); + /*%%%*/ + $$ = NEW_LIST($1, &@$); + /*% %*/ + /*% ripper: rb_ary_new3(1, get_value($1)) %*/ } | mrhs + { + /*%%%*/ + if (!($$ = splat_array($1))) $$ = $1; + /*% %*/ + /*% ripper: $1 %*/ + } | none ; @@ -1919,217 +4656,373 @@ exc_var : tASSOC lhs | none ; -opt_ensure : kENSURE compstmt +opt_ensure : k_ensure compstmt { - if ($2) - $$ = $2; - else - /* place holder */ - $$ = NEW_NIL(); + /*%%%*/ + $$ = $2; + /*% %*/ + /*% ripper: ensure!($2) %*/ } | none ; literal : numeric | symbol - { - $$ = NEW_LIT(ID2SYM($1)); - } - | dsym ; strings : string { + /*%%%*/ NODE *node = $1; if (!node) { - node = NEW_STR(rb_str_new(0, 0)); + node = NEW_STR(STR_NEW0(), &@$); + RB_OBJ_WRITTEN(p->ast, Qnil, node->nd_lit); } else { - node = evstr2dstr(node); + node = evstr2dstr(p, node); } $$ = node; + /*% %*/ + /*% ripper: $1 %*/ } ; -string : string1 +string : tCHAR + | string1 | string string1 { - $$ = literal_concat($1, $2); + /*%%%*/ + $$ = literal_concat(p, $1, $2, &@$); + /*% %*/ + /*% ripper: string_concat!($1, $2) %*/ } ; string1 : tSTRING_BEG string_contents tSTRING_END { - $$ = $2; + /*%%%*/ + $$ = heredoc_dedent(p, $2); + if ($$) nd_set_loc($$, &@$); + /*% %*/ + /*% ripper: string_literal!(heredoc_dedent(p, $2)) %*/ } ; xstring : tXSTRING_BEG xstring_contents tSTRING_END { - NODE *node = $2; - if (!node) { - node = NEW_XSTR(rb_str_new(0, 0)); - } - else { - switch (nd_type(node)) { - case NODE_STR: - nd_set_type(node, NODE_XSTR); - break; - case NODE_DSTR: - nd_set_type(node, NODE_DXSTR); - break; - default: - node = NEW_NODE(NODE_DXSTR, rb_str_new(0, 0), 1, NEW_LIST(node)); - break; - } - } - $$ = node; + /*%%%*/ + $$ = new_xstring(p, heredoc_dedent(p, $2), &@$); + /*% %*/ + /*% ripper: xstring_literal!(heredoc_dedent(p, $2)) %*/ } ; -regexp : tREGEXP_BEG xstring_contents tREGEXP_END +regexp : tREGEXP_BEG regexp_contents tREGEXP_END { - int options = $3; - NODE *node = $2; - if (!node) { - node = NEW_LIT(rb_reg_new("", 0, options & ~RE_OPTION_ONCE)); - } - else switch (nd_type(node)) { - case NODE_STR: - { - VALUE src = node->nd_lit; - nd_set_type(node, NODE_LIT); - node->nd_lit = rb_reg_new(RSTRING(src)->ptr, - RSTRING(src)->len, - options & ~RE_OPTION_ONCE); - } - break; - default: - node = NEW_NODE(NODE_DSTR, rb_str_new(0, 0), 1, NEW_LIST(node)); - case NODE_DSTR: - if (options & RE_OPTION_ONCE) { - nd_set_type(node, NODE_DREGX_ONCE); - } - else { - nd_set_type(node, NODE_DREGX); - } - node->nd_cflag = options & ~RE_OPTION_ONCE; - break; - } - $$ = node; + $$ = new_regexp(p, $2, $3, &@$); } ; -words : tWORDS_BEG ' ' tSTRING_END - { - $$ = NEW_ZARRAY(); - } - | tWORDS_BEG word_list tSTRING_END +words : tWORDS_BEG ' ' word_list tSTRING_END { - $$ = $2; + /*%%%*/ + $$ = make_list($3, &@$); + /*% %*/ + /*% ripper: array!($3) %*/ } ; word_list : /* none */ { + /*%%%*/ $$ = 0; + /*% %*/ + /*% ripper: words_new! %*/ } | word_list word ' ' { - $$ = list_append($1, evstr2dstr($2)); + /*%%%*/ + $$ = list_append(p, $1, evstr2dstr(p, $2)); + /*% %*/ + /*% ripper: words_add!($1, $2) %*/ } ; word : string_content + /*% ripper[brace]: word_add!(word_new!, $1) %*/ | word string_content { - $$ = literal_concat($1, $2); + /*%%%*/ + $$ = literal_concat(p, $1, $2, &@$); + /*% %*/ + /*% ripper: word_add!($1, $2) %*/ } ; -qwords : tQWORDS_BEG ' ' tSTRING_END +symbols : tSYMBOLS_BEG ' ' symbol_list tSTRING_END { - $$ = NEW_ZARRAY(); + /*%%%*/ + $$ = make_list($3, &@$); + /*% %*/ + /*% ripper: array!($3) %*/ } - | tQWORDS_BEG qword_list tSTRING_END + ; + +symbol_list : /* none */ { - $$ = $2; + /*%%%*/ + $$ = 0; + /*% %*/ + /*% ripper: symbols_new! %*/ + } + | symbol_list word ' ' + { + /*%%%*/ + $$ = symbol_append(p, $1, evstr2dstr(p, $2)); + /*% %*/ + /*% ripper: symbols_add!($1, $2) %*/ + } + ; + +qwords : tQWORDS_BEG ' ' qword_list tSTRING_END + { + /*%%%*/ + $$ = make_list($3, &@$); + /*% %*/ + /*% ripper: array!($3) %*/ + } + ; + +qsymbols : tQSYMBOLS_BEG ' ' qsym_list tSTRING_END + { + /*%%%*/ + $$ = make_list($3, &@$); + /*% %*/ + /*% ripper: array!($3) %*/ } ; qword_list : /* none */ { + /*%%%*/ $$ = 0; + /*% %*/ + /*% ripper: qwords_new! %*/ } | qword_list tSTRING_CONTENT ' ' { - $$ = list_append($1, $2); + /*%%%*/ + $$ = list_append(p, $1, $2); + /*% %*/ + /*% ripper: qwords_add!($1, $2) %*/ + } + ; + +qsym_list : /* none */ + { + /*%%%*/ + $$ = 0; + /*% %*/ + /*% ripper: qsymbols_new! %*/ + } + | qsym_list tSTRING_CONTENT ' ' + { + /*%%%*/ + $$ = symbol_append(p, $1, $2); + /*% %*/ + /*% ripper: qsymbols_add!($1, $2) %*/ } ; string_contents : /* none */ { + /*%%%*/ $$ = 0; + /*% %*/ + /*% ripper: string_content! %*/ + /*%%%*/ + /*% + $$ = ripper_new_yylval(p, 0, $$, 0); + %*/ } | string_contents string_content { - $$ = literal_concat($1, $2); + /*%%%*/ + $$ = literal_concat(p, $1, $2, &@$); + /*% %*/ + /*% ripper: string_add!($1, $2) %*/ + /*%%%*/ + /*% + if (ripper_is_node_yylval($1) && ripper_is_node_yylval($2) && + !RNODE($1)->nd_cval) { + RNODE($1)->nd_cval = RNODE($2)->nd_cval; + RNODE($1)->nd_rval = add_mark_object(p, $$); + $$ = $1; + } + %*/ } ; xstring_contents: /* none */ { + /*%%%*/ $$ = 0; + /*% %*/ + /*% ripper: xstring_new! %*/ } | xstring_contents string_content { - $$ = literal_concat($1, $2); + /*%%%*/ + $$ = literal_concat(p, $1, $2, &@$); + /*% %*/ + /*% ripper: xstring_add!($1, $2) %*/ + } + ; + +regexp_contents: /* none */ + { + /*%%%*/ + $$ = 0; + /*% %*/ + /*% ripper: regexp_new! %*/ + /*%%%*/ + /*% + $$ = ripper_new_yylval(p, 0, $$, 0); + %*/ + } + | regexp_contents string_content + { + /*%%%*/ + NODE *head = $1, *tail = $2; + if (!head) { + $$ = tail; + } + else if (!tail) { + $$ = head; + } + else { + switch (nd_type(head)) { + case NODE_STR: + nd_set_type(head, NODE_DSTR); + break; + case NODE_DSTR: + break; + default: + head = list_append(p, NEW_DSTR(Qnil, &@$), head); + break; + } + $$ = list_append(p, head, tail); + } + /*% + VALUE s1 = 1, s2 = 0, n1 = $1, n2 = $2; + if (ripper_is_node_yylval(n1)) { + s1 = RNODE(n1)->nd_cval; + n1 = RNODE(n1)->nd_rval; + } + if (ripper_is_node_yylval(n2)) { + s2 = RNODE(n2)->nd_cval; + n2 = RNODE(n2)->nd_rval; + } + $$ = dispatch2(regexp_add, n1, n2); + if (!s1 && s2) { + $$ = ripper_new_yylval(p, 0, $$, s2); + } + %*/ } ; string_content : tSTRING_CONTENT + /*% ripper[brace]: ripper_new_yylval(p, 0, get_value($1), $1) %*/ | tSTRING_DVAR { - $<node>$ = lex_strterm; - lex_strterm = 0; - lex_state = EXPR_BEG; + /* need to backup p->lex.strterm so that a string literal `%&foo,#$&,bar&` can be parsed */ + $<strterm>$ = p->lex.strterm; + p->lex.strterm = 0; + SET_LEX_STATE(EXPR_BEG); } string_dvar { - lex_strterm = $<node>2; - $$ = NEW_EVSTR($3); + p->lex.strterm = $<strterm>2; + /*%%%*/ + $$ = NEW_EVSTR($3, &@$); + nd_set_line($$, @3.end_pos.lineno); + /*% %*/ + /*% ripper: string_dvar!($3) %*/ } | tSTRING_DBEG { - $<node>$ = lex_strterm; - lex_strterm = 0; - lex_state = EXPR_BEG; - COND_PUSH(0); CMDARG_PUSH(0); + COND_PUSH(0); } - compstmt '}' { - lex_strterm = $<node>2; - COND_LEXPOP(); - CMDARG_LEXPOP(); - if (($$ = $3) && nd_type($$) == NODE_NEWLINE) { - $$ = $$->nd_next; - rb_gc_force_recycle((VALUE)$3); - } - $$ = new_evstr($$); + /* need to backup p->lex.strterm so that a string literal `%!foo,#{ !0 },bar!` can be parsed */ + $<strterm>$ = p->lex.strterm; + p->lex.strterm = 0; + } + { + $<num>$ = p->lex.state; + SET_LEX_STATE(EXPR_BEG); + } + { + $<num>$ = p->lex.brace_nest; + p->lex.brace_nest = 0; + } + { + $<num>$ = p->heredoc_indent; + p->heredoc_indent = 0; + } + compstmt tSTRING_DEND + { + COND_POP(); + CMDARG_POP(); + p->lex.strterm = $<strterm>3; + SET_LEX_STATE($<num>4); + p->lex.brace_nest = $<num>5; + p->heredoc_indent = $<num>6; + p->heredoc_line_indent = -1; + /*%%%*/ + if ($7) $7->flags &= ~NODE_FL_NEWLINE; + $$ = new_evstr(p, $7, &@$); + /*% %*/ + /*% ripper: string_embexpr!($7) %*/ } ; -string_dvar : tGVAR {$$ = NEW_GVAR($1);} - | tIVAR {$$ = NEW_IVAR($1);} - | tCVAR {$$ = NEW_CVAR($1);} +string_dvar : tGVAR + { + /*%%%*/ + $$ = NEW_GVAR($1, &@$); + /*% %*/ + /*% ripper: var_ref!($1) %*/ + } + | tIVAR + { + /*%%%*/ + $$ = NEW_IVAR($1, &@$); + /*% %*/ + /*% ripper: var_ref!($1) %*/ + } + | tCVAR + { + /*%%%*/ + $$ = NEW_CVAR($1, &@$); + /*% %*/ + /*% ripper: var_ref!($1) %*/ + } | backref ; -symbol : tSYMBEG sym +symbol : ssym + | dsym + ; + +ssym : tSYMBEG sym { - lex_state = EXPR_END; - $$ = $2; + SET_LEX_STATE(EXPR_END); + /*%%%*/ + $$ = NEW_LIT(ID2SYM($2), &@$); + /*% %*/ + /*% ripper: symbol_literal!(symbol!($2)) %*/ } ; @@ -2139,74 +5032,89 @@ sym : fname | tCVAR ; -dsym : tSYMBEG xstring_contents tSTRING_END +dsym : tSYMBEG string_contents tSTRING_END { - lex_state = EXPR_END; - if (!($$ = $2)) { - $$ = NEW_NIL(); - yyerror("empty symbol literal"); - } - else { - VALUE lit; - - switch (nd_type($$)) { - case NODE_DSTR: - nd_set_type($$, NODE_DSYM); - break; - case NODE_STR: - lit = $$->nd_lit; - if (RSTRING(lit)->len == 0) { - yyerror("empty symbol literal"); - break; - } - if (strlen(RSTRING(lit)->ptr) == RSTRING(lit)->len) { - $$->nd_lit = ID2SYM(rb_intern(RSTRING($$->nd_lit)->ptr)); - nd_set_type($$, NODE_LIT); - break; - } - /* fall through */ - default: - $$ = NEW_NODE(NODE_DSYM, rb_str_new(0, 0), 1, NEW_LIST($$)); - break; - } - } + SET_LEX_STATE(EXPR_END); + /*%%%*/ + $$ = dsym_node(p, $2, &@$); + /*% %*/ + /*% ripper: dyna_symbol!($2) %*/ } ; -numeric : tINTEGER - | tFLOAT - | tUMINUS_NUM tINTEGER %prec tLOWEST - { - $$ = negate_lit($2); - } - | tUMINUS_NUM tFLOAT %prec tLOWEST +numeric : simple_numeric + | tUMINUS_NUM simple_numeric %prec tLOWEST { - $$ = negate_lit($2); + /*%%%*/ + $$ = $2; + RB_OBJ_WRITE(p->ast, &$$->nd_lit, negate_lit(p, $$->nd_lit)); + /*% %*/ + /*% ripper: unary!(ID2VAL(idUMinus), $2) %*/ } ; -variable : tIDENTIFIER +simple_numeric : tINTEGER + | tFLOAT + | tRATIONAL + | tIMAGINARY + ; + +nonlocal_var : tIVAR + | tGVAR + | tCVAR + ; + +user_variable : tIDENTIFIER | tIVAR | tGVAR | tCONSTANT | tCVAR - | kNIL {$$ = kNIL;} - | kSELF {$$ = kSELF;} - | kTRUE {$$ = kTRUE;} - | kFALSE {$$ = kFALSE;} - | k__FILE__ {$$ = k__FILE__;} - | k__LINE__ {$$ = k__LINE__;} ; -var_ref : variable +keyword_variable: keyword_nil {$$ = KWD2EID(nil, $1);} + | keyword_self {$$ = KWD2EID(self, $1);} + | keyword_true {$$ = KWD2EID(true, $1);} + | keyword_false {$$ = KWD2EID(false, $1);} + | keyword__FILE__ {$$ = KWD2EID(_FILE__, $1);} + | keyword__LINE__ {$$ = KWD2EID(_LINE__, $1);} + | keyword__ENCODING__ {$$ = KWD2EID(_ENCODING__, $1);} + ; + +var_ref : user_variable + { + /*%%%*/ + if (!($$ = gettable(p, $1, &@$))) $$ = NEW_BEGIN(0, &@$); + /*% + if (id_is_var(p, get_id($1))) { + $$ = dispatch1(var_ref, $1); + } + else { + $$ = dispatch1(vcall, $1); + } + %*/ + } + | keyword_variable { - $$ = gettable($1); + /*%%%*/ + if (!($$ = gettable(p, $1, &@$))) $$ = NEW_BEGIN(0, &@$); + /*% %*/ + /*% ripper: var_ref!($1) %*/ } ; -var_lhs : variable +var_lhs : user_variable { - $$ = assignable($1, 0); + /*%%%*/ + $$ = assignable(p, $1, 0, &@$); + /*% %*/ + /*% ripper: assignable(p, var_field(p, $1)) %*/ + } + | keyword_variable + { + /*%%%*/ + $$ = assignable(p, $1, 0, &@$); + /*% %*/ + /*% ripper: assignable(p, var_field(p, $1)) %*/ } ; @@ -2214,123 +5122,429 @@ backref : tNTH_REF | tBACK_REF ; -superclass : term +superclass : '<' + { + SET_LEX_STATE(EXPR_BEG); + p->command_start = TRUE; + } + expr_value term + { + $$ = $3; + } + | /* none */ { + /*%%%*/ $$ = 0; + /*% %*/ + /*% ripper: Qnil %*/ } - | '<' + ; + +f_opt_paren_args: f_paren_args + | none { - lex_state = EXPR_BEG; + p->ctxt.in_argdef = 0; + $$ = new_args_tail(p, Qnone, Qnone, Qnone, &@0); + $$ = new_args(p, Qnone, Qnone, Qnone, Qnone, $$, &@0); } - expr_value term + ; + +f_paren_args : '(' f_args rparen { - $$ = $3; + /*%%%*/ + $$ = $2; + /*% %*/ + /*% ripper: paren!($2) %*/ + SET_LEX_STATE(EXPR_BEG); + p->command_start = TRUE; + p->ctxt.in_argdef = 0; + } + ; + +f_arglist : f_paren_args + | { + $<ctxt>$ = p->ctxt; + p->ctxt.in_kwarg = 1; + p->ctxt.in_argdef = 1; + SET_LEX_STATE(p->lex.state|EXPR_LABEL); /* force for args */ + } + f_args term + { + p->ctxt.in_kwarg = $<ctxt>1.in_kwarg; + p->ctxt.in_argdef = 0; + $$ = $2; + SET_LEX_STATE(EXPR_BEG); + p->command_start = TRUE; + } + ; + +args_tail : f_kwarg ',' f_kwrest opt_f_block_arg + { + $$ = new_args_tail(p, $1, $3, $4, &@3); + } + | f_kwarg opt_f_block_arg + { + $$ = new_args_tail(p, $1, Qnone, $2, &@1); + } + | f_any_kwrest opt_f_block_arg + { + $$ = new_args_tail(p, Qnone, $1, $2, &@1); + } + | f_block_arg + { + $$ = new_args_tail(p, Qnone, Qnone, $1, &@1); + } + | args_forward + { + add_forwarding_args(p); + $$ = new_args_tail(p, Qnone, $1, ID2VAL(idFWD_BLOCK), &@1); } - | error term {yyerrok; $$ = 0;} ; -f_arglist : '(' f_args opt_nl ')' +opt_args_tail : ',' args_tail { $$ = $2; - lex_state = EXPR_BEG; - command_start = Qtrue; } - | f_args term + | /* none */ { - $$ = $1; + $$ = new_args_tail(p, Qnone, Qnone, Qnone, &@0); } ; -f_args : f_arg ',' f_optarg ',' f_rest_arg opt_f_block_arg +f_args : f_arg ',' f_optarg ',' f_rest_arg opt_args_tail { - $$ = block_append(NEW_ARGS($1, $3, $5), $6); + $$ = new_args(p, $1, $3, $5, Qnone, $6, &@$); } - | f_arg ',' f_optarg opt_f_block_arg + | f_arg ',' f_optarg ',' f_rest_arg ',' f_arg opt_args_tail { - $$ = block_append(NEW_ARGS($1, $3, 0), $4); + $$ = new_args(p, $1, $3, $5, $7, $8, &@$); } - | f_arg ',' f_rest_arg opt_f_block_arg + | f_arg ',' f_optarg opt_args_tail { - $$ = block_append(NEW_ARGS($1, 0, $3), $4); + $$ = new_args(p, $1, $3, Qnone, Qnone, $4, &@$); } - | f_arg opt_f_block_arg + | f_arg ',' f_optarg ',' f_arg opt_args_tail { - $$ = block_append(NEW_ARGS($1, 0, 0), $2); + $$ = new_args(p, $1, $3, Qnone, $5, $6, &@$); } - | f_optarg ',' f_rest_arg opt_f_block_arg + | f_arg ',' f_rest_arg opt_args_tail { - $$ = block_append(NEW_ARGS(0, $1, $3), $4); + $$ = new_args(p, $1, Qnone, $3, Qnone, $4, &@$); } - | f_optarg opt_f_block_arg + | f_arg ',' f_rest_arg ',' f_arg opt_args_tail { - $$ = block_append(NEW_ARGS(0, $1, 0), $2); + $$ = new_args(p, $1, Qnone, $3, $5, $6, &@$); } - | f_rest_arg opt_f_block_arg + | f_arg opt_args_tail { - $$ = block_append(NEW_ARGS(0, 0, $1), $2); + $$ = new_args(p, $1, Qnone, Qnone, Qnone, $2, &@$); } - | f_block_arg + | f_optarg ',' f_rest_arg opt_args_tail + { + $$ = new_args(p, Qnone, $1, $3, Qnone, $4, &@$); + } + | f_optarg ',' f_rest_arg ',' f_arg opt_args_tail + { + $$ = new_args(p, Qnone, $1, $3, $5, $6, &@$); + } + | f_optarg opt_args_tail + { + $$ = new_args(p, Qnone, $1, Qnone, Qnone, $2, &@$); + } + | f_optarg ',' f_arg opt_args_tail { - $$ = block_append(NEW_ARGS(0, 0, 0), $1); + $$ = new_args(p, Qnone, $1, Qnone, $3, $4, &@$); + } + | f_rest_arg opt_args_tail + { + $$ = new_args(p, Qnone, Qnone, $1, Qnone, $2, &@$); + } + | f_rest_arg ',' f_arg opt_args_tail + { + $$ = new_args(p, Qnone, Qnone, $1, $3, $4, &@$); + } + | args_tail + { + $$ = new_args(p, Qnone, Qnone, Qnone, Qnone, $1, &@$); } | /* none */ { - $$ = NEW_ARGS(0, 0, 0); + $$ = new_args_tail(p, Qnone, Qnone, Qnone, &@0); + $$ = new_args(p, Qnone, Qnone, Qnone, Qnone, $$, &@0); + } + ; + +args_forward : tBDOT3 + { + /*%%%*/ + $$ = idFWD_KWREST; + /*% %*/ + /*% ripper: args_forward! %*/ } ; -f_norm_arg : tCONSTANT +f_bad_arg : tCONSTANT { - yyerror("formal argument cannot be a constant"); + static const char mesg[] = "formal argument cannot be a constant"; + /*%%%*/ + yyerror1(&@1, mesg); + $$ = 0; + /*% %*/ + /*% ripper[error]: param_error!(ERR_MESG(), $1) %*/ } - | tIVAR + | tIVAR { - yyerror("formal argument cannot be an instance variable"); + static const char mesg[] = "formal argument cannot be an instance variable"; + /*%%%*/ + yyerror1(&@1, mesg); + $$ = 0; + /*% %*/ + /*% ripper[error]: param_error!(ERR_MESG(), $1) %*/ } - | tGVAR + | tGVAR { - yyerror("formal argument cannot be a global variable"); + static const char mesg[] = "formal argument cannot be a global variable"; + /*%%%*/ + yyerror1(&@1, mesg); + $$ = 0; + /*% %*/ + /*% ripper[error]: param_error!(ERR_MESG(), $1) %*/ } - | tCVAR + | tCVAR { - yyerror("formal argument cannot be a class variable"); + static const char mesg[] = "formal argument cannot be a class variable"; + /*%%%*/ + yyerror1(&@1, mesg); + $$ = 0; + /*% %*/ + /*% ripper[error]: param_error!(ERR_MESG(), $1) %*/ } + ; + +f_norm_arg : f_bad_arg | tIDENTIFIER { - if (!is_local_id($1)) - yyerror("formal argument must be local variable"); - else if (local_id($1)) - yyerror("duplicate argument name"); - local_cnt($1); - $$ = 1; + formal_argument(p, $1); + p->max_numparam = ORDINAL_PARAM; + $$ = $1; } ; -f_arg : f_norm_arg - | f_arg ',' f_norm_arg +f_arg_asgn : f_norm_arg { - $$ += 1; + ID id = get_id($1); + arg_var(p, id); + p->cur_arg = id; + $$ = $1; } ; -f_opt : tIDENTIFIER '=' arg_value +f_arg_item : f_arg_asgn { - if (!is_local_id($1)) - yyerror("formal argument must be local variable"); - else if (local_id($1)) - yyerror("duplicate optional argument name"); - $$ = assignable($1, $3); + p->cur_arg = 0; + /*%%%*/ + $$ = NEW_ARGS_AUX($1, 1, &NULL_LOC); + /*% %*/ + /*% ripper: get_value($1) %*/ + } + | tLPAREN f_margs rparen + { + /*%%%*/ + ID tid = internal_id(p); + YYLTYPE loc; + loc.beg_pos = @2.beg_pos; + loc.end_pos = @2.beg_pos; + arg_var(p, tid); + if (dyna_in_block(p)) { + $2->nd_value = NEW_DVAR(tid, &loc); + } + else { + $2->nd_value = NEW_LVAR(tid, &loc); + } + $$ = NEW_ARGS_AUX(tid, 1, &NULL_LOC); + $$->nd_next = $2; + /*% %*/ + /*% ripper: mlhs_paren!($2) %*/ + } + ; + +f_arg : f_arg_item + /*% ripper[brace]: rb_ary_new3(1, get_value($1)) %*/ + | f_arg ',' f_arg_item + { + /*%%%*/ + $$ = $1; + $$->nd_plen++; + $$->nd_next = block_append(p, $$->nd_next, $3->nd_next); + rb_discard_node(p, $3); + /*% %*/ + /*% ripper: rb_ary_push($1, get_value($3)) %*/ + } + ; + + +f_label : tLABEL + { + arg_var(p, formal_argument(p, $1)); + p->cur_arg = get_id($1); + p->max_numparam = ORDINAL_PARAM; + p->ctxt.in_argdef = 0; + $$ = $1; + } + ; + +f_kw : f_label arg_value + { + p->cur_arg = 0; + p->ctxt.in_argdef = 1; + /*%%%*/ + $$ = new_kw_arg(p, assignable(p, $1, $2, &@$), &@$); + /*% %*/ + /*% ripper: rb_assoc_new(get_value(assignable(p, $1)), get_value($2)) %*/ + } + | f_label + { + p->cur_arg = 0; + p->ctxt.in_argdef = 1; + /*%%%*/ + $$ = new_kw_arg(p, assignable(p, $1, NODE_SPECIAL_REQUIRED_KEYWORD, &@$), &@$); + /*% %*/ + /*% ripper: rb_assoc_new(get_value(assignable(p, $1)), 0) %*/ + } + ; + +f_block_kw : f_label primary_value + { + p->ctxt.in_argdef = 1; + /*%%%*/ + $$ = new_kw_arg(p, assignable(p, $1, $2, &@$), &@$); + /*% %*/ + /*% ripper: rb_assoc_new(get_value(assignable(p, $1)), get_value($2)) %*/ + } + | f_label + { + p->ctxt.in_argdef = 1; + /*%%%*/ + $$ = new_kw_arg(p, assignable(p, $1, NODE_SPECIAL_REQUIRED_KEYWORD, &@$), &@$); + /*% %*/ + /*% ripper: rb_assoc_new(get_value(assignable(p, $1)), 0) %*/ + } + ; + +f_block_kwarg : f_block_kw + { + /*%%%*/ + $$ = $1; + /*% %*/ + /*% ripper: rb_ary_new3(1, get_value($1)) %*/ + } + | f_block_kwarg ',' f_block_kw + { + /*%%%*/ + $$ = kwd_append($1, $3); + /*% %*/ + /*% ripper: rb_ary_push($1, get_value($3)) %*/ + } + ; + + +f_kwarg : f_kw + { + /*%%%*/ + $$ = $1; + /*% %*/ + /*% ripper: rb_ary_new3(1, get_value($1)) %*/ + } + | f_kwarg ',' f_kw + { + /*%%%*/ + $$ = kwd_append($1, $3); + /*% %*/ + /*% ripper: rb_ary_push($1, get_value($3)) %*/ + } + ; + +kwrest_mark : tPOW + | tDSTAR + ; + +f_no_kwarg : kwrest_mark keyword_nil + { + /*%%%*/ + /*% %*/ + /*% ripper: nokw_param!(Qnil) %*/ + } + ; + +f_kwrest : kwrest_mark tIDENTIFIER + { + arg_var(p, shadowing_lvar(p, get_id($2))); + /*%%%*/ + $$ = $2; + /*% %*/ + /*% ripper: kwrest_param!($2) %*/ + } + | kwrest_mark + { + /*%%%*/ + $$ = internal_id(p); + arg_var(p, $$); + /*% %*/ + /*% ripper: kwrest_param!(Qnil) %*/ + } + ; + +f_opt : f_arg_asgn f_eq arg_value + { + p->cur_arg = 0; + p->ctxt.in_argdef = 1; + /*%%%*/ + $$ = NEW_OPT_ARG(0, assignable(p, $1, $3, &@$), &@$); + /*% %*/ + /*% ripper: rb_assoc_new(get_value(assignable(p, $1)), get_value($3)) %*/ + } + ; + +f_block_opt : f_arg_asgn f_eq primary_value + { + p->cur_arg = 0; + p->ctxt.in_argdef = 1; + /*%%%*/ + $$ = NEW_OPT_ARG(0, assignable(p, $1, $3, &@$), &@$); + /*% %*/ + /*% ripper: rb_assoc_new(get_value(assignable(p, $1)), get_value($3)) %*/ + } + ; + +f_block_optarg : f_block_opt + { + /*%%%*/ + $$ = $1; + /*% %*/ + /*% ripper: rb_ary_new3(1, get_value($1)) %*/ + } + | f_block_optarg ',' f_block_opt + { + /*%%%*/ + $$ = opt_arg_append($1, $3); + /*% %*/ + /*% ripper: rb_ary_push($1, get_value($3)) %*/ } ; f_optarg : f_opt { - $$ = NEW_BLOCK($1); - $$->nd_end = $$; + /*%%%*/ + $$ = $1; + /*% %*/ + /*% ripper: rb_ary_new3(1, get_value($1)) %*/ } | f_optarg ',' f_opt { - $$ = block_append($1, $3); + /*%%%*/ + $$ = opt_arg_append($1, $3); + /*% %*/ + /*% ripper: rb_ary_push($1, get_value($3)) %*/ } ; @@ -2340,21 +5554,19 @@ restarg_mark : '*' f_rest_arg : restarg_mark tIDENTIFIER { - if (!is_local_id($2)) - yyerror("rest argument must be local variable"); - if (dyna_in_block()) { - rb_dvar_push($2, Qnil); - } - $$ = assignable($2, 0); + arg_var(p, shadowing_lvar(p, get_id($2))); + /*%%%*/ + $$ = $2; + /*% %*/ + /*% ripper: rest_param!($2) %*/ } | restarg_mark { - if (dyna_in_block()) { - $$ = NEW_DASGN_CURR(internal_id(), 0); - } - else { - $$ = NEW_NODE(NODE_LASGN,0,0,local_append(0)); - } + /*%%%*/ + $$ = internal_id(p); + arg_var(p, $$); + /*% %*/ + /*% ripper: rest_param!(Qnil) %*/ } ; @@ -2364,12 +5576,20 @@ blkarg_mark : '&' f_block_arg : blkarg_mark tIDENTIFIER { - if (!is_local_id($2)) - yyerror("block argument must be local variable"); - else if (local_id($2)) - yyerror("duplicate block argument name"); - $$ = NEW_BLOCK_ARG($2); - } + arg_var(p, shadowing_lvar(p, get_id($2))); + /*%%%*/ + $$ = $2; + /*% %*/ + /*% ripper: blockarg!($2) %*/ + } + | blkarg_mark + { + /*%%%*/ + arg_var(p, shadowing_lvar(p, get_id(ANON_BLOCK_ID))); + /*% + $$ = dispatch1(blockarg, Qnil); + %*/ + } ; opt_f_block_arg : ',' f_block_arg @@ -2377,62 +5597,126 @@ opt_f_block_arg : ',' f_block_arg $$ = $2; } | none + { + $$ = Qnull; + } ; singleton : var_ref { + value_expr($1); $$ = $1; - value_expr($$); } - | '(' {lex_state = EXPR_BEG;} expr opt_nl ')' + | '(' {SET_LEX_STATE(EXPR_BEG);} expr rparen { - if ($3 == 0) { - yyerror("can't define singleton method for ()."); - } - else { - switch (nd_type($3)) { - case NODE_STR: - case NODE_DSTR: - case NODE_XSTR: - case NODE_DXSTR: - case NODE_DREGX: - case NODE_LIT: - case NODE_ARRAY: - case NODE_ZARRAY: - yyerror("can't define singleton method for literals"); - default: - value_expr($3); - break; - } + /*%%%*/ + switch (nd_type($3)) { + case NODE_STR: + case NODE_DSTR: + case NODE_XSTR: + case NODE_DXSTR: + case NODE_DREGX: + case NODE_LIT: + case NODE_LIST: + case NODE_ZLIST: + yyerror1(&@3, "can't define singleton method for literals"); + break; + default: + value_expr($3); + break; } $$ = $3; + /*% %*/ + /*% ripper: paren!($3) %*/ } ; assoc_list : none | assocs trailer { + /*%%%*/ $$ = $1; - } - | args trailer - { - if ($1->nd_alen%2 != 0) { - yyerror("odd number list for Hash"); - } - $$ = $1; + /*% %*/ + /*% ripper: assoclist_from_args!($1) %*/ } ; assocs : assoc + /*% ripper[brace]: rb_ary_new3(1, get_value($1)) %*/ | assocs ',' assoc { - $$ = list_concat($1, $3); + /*%%%*/ + NODE *assocs = $1; + NODE *tail = $3; + if (!assocs) { + assocs = tail; + } + else if (tail) { + if (assocs->nd_head && + !tail->nd_head && nd_type_p(tail->nd_next, NODE_LIST) && + nd_type_p(tail->nd_next->nd_head, NODE_HASH)) { + /* DSTAR */ + tail = tail->nd_next->nd_head->nd_head; + } + assocs = list_concat(assocs, tail); + } + $$ = assocs; + /*% %*/ + /*% ripper: rb_ary_push($1, get_value($3)) %*/ } ; assoc : arg_value tASSOC arg_value { - $$ = list_append(NEW_LIST($1), $3); + /*%%%*/ + if (nd_type_p($1, NODE_STR)) { + nd_set_type($1, NODE_LIT); + RB_OBJ_WRITE(p->ast, &$1->nd_lit, rb_fstring($1->nd_lit)); + } + $$ = list_append(p, NEW_LIST($1, &@$), $3); + /*% %*/ + /*% ripper: assoc_new!($1, $3) %*/ + } + | tLABEL arg_value + { + /*%%%*/ + $$ = list_append(p, NEW_LIST(NEW_LIT(ID2SYM($1), &@1), &@$), $2); + /*% %*/ + /*% ripper: assoc_new!($1, $2) %*/ + } + | tLABEL + { + /*%%%*/ + NODE *val = gettable(p, $1, &@$); + if (!val) val = NEW_BEGIN(0, &@$); + $$ = list_append(p, NEW_LIST(NEW_LIT(ID2SYM($1), &@1), &@$), val); + /*% %*/ + /*% ripper: assoc_new!($1, Qnil) %*/ + } + | tSTRING_BEG string_contents tLABEL_END arg_value + { + /*%%%*/ + YYLTYPE loc = code_loc_gen(&@1, &@3); + $$ = list_append(p, NEW_LIST(dsym_node(p, $2, &loc), &loc), $4); + /*% %*/ + /*% ripper: assoc_new!(dyna_symbol!($2), $4) %*/ + } + | tDSTAR arg_value + { + /*%%%*/ + if (nd_type_p($2, NODE_HASH) && + !($2->nd_head && $2->nd_head->nd_alen)) { + static VALUE empty_hash; + if (!empty_hash) { + empty_hash = rb_obj_freeze(rb_hash_new()); + rb_gc_register_mark_object(empty_hash); + } + $$ = list_append(p, NEW_LIST(0, &@$), NEW_LIT(empty_hash, &@$)); + } + else + $$ = list_append(p, NEW_LIST(0, &@$), $2); + /*% %*/ + /*% ripper: assoc_splat!($2) %*/ } ; @@ -2456,6 +5740,14 @@ dot_or_colon : '.' | tCOLON2 ; +call_op : '.' + | tANDDOT + ; + +call_op2 : call_op + | tCOLON2 + ; + opt_terms : /* none */ | terms ; @@ -2464,321 +5756,1150 @@ opt_nl : /* none */ | '\n' ; +rparen : opt_nl ')' + ; + +rbracket : opt_nl ']' + ; + +rbrace : opt_nl '}' + ; + trailer : /* none */ | '\n' | ',' ; -term : ';' {yyerrok;} - | '\n' +term : ';' {yyerrok;token_flush(p);} + | '\n' {token_flush(p);} ; terms : term | terms ';' {yyerrok;} ; -none : /* none */ {$$ = 0;} +none : /* none */ + { + $$ = Qnull; + } ; %% -#ifdef yystacksize -#undef YYMALLOC +# undef p +# undef yylex +# undef yylval +# define yylval (*p->lval) + +static int regx_options(struct parser_params*); +static int tokadd_string(struct parser_params*,int,int,int,long*,rb_encoding**,rb_encoding**); +static void tokaddmbc(struct parser_params *p, int c, rb_encoding *enc); +static enum yytokentype parse_string(struct parser_params*,rb_strterm_literal_t*); +static enum yytokentype here_document(struct parser_params*,rb_strterm_heredoc_t*); + +#ifndef RIPPER +# define set_yylval_node(x) { \ + YYLTYPE _cur_loc; \ + rb_parser_set_location(p, &_cur_loc); \ + yylval.node = (x); \ +} +# define set_yylval_str(x) \ +do { \ + set_yylval_node(NEW_STR(x, &_cur_loc)); \ + RB_OBJ_WRITTEN(p->ast, Qnil, x); \ +} while(0) +# define set_yylval_literal(x) \ +do { \ + set_yylval_node(NEW_LIT(x, &_cur_loc)); \ + RB_OBJ_WRITTEN(p->ast, Qnil, x); \ +} while(0) +# define set_yylval_num(x) (yylval.num = (x)) +# define set_yylval_id(x) (yylval.id = (x)) +# define set_yylval_name(x) (yylval.id = (x)) +# define yylval_id() (yylval.id) +#else +static inline VALUE +ripper_yylval_id(struct parser_params *p, ID x) +{ + return ripper_new_yylval(p, x, ID2SYM(x), 0); +} +# define set_yylval_str(x) (yylval.val = add_mark_object(p, (x))) +# define set_yylval_num(x) (yylval.val = ripper_new_yylval(p, (x), 0, 0)) +# define set_yylval_id(x) (void)(x) +# define set_yylval_name(x) (void)(yylval.val = ripper_yylval_id(p, x)) +# define set_yylval_literal(x) add_mark_object(p, (x)) +# define set_yylval_node(x) (yylval.val = ripper_new_yylval(p, 0, 0, STR_NEW(p->lex.ptok, p->lex.pcur-p->lex.ptok))) +# define yylval_id() yylval.id +# define _cur_loc NULL_LOC /* dummy */ #endif -#include "regex.h" -#include "util.h" - -/* We remove any previous definition of `SIGN_EXTEND_CHAR', - since ours (we hope) works properly with all combinations of - machines, compilers, `char' and `unsigned char' argument types. - (Per Bothner suggested the basic approach.) */ -#undef SIGN_EXTEND_CHAR -#if __STDC__ -# define SIGN_EXTEND_CHAR(c) ((signed char)(c)) -#else /* not __STDC__ */ -/* As in Harbison and Steele. */ -# define SIGN_EXTEND_CHAR(c) ((((unsigned char)(c)) ^ 128) - 128) -#endif -#define is_identchar(c) (SIGN_EXTEND_CHAR(c)!=-1&&(ISALNUM(c) || (c) == '_' || ismbchar(c))) +#define set_yylval_noname() set_yylval_id(keyword_nil) + +#ifndef RIPPER +#define literal_flush(p, ptr) ((p)->lex.ptok = (ptr)) +#define dispatch_scan_event(p, t) ((void)0) +#define dispatch_delayed_token(p, t) ((void)0) +#define has_delayed_token(p) (0) +#else +#define literal_flush(p, ptr) ((void)(ptr)) + +#define yylval_rval (*(RB_TYPE_P(yylval.val, T_NODE) ? &yylval.node->nd_rval : &yylval.val)) + +static inline VALUE +intern_sym(const char *name) +{ + ID id = rb_intern_const(name); + return ID2SYM(id); +} + +static int +ripper_has_scan_event(struct parser_params *p) +{ + if (p->lex.pcur < p->lex.ptok) rb_raise(rb_eRuntimeError, "lex.pcur < lex.ptok"); + return p->lex.pcur > p->lex.ptok; +} + +static VALUE +ripper_scan_event_val(struct parser_params *p, enum yytokentype t) +{ + VALUE str = STR_NEW(p->lex.ptok, p->lex.pcur - p->lex.ptok); + VALUE rval = ripper_dispatch1(p, ripper_token2eventid(t), str); + token_flush(p); + return rval; +} + +static void +ripper_dispatch_scan_event(struct parser_params *p, enum yytokentype t) +{ + if (!ripper_has_scan_event(p)) return; + add_mark_object(p, yylval_rval = ripper_scan_event_val(p, t)); +} +#define dispatch_scan_event(p, t) ripper_dispatch_scan_event(p, t) + +static void +ripper_dispatch_delayed_token(struct parser_params *p, enum yytokentype t) +{ + int saved_line = p->ruby_sourceline; + const char *saved_tokp = p->lex.ptok; + + if (NIL_P(p->delayed.token)) return; + p->ruby_sourceline = p->delayed.line; + p->lex.ptok = p->lex.pbeg + p->delayed.col; + add_mark_object(p, yylval_rval = ripper_dispatch1(p, ripper_token2eventid(t), p->delayed.token)); + p->delayed.token = Qnil; + p->ruby_sourceline = saved_line; + p->lex.ptok = saved_tokp; +} +#define dispatch_delayed_token(p, t) ripper_dispatch_delayed_token(p, t) +#define has_delayed_token(p) (!NIL_P(p->delayed.token)) +#endif /* RIPPER */ + +static inline int +is_identchar(const char *ptr, const char *MAYBE_UNUSED(ptr_end), rb_encoding *enc) +{ + return rb_enc_isalnum((unsigned char)*ptr, enc) || *ptr == '_' || !ISASCII(*ptr); +} + +static inline int +parser_is_identchar(struct parser_params *p) +{ + return !(p)->eofp && is_identchar(p->lex.pcur-1, p->lex.pend, p->enc); +} + +static inline int +parser_isascii(struct parser_params *p) +{ + return ISASCII(*(p->lex.pcur-1)); +} + +static void +token_info_setup(token_info *ptinfo, const char *ptr, const rb_code_location_t *loc) +{ + int column = 1, nonspc = 0, i; + for (i = 0; i < loc->beg_pos.column; i++, ptr++) { + if (*ptr == '\t') { + column = (((column - 1) / TAB_WIDTH) + 1) * TAB_WIDTH; + } + column++; + if (*ptr != ' ' && *ptr != '\t') { + nonspc = 1; + } + } + + ptinfo->beg = loc->beg_pos; + ptinfo->indent = column; + ptinfo->nonspc = nonspc; +} + +static void +token_info_push(struct parser_params *p, const char *token, const rb_code_location_t *loc) +{ + token_info *ptinfo; + + if (!p->token_info_enabled) return; + ptinfo = ALLOC(token_info); + ptinfo->token = token; + ptinfo->next = p->token_info; + token_info_setup(ptinfo, p->lex.pbeg, loc); + + p->token_info = ptinfo; +} + +static void +token_info_pop(struct parser_params *p, const char *token, const rb_code_location_t *loc) +{ + token_info *ptinfo_beg = p->token_info; + + if (!ptinfo_beg) return; + p->token_info = ptinfo_beg->next; + + /* indentation check of matched keywords (begin..end, if..end, etc.) */ + token_info_warn(p, token, ptinfo_beg, 1, loc); + ruby_sized_xfree(ptinfo_beg, sizeof(*ptinfo_beg)); +} + +static void +token_info_drop(struct parser_params *p, const char *token, rb_code_position_t beg_pos) +{ + token_info *ptinfo_beg = p->token_info; + + if (!ptinfo_beg) return; + p->token_info = ptinfo_beg->next; + + if (ptinfo_beg->beg.lineno != beg_pos.lineno || + ptinfo_beg->beg.column != beg_pos.column || + strcmp(ptinfo_beg->token, token)) { + compile_error(p, "token position mismatch: %d:%d:%s expected but %d:%d:%s", + beg_pos.lineno, beg_pos.column, token, + ptinfo_beg->beg.lineno, ptinfo_beg->beg.column, + ptinfo_beg->token); + } + + ruby_sized_xfree(ptinfo_beg, sizeof(*ptinfo_beg)); +} + +static void +token_info_warn(struct parser_params *p, const char *token, token_info *ptinfo_beg, int same, const rb_code_location_t *loc) +{ + token_info ptinfo_end_body, *ptinfo_end = &ptinfo_end_body; + if (!p->token_info_enabled) return; + if (!ptinfo_beg) return; + token_info_setup(ptinfo_end, p->lex.pbeg, loc); + if (ptinfo_beg->beg.lineno == ptinfo_end->beg.lineno) return; /* ignore one-line block */ + if (ptinfo_beg->nonspc || ptinfo_end->nonspc) return; /* ignore keyword in the middle of a line */ + if (ptinfo_beg->indent == ptinfo_end->indent) return; /* the indents are matched */ + if (!same && ptinfo_beg->indent < ptinfo_end->indent) return; + rb_warn3L(ptinfo_end->beg.lineno, + "mismatched indentations at '%s' with '%s' at %d", + WARN_S(token), WARN_S(ptinfo_beg->token), WARN_I(ptinfo_beg->beg.lineno)); +} + +static int +parser_precise_mbclen(struct parser_params *p, const char *ptr) +{ + int len = rb_enc_precise_mbclen(ptr, p->lex.pend, p->enc); + if (!MBCLEN_CHARFOUND_P(len)) { + compile_error(p, "invalid multibyte char (%s)", rb_enc_name(p->enc)); + return -1; + } + return len; +} -static char *tokenbuf = NULL; -static int tokidx, toksiz = 0; +#ifndef RIPPER +static void ruby_show_error_line(VALUE errbuf, const YYLTYPE *yylloc, int lineno, VALUE str); -#define LEAVE_BS 1 +static inline void +parser_show_error_line(struct parser_params *p, const YYLTYPE *yylloc) +{ + VALUE str; + int lineno = p->ruby_sourceline; + if (!yylloc) { + return; + } + else if (yylloc->beg_pos.lineno == lineno) { + str = p->lex.lastline; + } + else { + return; + } + ruby_show_error_line(p->error_buffer, yylloc, lineno, str); +} + +static int +parser_yyerror(struct parser_params *p, const YYLTYPE *yylloc, const char *msg) +{ +#if 0 + YYLTYPE current; -static VALUE (*lex_gets)(); /* gets function */ -static VALUE lex_input; /* non-nil if File */ -static VALUE lex_lastline; /* gc protect */ -static char *lex_pbeg; -static char *lex_p; -static char *lex_pend; + if (!yylloc) { + yylloc = RUBY_SET_YYLLOC(current); + } + else if ((p->ruby_sourceline != yylloc->beg_pos.lineno && + p->ruby_sourceline != yylloc->end_pos.lineno)) { + yylloc = 0; + } +#endif + compile_error(p, "%s", msg); + parser_show_error_line(p, yylloc); + return 0; +} static int -yyerror(msg) - const char *msg; +parser_yyerror0(struct parser_params *p, const char *msg) { - char *p, *pe, *buf; - int len, i; + YYLTYPE current; + return parser_yyerror(p, RUBY_SET_YYLLOC(current), msg); +} + +static void +ruby_show_error_line(VALUE errbuf, const YYLTYPE *yylloc, int lineno, VALUE str) +{ + VALUE mesg; + const int max_line_margin = 30; + const char *ptr, *ptr_end, *pt, *pb; + const char *pre = "", *post = "", *pend; + const char *code = "", *caret = ""; + const char *lim; + const char *const pbeg = RSTRING_PTR(str); + char *buf; + long len; + int i; - rb_compile_error("%s", msg); - p = lex_p; - while (lex_pbeg <= p) { - if (*p == '\n') break; - p--; + if (!yylloc) return; + pend = RSTRING_END(str); + if (pend > pbeg && pend[-1] == '\n') { + if (--pend > pbeg && pend[-1] == '\r') --pend; } - p++; - pe = lex_p; - while (pe < lex_pend) { - if (*pe == '\n') break; - pe++; + pt = pend; + if (lineno == yylloc->end_pos.lineno && + (pend - pbeg) > yylloc->end_pos.column) { + pt = pbeg + yylloc->end_pos.column; } - len = pe - p; - if (len > 4) { - buf = ALLOCA_N(char, len+2); - MEMCPY(buf, p, char, len); - buf[len] = '\0'; - rb_compile_error_append("%s", buf); + ptr = ptr_end = pt; + lim = ptr - pbeg > max_line_margin ? ptr - max_line_margin : pbeg; + while ((lim < ptr) && (*(ptr-1) != '\n')) ptr--; - i = lex_p - p; - p = buf; pe = p + len; + lim = pend - ptr_end > max_line_margin ? ptr_end + max_line_margin : pend; + while ((ptr_end < lim) && (*ptr_end != '\n') && (*ptr_end != '\r')) ptr_end++; - while (p < pe) { - if (*p != '\t') *p = ' '; - p++; + len = ptr_end - ptr; + if (len > 4) { + if (ptr > pbeg) { + ptr = rb_enc_prev_char(pbeg, ptr, pt, rb_enc_get(str)); + if (ptr > pbeg) pre = "..."; + } + if (ptr_end < pend) { + ptr_end = rb_enc_prev_char(pt, ptr_end, pend, rb_enc_get(str)); + if (ptr_end < pend) post = "..."; + } + } + pb = pbeg; + if (lineno == yylloc->beg_pos.lineno) { + pb += yylloc->beg_pos.column; + if (pb > pt) pb = pt; + } + if (pb < ptr) pb = ptr; + if (len <= 4 && yylloc->beg_pos.lineno == yylloc->end_pos.lineno) { + return; + } + if (RTEST(errbuf)) { + mesg = rb_attr_get(errbuf, idMesg); + if (RSTRING_LEN(mesg) > 0 && *(RSTRING_END(mesg)-1) != '\n') + rb_str_cat_cstr(mesg, "\n"); + } + else { + mesg = rb_enc_str_new(0, 0, rb_enc_get(str)); + } + if (!errbuf && rb_stderr_tty_p()) { +#define CSI_BEGIN "\033[" +#define CSI_SGR "m" + rb_str_catf(mesg, + CSI_BEGIN""CSI_SGR"%s" /* pre */ + CSI_BEGIN"1"CSI_SGR"%.*s" + CSI_BEGIN"1;4"CSI_SGR"%.*s" + CSI_BEGIN";1"CSI_SGR"%.*s" + CSI_BEGIN""CSI_SGR"%s" /* post */ + "\n", + pre, + (int)(pb - ptr), ptr, + (int)(pt - pb), pb, + (int)(ptr_end - pt), pt, + post); + } + else { + char *p2; + + len = ptr_end - ptr; + lim = pt < pend ? pt : pend; + i = (int)(lim - ptr); + buf = ALLOCA_N(char, i+2); + code = ptr; + caret = p2 = buf; + if (ptr <= pb) { + while (ptr < pb) { + *p2++ = *ptr++ == '\t' ? '\t' : ' '; + } + *p2++ = '^'; + ptr++; } - buf[i] = '^'; - buf[i+1] = '\0'; - rb_compile_error_append("%s", buf); + if (lim > ptr) { + memset(p2, '~', (lim - ptr)); + p2 += (lim - ptr); + } + *p2 = '\0'; + rb_str_catf(mesg, "%s%.*s%s\n""%s%s\n", + pre, (int)len, code, post, + pre, caret); + } + if (!errbuf) rb_write_error_str(mesg); +} +#else +static int +parser_yyerror(struct parser_params *p, const YYLTYPE *yylloc, const char *msg) +{ + const char *pcur = 0, *ptok = 0; + if (p->ruby_sourceline == yylloc->beg_pos.lineno && + p->ruby_sourceline == yylloc->end_pos.lineno) { + pcur = p->lex.pcur; + ptok = p->lex.ptok; + p->lex.ptok = p->lex.pbeg + yylloc->beg_pos.column; + p->lex.pcur = p->lex.pbeg + yylloc->end_pos.column; + } + parser_yyerror0(p, msg); + if (pcur) { + p->lex.ptok = ptok; + p->lex.pcur = pcur; } + return 0; +} +static int +parser_yyerror0(struct parser_params *p, const char *msg) +{ + dispatch1(parse_error, STR_NEW2(msg)); + ripper_error(p); return 0; } -static int heredoc_end; +static inline void +parser_show_error_line(struct parser_params *p, const YYLTYPE *yylloc) +{ +} +#endif /* !RIPPER */ -int ruby_in_compile = 0; -int ruby__end__seen; +#ifndef RIPPER +static int +vtable_size(const struct vtable *tbl) +{ + if (!DVARS_TERMINAL_P(tbl)) { + return tbl->pos; + } + else { + return 0; + } +} +#endif -static VALUE ruby_debug_lines; -#ifdef YYMALLOC -static NODE *parser_heap; +static struct vtable * +vtable_alloc_gen(struct parser_params *p, int line, struct vtable *prev) +{ + struct vtable *tbl = ALLOC(struct vtable); + tbl->pos = 0; + tbl->capa = 8; + tbl->tbl = ALLOC_N(ID, tbl->capa); + tbl->prev = prev; +#ifndef RIPPER + if (p->debug) { + rb_parser_printf(p, "vtable_alloc:%d: %p\n", line, (void *)tbl); + } #endif + return tbl; +} +#define vtable_alloc(prev) vtable_alloc_gen(p, __LINE__, prev) -static NODE* -yycompile(f, line) - char *f; - int line; +static void +vtable_free_gen(struct parser_params *p, int line, const char *name, + struct vtable *tbl) { - int n; - NODE *node = 0; - struct RVarmap *vp, *vars = ruby_dyna_vars; - - ruby_in_compile = 1; - if (!compile_for_eval && rb_safe_level() == 0 && - rb_const_defined(rb_cObject, rb_intern("SCRIPT_LINES__"))) { - VALUE hash, fname; - - hash = rb_const_get(rb_cObject, rb_intern("SCRIPT_LINES__")); - if (TYPE(hash) == T_HASH) { - fname = rb_str_new2(f); - ruby_debug_lines = rb_hash_aref(hash, fname); - if (NIL_P(ruby_debug_lines)) { - ruby_debug_lines = rb_ary_new(); - rb_hash_aset(hash, fname, ruby_debug_lines); - } - } - if (line > 1) { - VALUE str = rb_str_new(0,0); - while (line > 1) { - rb_ary_push(ruby_debug_lines, str); - line--; - } - } - } - - ruby__end__seen = 0; - ruby_eval_tree = 0; - heredoc_end = 0; - lex_strterm = 0; - ruby_current_node = 0; - ruby_sourcefile = rb_source_filename(f); - n = yyparse(); - ruby_debug_lines = 0; - compile_for_eval = 0; - ruby_in_compile = 0; - cond_stack = 0; - cmdarg_stack = 0; - command_start = 1; - class_nest = 0; - in_single = 0; - in_def = 0; - cur_mid = 0; - - vp = ruby_dyna_vars; - ruby_dyna_vars = vars; - lex_strterm = 0; - while (vp && vp != vars) { - struct RVarmap *tmp = vp; - vp = vp->next; - rb_gc_force_recycle((VALUE)tmp); - } - if (n == 0) node = ruby_eval_tree; - else ruby_eval_tree_begin = 0; - return node; +#ifndef RIPPER + if (p->debug) { + rb_parser_printf(p, "vtable_free:%d: %s(%p)\n", line, name, (void *)tbl); + } +#endif + if (!DVARS_TERMINAL_P(tbl)) { + if (tbl->tbl) { + ruby_sized_xfree(tbl->tbl, tbl->capa * sizeof(ID)); + } + ruby_sized_xfree(tbl, sizeof(*tbl)); + } } +#define vtable_free(tbl) vtable_free_gen(p, __LINE__, #tbl, tbl) -static int lex_gets_ptr; +static void +vtable_add_gen(struct parser_params *p, int line, const char *name, + struct vtable *tbl, ID id) +{ +#ifndef RIPPER + if (p->debug) { + rb_parser_printf(p, "vtable_add:%d: %s(%p), %s\n", + line, name, (void *)tbl, rb_id2name(id)); + } +#endif + if (DVARS_TERMINAL_P(tbl)) { + rb_parser_fatal(p, "vtable_add: vtable is not allocated (%p)", (void *)tbl); + return; + } + if (tbl->pos == tbl->capa) { + tbl->capa = tbl->capa * 2; + SIZED_REALLOC_N(tbl->tbl, ID, tbl->capa, tbl->pos); + } + tbl->tbl[tbl->pos++] = id; +} +#define vtable_add(tbl, id) vtable_add_gen(p, __LINE__, #tbl, tbl, id) -static VALUE -lex_get_str(s) - VALUE s; +#ifndef RIPPER +static void +vtable_pop_gen(struct parser_params *p, int line, const char *name, + struct vtable *tbl, int n) +{ + if (p->debug) { + rb_parser_printf(p, "vtable_pop:%d: %s(%p), %d\n", + line, name, (void *)tbl, n); + } + if (tbl->pos < n) { + rb_parser_fatal(p, "vtable_pop: unreachable (%d < %d)", tbl->pos, n); + return; + } + tbl->pos -= n; +} +#define vtable_pop(tbl, n) vtable_pop_gen(p, __LINE__, #tbl, tbl, n) +#endif + +static int +vtable_included(const struct vtable * tbl, ID id) { - char *beg, *end, *pend; + int i; - beg = RSTRING(s)->ptr; - if (lex_gets_ptr) { - if (RSTRING(s)->len == lex_gets_ptr) return Qnil; - beg += lex_gets_ptr; + if (!DVARS_TERMINAL_P(tbl)) { + for (i = 0; i < tbl->pos; i++) { + if (tbl->tbl[i] == id) { + return i+1; + } + } } - pend = RSTRING(s)->ptr + RSTRING(s)->len; - end = beg; - while (end < pend) { - if (*end++ == '\n') break; + return 0; +} + +static void parser_prepare(struct parser_params *p); + +#ifndef RIPPER +static NODE *parser_append_options(struct parser_params *p, NODE *node); + +static VALUE +debug_lines(VALUE fname) +{ + ID script_lines; + CONST_ID(script_lines, "SCRIPT_LINES__"); + if (rb_const_defined_at(rb_cObject, script_lines)) { + VALUE hash = rb_const_get_at(rb_cObject, script_lines); + if (RB_TYPE_P(hash, T_HASH)) { + VALUE lines = rb_ary_new(); + rb_hash_aset(hash, fname, lines); + return lines; + } } - lex_gets_ptr = end - RSTRING(s)->ptr; - return rb_str_new(beg, end - beg); + return 0; +} + +static int +e_option_supplied(struct parser_params *p) +{ + return strcmp(p->ruby_sourcefile, "-e") == 0; } static VALUE -lex_getline() +yycompile0(VALUE arg) +{ + int n; + NODE *tree; + struct parser_params *p = (struct parser_params *)arg; + VALUE cov = Qfalse; + + if (!compile_for_eval && !NIL_P(p->ruby_sourcefile_string)) { + p->debug_lines = debug_lines(p->ruby_sourcefile_string); + if (p->debug_lines && p->ruby_sourceline > 0) { + VALUE str = rb_default_rs; + n = p->ruby_sourceline; + do { + rb_ary_push(p->debug_lines, str); + } while (--n); + } + + if (!e_option_supplied(p)) { + cov = Qtrue; + } + } + + if (p->keep_script_lines || ruby_vm_keep_script_lines) { + if (!p->debug_lines) { + p->debug_lines = rb_ary_new(); + } + + RB_OBJ_WRITE(p->ast, &p->ast->body.script_lines, p->debug_lines); + } + + parser_prepare(p); +#define RUBY_DTRACE_PARSE_HOOK(name) \ + if (RUBY_DTRACE_PARSE_##name##_ENABLED()) { \ + RUBY_DTRACE_PARSE_##name(p->ruby_sourcefile, p->ruby_sourceline); \ + } + RUBY_DTRACE_PARSE_HOOK(BEGIN); + n = yyparse(p); + RUBY_DTRACE_PARSE_HOOK(END); + p->debug_lines = 0; + + p->lex.strterm = 0; + p->lex.pcur = p->lex.pbeg = p->lex.pend = 0; + p->lex.prevline = p->lex.lastline = p->lex.nextline = 0; + if (n || p->error_p) { + VALUE mesg = p->error_buffer; + if (!mesg) { + mesg = rb_class_new_instance(0, 0, rb_eSyntaxError); + } + rb_set_errinfo(mesg); + return FALSE; + } + tree = p->eval_tree; + if (!tree) { + tree = NEW_NIL(&NULL_LOC); + } + else { + VALUE opt = p->compile_option; + NODE *prelude; + NODE *body = parser_append_options(p, tree->nd_body); + if (!opt) opt = rb_obj_hide(rb_ident_hash_new()); + rb_hash_aset(opt, rb_sym_intern_ascii_cstr("coverage_enabled"), cov); + prelude = block_append(p, p->eval_tree_begin, body); + tree->nd_body = prelude; + RB_OBJ_WRITE(p->ast, &p->ast->body.compile_option, opt); + } + p->ast->body.root = tree; + if (!p->ast->body.script_lines) p->ast->body.script_lines = INT2FIX(p->line_count); + return TRUE; +} + +static rb_ast_t * +yycompile(VALUE vparser, struct parser_params *p, VALUE fname, int line) +{ + rb_ast_t *ast; + if (NIL_P(fname)) { + p->ruby_sourcefile_string = Qnil; + p->ruby_sourcefile = "(none)"; + } + else { + p->ruby_sourcefile_string = rb_fstring(fname); + p->ruby_sourcefile = StringValueCStr(fname); + } + p->ruby_sourceline = line - 1; + + p->lvtbl = NULL; + + p->ast = ast = rb_ast_new(); + rb_suppress_tracing(yycompile0, (VALUE)p); + p->ast = 0; + RB_GC_GUARD(vparser); /* prohibit tail call optimization */ + + while (p->lvtbl) { + local_pop(p); + } + + return ast; +} +#endif /* !RIPPER */ + +static rb_encoding * +must_be_ascii_compatible(VALUE s) { - VALUE line = (*lex_gets)(lex_input); - if (ruby_debug_lines && !NIL_P(line)) { - rb_ary_push(ruby_debug_lines, line); + rb_encoding *enc = rb_enc_get(s); + if (!rb_enc_asciicompat(enc)) { + rb_raise(rb_eArgError, "invalid source encoding"); } + return enc; +} + +static VALUE +lex_get_str(struct parser_params *p, VALUE s) +{ + char *beg, *end, *start; + long len; + + beg = RSTRING_PTR(s); + len = RSTRING_LEN(s); + start = beg; + if (p->lex.gets_.ptr) { + if (len == p->lex.gets_.ptr) return Qnil; + beg += p->lex.gets_.ptr; + len -= p->lex.gets_.ptr; + } + end = memchr(beg, '\n', len); + if (end) len = ++end - beg; + p->lex.gets_.ptr += len; + return rb_str_subseq(s, beg - start, len); +} + +static VALUE +lex_getline(struct parser_params *p) +{ + VALUE line = (*p->lex.gets)(p, p->lex.input); + if (NIL_P(line)) return line; + must_be_ascii_compatible(line); + if (RB_OBJ_FROZEN(line)) line = rb_str_dup(line); // needed for RubyVM::AST.of because script_lines in iseq is deep-frozen + p->line_count++; return line; } -NODE* -rb_compile_string(f, s, line) - const char *f; - VALUE s; - int line; +static const rb_data_type_t parser_data_type; + +#ifndef RIPPER +static rb_ast_t* +parser_compile_string(VALUE vparser, VALUE fname, VALUE s, int line) +{ + struct parser_params *p; + + TypedData_Get_Struct(vparser, struct parser_params, &parser_data_type, p); + + p->lex.gets = lex_get_str; + p->lex.gets_.ptr = 0; + p->lex.input = rb_str_new_frozen(s); + p->lex.pbeg = p->lex.pcur = p->lex.pend = 0; + + return yycompile(vparser, p, fname, line); +} + +rb_ast_t* +rb_parser_compile_string(VALUE vparser, const char *f, VALUE s, int line) { - lex_gets = lex_get_str; - lex_gets_ptr = 0; - lex_input = s; - lex_pbeg = lex_p = lex_pend = 0; - ruby_sourceline = line - 1; - compile_for_eval = ruby_in_eval; + return rb_parser_compile_string_path(vparser, rb_filesystem_str_new_cstr(f), s, line); +} - return yycompile(f, line); +rb_ast_t* +rb_parser_compile_string_path(VALUE vparser, VALUE f, VALUE s, int line) +{ + must_be_ascii_compatible(s); + return parser_compile_string(vparser, f, s, line); } -NODE* -rb_compile_cstr(f, s, len, line) - const char *f, *s; - int len, line; +VALUE rb_io_gets_internal(VALUE io); + +static VALUE +lex_io_gets(struct parser_params *p, VALUE io) { - return rb_compile_string(f, rb_str_new(s, len), line); + return rb_io_gets_internal(io); } -NODE* -rb_compile_file(f, file, start) - const char *f; - VALUE file; - int start; +rb_ast_t* +rb_parser_compile_file_path(VALUE vparser, VALUE fname, VALUE file, int start) { - lex_gets = rb_io_gets; - lex_input = file; - lex_pbeg = lex_p = lex_pend = 0; - ruby_sourceline = start - 1; + struct parser_params *p; - return yycompile(f, start); + TypedData_Get_Struct(vparser, struct parser_params, &parser_data_type, p); + + p->lex.gets = lex_io_gets; + p->lex.input = file; + p->lex.pbeg = p->lex.pcur = p->lex.pend = 0; + + return yycompile(vparser, p, fname, start); } -static inline int -nextc() +static VALUE +lex_generic_gets(struct parser_params *p, VALUE input) { - int c; + return (*p->lex.gets_.call)(input, p->line_count); +} - if (lex_p == lex_pend) { - if (lex_input) { - VALUE v = lex_getline(); +rb_ast_t* +rb_parser_compile_generic(VALUE vparser, VALUE (*lex_gets)(VALUE, int), VALUE fname, VALUE input, int start) +{ + struct parser_params *p; - if (NIL_P(v)) return -1; - if (heredoc_end > 0) { - ruby_sourceline = heredoc_end; - heredoc_end = 0; - } - ruby_sourceline++; - lex_pbeg = lex_p = RSTRING(v)->ptr; - lex_pend = lex_p + RSTRING(v)->len; - lex_lastline = v; + TypedData_Get_Struct(vparser, struct parser_params, &parser_data_type, p); + + p->lex.gets = lex_generic_gets; + p->lex.gets_.call = lex_gets; + p->lex.input = input; + p->lex.pbeg = p->lex.pcur = p->lex.pend = 0; + + return yycompile(vparser, p, fname, start); +} +#endif /* !RIPPER */ + +#define STR_FUNC_ESCAPE 0x01 +#define STR_FUNC_EXPAND 0x02 +#define STR_FUNC_REGEXP 0x04 +#define STR_FUNC_QWORDS 0x08 +#define STR_FUNC_SYMBOL 0x10 +#define STR_FUNC_INDENT 0x20 +#define STR_FUNC_LABEL 0x40 +#define STR_FUNC_LIST 0x4000 +#define STR_FUNC_TERM 0x8000 + +enum string_type { + str_label = STR_FUNC_LABEL, + str_squote = (0), + str_dquote = (STR_FUNC_EXPAND), + str_xquote = (STR_FUNC_EXPAND), + str_regexp = (STR_FUNC_REGEXP|STR_FUNC_ESCAPE|STR_FUNC_EXPAND), + str_sword = (STR_FUNC_QWORDS|STR_FUNC_LIST), + str_dword = (STR_FUNC_QWORDS|STR_FUNC_EXPAND|STR_FUNC_LIST), + str_ssym = (STR_FUNC_SYMBOL), + str_dsym = (STR_FUNC_SYMBOL|STR_FUNC_EXPAND) +}; + +static VALUE +parser_str_new(const char *ptr, long len, rb_encoding *enc, int func, rb_encoding *enc0) +{ + VALUE str; + + str = rb_enc_str_new(ptr, len, enc); + if (!(func & STR_FUNC_REGEXP) && rb_enc_asciicompat(enc)) { + if (rb_enc_str_coderange(str) == ENC_CODERANGE_7BIT) { } - else { - lex_lastline = 0; - return -1; + else if (enc0 == rb_usascii_encoding() && enc != rb_utf8_encoding()) { + rb_enc_associate(str, rb_ascii8bit_encoding()); + } + } + + return str; +} + +#define lex_goto_eol(p) ((p)->lex.pcur = (p)->lex.pend) +#define lex_eol_p(p) ((p)->lex.pcur >= (p)->lex.pend) +#define lex_eol_n_p(p,n) ((p)->lex.pcur+(n) >= (p)->lex.pend) +#define peek(p,c) peek_n(p, (c), 0) +#define peek_n(p,c,n) (!lex_eol_n_p(p, n) && (c) == (unsigned char)(p)->lex.pcur[n]) +#define peekc(p) peekc_n(p, 0) +#define peekc_n(p,n) (lex_eol_n_p(p, n) ? -1 : (unsigned char)(p)->lex.pcur[n]) + +#ifdef RIPPER +static void +add_delayed_token(struct parser_params *p, const char *tok, const char *end) +{ + if (tok < end) { + if (!has_delayed_token(p)) { + p->delayed.token = rb_str_buf_new(end - tok); + rb_enc_associate(p->delayed.token, p->enc); + p->delayed.line = p->ruby_sourceline; + p->delayed.col = rb_long2int(tok - p->lex.pbeg); } + rb_str_buf_cat(p->delayed.token, tok, end - tok); + p->lex.ptok = end; } - c = (unsigned char)*lex_p++; - if (c == '\r' && lex_p < lex_pend && *lex_p == '\n') { - lex_p++; +} +#else +#define add_delayed_token(p, tok, end) ((void)(tok), (void)(end)) +#endif + +static int +nextline(struct parser_params *p, int set_encoding) +{ + VALUE v = p->lex.nextline; + p->lex.nextline = 0; + if (!v) { + if (p->eofp) + return -1; + + if (p->lex.pend > p->lex.pbeg && *(p->lex.pend-1) != '\n') { + goto end_of_input; + } + + if (!p->lex.input || NIL_P(v = lex_getline(p))) { + end_of_input: + p->eofp = 1; + lex_goto_eol(p); + return -1; + } +#ifndef RIPPER + if (p->debug_lines) { + if (set_encoding) rb_enc_associate(v, p->enc); + rb_ary_push(p->debug_lines, v); + } +#endif + p->cr_seen = FALSE; + } + else if (NIL_P(v)) { + /* after here-document without terminator */ + goto end_of_input; + } + add_delayed_token(p, p->lex.ptok, p->lex.pend); + if (p->heredoc_end > 0) { + p->ruby_sourceline = p->heredoc_end; + p->heredoc_end = 0; + } + p->ruby_sourceline++; + p->lex.pbeg = p->lex.pcur = RSTRING_PTR(v); + p->lex.pend = p->lex.pcur + RSTRING_LEN(v); + token_flush(p); + p->lex.prevline = p->lex.lastline; + p->lex.lastline = v; + return 0; +} + +static int +parser_cr(struct parser_params *p, int c) +{ + if (peek(p, '\n')) { + p->lex.pcur++; c = '\n'; } + return c; +} + +static inline int +nextc0(struct parser_params *p, int set_encoding) +{ + int c; + + if (UNLIKELY((p->lex.pcur == p->lex.pend) || p->eofp || RTEST(p->lex.nextline))) { + if (nextline(p, set_encoding)) return -1; + } + c = (unsigned char)*p->lex.pcur++; + if (UNLIKELY(c == '\r')) { + c = parser_cr(p, c); + } return c; } +#define nextc(p) nextc0(p, TRUE) static void -pushback(c) - int c; +pushback(struct parser_params *p, int c) { if (c == -1) return; - lex_p--; + p->lex.pcur--; + if (p->lex.pcur > p->lex.pbeg && p->lex.pcur[0] == '\n' && p->lex.pcur[-1] == '\r') { + p->lex.pcur--; + } } -#define was_bol() (lex_p == lex_pbeg + 1) -#define peek(c) (lex_p != lex_pend && (c) == *lex_p) +#define was_bol(p) ((p)->lex.pcur == (p)->lex.pbeg + 1) + +#define tokfix(p) ((p)->tokenbuf[(p)->tokidx]='\0') +#define tok(p) (p)->tokenbuf +#define toklen(p) (p)->tokidx -#define tokfix() (tokenbuf[tokidx]='\0') -#define tok() tokenbuf -#define toklen() tokidx -#define toklast() (tokidx>0?tokenbuf[tokidx-1]:0) +static int +looking_at_eol_p(struct parser_params *p) +{ + const char *ptr = p->lex.pcur; + while (ptr < p->lex.pend) { + int c = (unsigned char)*ptr++; + int eol = (c == '\n' || c == '#'); + if (eol || !ISSPACE(c)) { + return eol; + } + } + return TRUE; +} static char* -newtok() +newtok(struct parser_params *p) +{ + p->tokidx = 0; + p->tokline = p->ruby_sourceline; + if (!p->tokenbuf) { + p->toksiz = 60; + p->tokenbuf = ALLOC_N(char, 60); + } + if (p->toksiz > 4096) { + p->toksiz = 60; + REALLOC_N(p->tokenbuf, char, 60); + } + return p->tokenbuf; +} + +static char * +tokspace(struct parser_params *p, int n) +{ + p->tokidx += n; + + if (p->tokidx >= p->toksiz) { + do {p->toksiz *= 2;} while (p->toksiz < p->tokidx); + REALLOC_N(p->tokenbuf, char, p->toksiz); + } + return &p->tokenbuf[p->tokidx-n]; +} + +static void +tokadd(struct parser_params *p, int c) +{ + p->tokenbuf[p->tokidx++] = (char)c; + if (p->tokidx >= p->toksiz) { + p->toksiz *= 2; + REALLOC_N(p->tokenbuf, char, p->toksiz); + } +} + +static int +tok_hex(struct parser_params *p, size_t *numlen) +{ + int c; + + c = scan_hex(p->lex.pcur, 2, numlen); + if (!*numlen) { + yyerror0("invalid hex escape"); + token_flush(p); + return 0; + } + p->lex.pcur += *numlen; + return c; +} + +#define tokcopy(p, n) memcpy(tokspace(p, n), (p)->lex.pcur - (n), (n)) + +static int +escaped_control_code(int c) +{ + int c2 = 0; + switch (c) { + case ' ': + c2 = 's'; + break; + case '\n': + c2 = 'n'; + break; + case '\t': + c2 = 't'; + break; + case '\v': + c2 = 'v'; + break; + case '\r': + c2 = 'r'; + break; + case '\f': + c2 = 'f'; + break; + } + return c2; +} + +#define WARN_SPACE_CHAR(c, prefix) \ + rb_warn1("invalid character syntax; use "prefix"\\%c", WARN_I(c2)) + +static int +tokadd_codepoint(struct parser_params *p, rb_encoding **encp, + int regexp_literal, int wide) { - tokidx = 0; - if (!tokenbuf) { - toksiz = 60; - tokenbuf = ALLOC_N(char, 60); + size_t numlen; + int codepoint = scan_hex(p->lex.pcur, wide ? p->lex.pend - p->lex.pcur : 4, &numlen); + literal_flush(p, p->lex.pcur); + p->lex.pcur += numlen; + if (wide ? (numlen == 0 || numlen > 6) : (numlen < 4)) { + yyerror0("invalid Unicode escape"); + return wide && numlen > 0; + } + if (codepoint > 0x10ffff) { + yyerror0("invalid Unicode codepoint (too large)"); + return wide; + } + if ((codepoint & 0xfffff800) == 0xd800) { + yyerror0("invalid Unicode codepoint"); + return wide; + } + if (regexp_literal) { + tokcopy(p, (int)numlen); + } + else if (codepoint >= 0x80) { + rb_encoding *utf8 = rb_utf8_encoding(); + if (*encp && utf8 != *encp) { + YYLTYPE loc = RUBY_INIT_YYLLOC(); + compile_error(p, "UTF-8 mixed within %s source", rb_enc_name(*encp)); + parser_show_error_line(p, &loc); + return wide; + } + *encp = utf8; + tokaddmbc(p, codepoint, *encp); } - if (toksiz > 4096) { - toksiz = 60; - REALLOC_N(tokenbuf, char, 60); + else { + tokadd(p, codepoint); } - return tokenbuf; + return TRUE; } +/* return value is for ?\u3042 */ static void -tokadd(c) - char c; +tokadd_utf8(struct parser_params *p, rb_encoding **encp, + int term, int symbol_literal, int regexp_literal) { - tokenbuf[tokidx++] = c; - if (tokidx >= toksiz) { - toksiz *= 2; - REALLOC_N(tokenbuf, char, toksiz); + /* + * If `term` is not -1, then we allow multiple codepoints in \u{} + * upto `term` byte, otherwise we're parsing a character literal. + * And then add the codepoints to the current token. + */ + static const char multiple_codepoints[] = "Multiple codepoints at single character literal"; + + const int open_brace = '{', close_brace = '}'; + + if (regexp_literal) { tokadd(p, '\\'); tokadd(p, 'u'); } + + if (peek(p, open_brace)) { /* handle \u{...} form */ + const char *second = NULL; + int c, last = nextc(p); + if (p->lex.pcur >= p->lex.pend) goto unterminated; + while (ISSPACE(c = *p->lex.pcur) && ++p->lex.pcur < p->lex.pend); + while (c != close_brace) { + if (c == term) goto unterminated; + if (second == multiple_codepoints) + second = p->lex.pcur; + if (regexp_literal) tokadd(p, last); + if (!tokadd_codepoint(p, encp, regexp_literal, TRUE)) { + break; + } + while (ISSPACE(c = *p->lex.pcur)) { + if (++p->lex.pcur >= p->lex.pend) goto unterminated; + last = c; + } + if (term == -1 && !second) + second = multiple_codepoints; + } + + if (c != close_brace) { + unterminated: + token_flush(p); + yyerror0("unterminated Unicode escape"); + return; + } + if (second && second != multiple_codepoints) { + const char *pcur = p->lex.pcur; + p->lex.pcur = second; + dispatch_scan_event(p, tSTRING_CONTENT); + token_flush(p); + p->lex.pcur = pcur; + yyerror0(multiple_codepoints); + token_flush(p); + } + + if (regexp_literal) tokadd(p, close_brace); + nextc(p); + } + else { /* handle \uxxxx form */ + if (!tokadd_codepoint(p, encp, regexp_literal, FALSE)) { + token_flush(p); + return; + } } } +#define ESCAPE_CONTROL 1 +#define ESCAPE_META 2 + static int -read_escape() +read_escape(struct parser_params *p, int flags, rb_encoding **encp) { int c; + size_t numlen; - switch (c = nextc()) { + switch (c = nextc(p)) { case '\\': /* Backslash */ return c; @@ -2805,26 +6926,14 @@ read_escape() case '0': case '1': case '2': case '3': /* octal constant */ case '4': case '5': case '6': case '7': - { - int numlen; - - pushback(c); - c = scan_oct(lex_p, 3, &numlen); - lex_p += numlen; - } + pushback(p, c); + c = scan_oct(p->lex.pcur, 3, &numlen); + p->lex.pcur += numlen; return c; case 'x': /* hex constant */ - { - int numlen; - - c = scan_hex(lex_p, 2, &numlen); - if (numlen == 0) { - yyerror("Invalid escape character syntax"); - return 0; - } - lex_p += numlen; - } + c = tok_hex(p, &numlen); + if (numlen == 0) return 0; return c; case 'b': /* backspace */ @@ -2834,37 +6943,78 @@ read_escape() return ' '; case 'M': - if ((c = nextc()) != '-') { - yyerror("Invalid escape character syntax"); - pushback(c); - return '\0'; - } - if ((c = nextc()) == '\\') { - return read_escape() | 0x80; + if (flags & ESCAPE_META) goto eof; + if ((c = nextc(p)) != '-') { + goto eof; + } + if ((c = nextc(p)) == '\\') { + switch (peekc(p)) { + case 'u': case 'U': + nextc(p); + goto eof; + } + return read_escape(p, flags|ESCAPE_META, encp) | 0x80; } - else if (c == -1) goto eof; + else if (c == -1 || !ISASCII(c)) goto eof; else { + int c2 = escaped_control_code(c); + if (c2) { + if (ISCNTRL(c) || !(flags & ESCAPE_CONTROL)) { + WARN_SPACE_CHAR(c2, "\\M-"); + } + else { + WARN_SPACE_CHAR(c2, "\\C-\\M-"); + } + } + else if (ISCNTRL(c)) goto eof; return ((c & 0xff) | 0x80); } case 'C': - if ((c = nextc()) != '-') { - yyerror("Invalid escape character syntax"); - pushback(c); - return '\0'; + if ((c = nextc(p)) != '-') { + goto eof; } case 'c': - if ((c = nextc())== '\\') { - c = read_escape(); + if (flags & ESCAPE_CONTROL) goto eof; + if ((c = nextc(p))== '\\') { + switch (peekc(p)) { + case 'u': case 'U': + nextc(p); + goto eof; + } + c = read_escape(p, flags|ESCAPE_CONTROL, encp); } else if (c == '?') return 0177; - else if (c == -1) goto eof; + else if (c == -1 || !ISASCII(c)) goto eof; + else { + int c2 = escaped_control_code(c); + if (c2) { + if (ISCNTRL(c)) { + if (flags & ESCAPE_META) { + WARN_SPACE_CHAR(c2, "\\M-"); + } + else { + WARN_SPACE_CHAR(c2, ""); + } + } + else { + if (flags & ESCAPE_META) { + WARN_SPACE_CHAR(c2, "\\M-\\C-"); + } + else { + WARN_SPACE_CHAR(c2, "\\C-"); + } + } + } + else if (ISCNTRL(c)) goto eof; + } return c & 0x9f; eof: case -1: - yyerror("Invalid escape character syntax"); + yyerror0("Invalid escape character syntax"); + token_flush(p); return '\0'; default: @@ -2872,308 +7022,533 @@ read_escape() } } +static void +tokaddmbc(struct parser_params *p, int c, rb_encoding *enc) +{ + int len = rb_enc_codelen(c, enc); + rb_enc_mbcput(c, tokspace(p, len), enc); +} + static int -tokadd_escape(term) - int term; +tokadd_escape(struct parser_params *p, rb_encoding **encp) { int c; + size_t numlen; - switch (c = nextc()) { + switch (c = nextc(p)) { case '\n': return 0; /* just ignore */ case '0': case '1': case '2': case '3': /* octal constant */ case '4': case '5': case '6': case '7': { - int i; - - tokadd('\\'); - tokadd(c); - for (i=0; i<2; i++) { - c = nextc(); - if (c == -1) goto eof; - if (c < '0' || '7' < c) { - pushback(c); - break; - } - tokadd(c); - } + ruby_scan_oct(--p->lex.pcur, 3, &numlen); + if (numlen == 0) goto eof; + p->lex.pcur += numlen; + tokcopy(p, (int)numlen + 1); } return 0; case 'x': /* hex constant */ { - int numlen; - - tokadd('\\'); - tokadd(c); - scan_hex(lex_p, 2, &numlen); - if (numlen == 0) { - yyerror("Invalid escape character syntax"); - return -1; - } - while (numlen--) - tokadd(nextc()); - } - return 0; - - case 'M': - if ((c = nextc()) != '-') { - yyerror("Invalid escape character syntax"); - pushback(c); - return 0; - } - tokadd('\\'); tokadd('M'); tokadd('-'); - goto escaped; - - case 'C': - if ((c = nextc()) != '-') { - yyerror("Invalid escape character syntax"); - pushback(c); - return 0; - } - tokadd('\\'); tokadd('C'); tokadd('-'); - goto escaped; - - case 'c': - tokadd('\\'); tokadd('c'); - escaped: - if ((c = nextc()) == '\\') { - return tokadd_escape(term); + tok_hex(p, &numlen); + if (numlen == 0) return -1; + tokcopy(p, (int)numlen + 2); } - else if (c == -1) goto eof; - tokadd(c); return 0; eof: case -1: - yyerror("Invalid escape character syntax"); + yyerror0("Invalid escape character syntax"); + token_flush(p); return -1; default: - if (c != '\\' || c != term) - tokadd('\\'); - tokadd(c); + tokadd(p, '\\'); + tokadd(p, c); } return 0; } static int -regx_options() +regx_options(struct parser_params *p) { - char kcode = 0; + int kcode = 0; + int kopt = 0; int options = 0; - int c; + int c, opt, kc; - newtok(); - while (c = nextc(), ISALPHA(c)) { - switch (c) { - case 'i': - options |= RE_OPTION_IGNORECASE; - break; - case 'x': - options |= RE_OPTION_EXTENDED; - break; - case 'm': - options |= RE_OPTION_MULTILINE; - break; - case 'o': - options |= RE_OPTION_ONCE; - break; - case 'n': - kcode = 16; - break; - case 'e': - kcode = 32; - break; - case 's': - kcode = 48; - break; - case 'u': - kcode = 64; - break; - default: - tokadd(c); - break; - } + newtok(p); + while (c = nextc(p), ISALPHA(c)) { + if (c == 'o') { + options |= RE_OPTION_ONCE; + } + else if (rb_char_to_option_kcode(c, &opt, &kc)) { + if (kc >= 0) { + if (kc != rb_ascii8bit_encindex()) kcode = c; + kopt = opt; + } + else { + options |= opt; + } + } + else { + tokadd(p, c); + } } - pushback(c); - if (toklen()) { - tokfix(); - rb_compile_error("unknown regexp option%s - %s", - toklen() > 1 ? "s" : "", tok()); + options |= kopt; + pushback(p, c); + if (toklen(p)) { + YYLTYPE loc = RUBY_INIT_YYLLOC(); + tokfix(p); + compile_error(p, "unknown regexp option%s - %*s", + toklen(p) > 1 ? "s" : "", toklen(p), tok(p)); + parser_show_error_line(p, &loc); } - return options | kcode; + return options | RE_OPTION_ENCODING(kcode); } -#define STR_FUNC_ESCAPE 0x01 -#define STR_FUNC_EXPAND 0x02 -#define STR_FUNC_REGEXP 0x04 -#define STR_FUNC_QWORDS 0x08 -#define STR_FUNC_SYMBOL 0x10 -#define STR_FUNC_INDENT 0x20 +static int +tokadd_mbchar(struct parser_params *p, int c) +{ + int len = parser_precise_mbclen(p, p->lex.pcur-1); + if (len < 0) return -1; + tokadd(p, c); + p->lex.pcur += --len; + if (len > 0) tokcopy(p, len); + return c; +} -enum string_type { - str_squote = (0), - str_dquote = (STR_FUNC_EXPAND), - str_xquote = (STR_FUNC_EXPAND), - str_regexp = (STR_FUNC_REGEXP|STR_FUNC_ESCAPE|STR_FUNC_EXPAND), - str_sword = (STR_FUNC_QWORDS), - str_dword = (STR_FUNC_QWORDS|STR_FUNC_EXPAND), - str_ssym = (STR_FUNC_SYMBOL), - str_dsym = (STR_FUNC_SYMBOL|STR_FUNC_EXPAND), -}; +static inline int +simple_re_meta(int c) +{ + switch (c) { + case '$': case '*': case '+': case '.': + case '?': case '^': case '|': + case ')': case ']': case '}': case '>': + return TRUE; + default: + return FALSE; + } +} + +static int +parser_update_heredoc_indent(struct parser_params *p, int c) +{ + if (p->heredoc_line_indent == -1) { + if (c == '\n') p->heredoc_line_indent = 0; + } + else { + if (c == ' ') { + p->heredoc_line_indent++; + return TRUE; + } + else if (c == '\t') { + int w = (p->heredoc_line_indent / TAB_WIDTH) + 1; + p->heredoc_line_indent = w * TAB_WIDTH; + return TRUE; + } + else if (c != '\n') { + if (p->heredoc_indent > p->heredoc_line_indent) { + p->heredoc_indent = p->heredoc_line_indent; + } + p->heredoc_line_indent = -1; + } + } + return FALSE; +} static void -dispose_string(str) - VALUE str; +parser_mixed_error(struct parser_params *p, rb_encoding *enc1, rb_encoding *enc2) +{ + YYLTYPE loc = RUBY_INIT_YYLLOC(); + const char *n1 = rb_enc_name(enc1), *n2 = rb_enc_name(enc2); + compile_error(p, "%s mixed within %s source", n1, n2); + parser_show_error_line(p, &loc); +} + +static void +parser_mixed_escape(struct parser_params *p, const char *beg, rb_encoding *enc1, rb_encoding *enc2) { - xfree(RSTRING(str)->ptr); - rb_gc_force_recycle(str); + const char *pos = p->lex.pcur; + p->lex.pcur = beg; + parser_mixed_error(p, enc1, enc2); + p->lex.pcur = pos; } static int -tokadd_string(func, term, paren, nest) - int func, term, paren, *nest; +tokadd_string(struct parser_params *p, + int func, int term, int paren, long *nest, + rb_encoding **encp, rb_encoding **enc) { int c; + bool erred = false; + +#define mixed_error(enc1, enc2) \ + (void)(erred || (parser_mixed_error(p, enc1, enc2), erred = true)) +#define mixed_escape(beg, enc1, enc2) \ + (void)(erred || (parser_mixed_escape(p, beg, enc1, enc2), erred = true)) + + while ((c = nextc(p)) != -1) { + if (p->heredoc_indent > 0) { + parser_update_heredoc_indent(p, c); + } - while ((c = nextc()) != -1) { if (paren && c == paren) { ++*nest; } else if (c == term) { if (!nest || !*nest) { - pushback(c); + pushback(p, c); break; } --*nest; } - else if ((func & STR_FUNC_EXPAND) && c == '#' && lex_p < lex_pend) { - int c2 = *lex_p; + else if ((func & STR_FUNC_EXPAND) && c == '#' && p->lex.pcur < p->lex.pend) { + int c2 = *p->lex.pcur; if (c2 == '$' || c2 == '@' || c2 == '{') { - pushback(c); + pushback(p, c); break; } } else if (c == '\\') { - c = nextc(); + literal_flush(p, p->lex.pcur - 1); + c = nextc(p); switch (c) { case '\n': if (func & STR_FUNC_QWORDS) break; - if (func & STR_FUNC_EXPAND) continue; - tokadd('\\'); + if (func & STR_FUNC_EXPAND) { + if (!(func & STR_FUNC_INDENT) || (p->heredoc_indent < 0)) + continue; + if (c == term) { + c = '\\'; + goto terminate; + } + } + tokadd(p, '\\'); break; case '\\': - if (func & STR_FUNC_ESCAPE) tokadd(c); + if (func & STR_FUNC_ESCAPE) tokadd(p, c); break; + case 'u': + if ((func & STR_FUNC_EXPAND) == 0) { + tokadd(p, '\\'); + break; + } + tokadd_utf8(p, enc, term, + func & STR_FUNC_SYMBOL, + func & STR_FUNC_REGEXP); + continue; + default: + if (c == -1) return -1; + if (!ISASCII(c)) { + if ((func & STR_FUNC_EXPAND) == 0) tokadd(p, '\\'); + goto non_ascii; + } if (func & STR_FUNC_REGEXP) { - pushback(c); - if (tokadd_escape(term) < 0) + switch (c) { + case 'c': + case 'C': + case 'M': { + pushback(p, c); + c = read_escape(p, 0, enc); + + int i; + char escbuf[5]; + snprintf(escbuf, sizeof(escbuf), "\\x%02X", c); + for (i = 0; i < 4; i++) { + tokadd(p, escbuf[i]); + } + continue; + } + } + + if (c == term && !simple_re_meta(c)) { + tokadd(p, c); + continue; + } + pushback(p, c); + if ((c = tokadd_escape(p, enc)) < 0) return -1; + if (*enc && *enc != *encp) { + mixed_escape(p->lex.ptok+2, *enc, *encp); + } continue; } else if (func & STR_FUNC_EXPAND) { - pushback(c); - if (func & STR_FUNC_ESCAPE) tokadd('\\'); - c = read_escape(); + pushback(p, c); + if (func & STR_FUNC_ESCAPE) tokadd(p, '\\'); + c = read_escape(p, 0, enc); } else if ((func & STR_FUNC_QWORDS) && ISSPACE(c)) { /* ignore backslashed spaces in %w */ } else if (c != term && !(paren && c == paren)) { - tokadd('\\'); + tokadd(p, '\\'); + pushback(p, c); + continue; } } } - else if (ismbchar(c)) { - int i, len = mbclen(c)-1; - - for (i = 0; i < len; i++) { - tokadd(c); - c = nextc(); + else if (!parser_isascii(p)) { + non_ascii: + if (!*enc) { + *enc = *encp; + } + else if (*enc != *encp) { + mixed_error(*enc, *encp); + continue; } + if (tokadd_mbchar(p, c) == -1) return -1; + continue; } else if ((func & STR_FUNC_QWORDS) && ISSPACE(c)) { - pushback(c); + pushback(p, c); break; } - if (!c && (func & STR_FUNC_SYMBOL)) { - func &= ~STR_FUNC_SYMBOL; - rb_compile_error("symbol cannot contain '\\0'"); - continue; - } - tokadd(c); + if (c & 0x80) { + if (!*enc) { + *enc = *encp; + } + else if (*enc != *encp) { + mixed_error(*enc, *encp); + continue; + } + } + tokadd(p, c); } + terminate: + if (*enc) *encp = *enc; return c; } +static inline rb_strterm_t * +new_strterm(VALUE v1, VALUE v2, VALUE v3, VALUE v0) +{ + return (rb_strterm_t*)rb_imemo_new(imemo_parser_strterm, v1, v2, v3, v0); +} + +/* imemo_parser_strterm for literal */ #define NEW_STRTERM(func, term, paren) \ - rb_node_newnode(NODE_STRTERM, (func), (term) | ((paren) << (CHAR_BIT * 2)), 0) + new_strterm((VALUE)(func), (VALUE)(paren), (VALUE)(term), 0) -static int -parse_string(quote) - NODE *quote; +#ifdef RIPPER +static void +flush_string_content(struct parser_params *p, rb_encoding *enc) +{ + VALUE content = yylval.val; + if (!ripper_is_node_yylval(content)) + content = ripper_new_yylval(p, 0, 0, content); + if (has_delayed_token(p)) { + ptrdiff_t len = p->lex.pcur - p->lex.ptok; + if (len > 0) { + rb_enc_str_buf_cat(p->delayed.token, p->lex.ptok, len, enc); + } + dispatch_delayed_token(p, tSTRING_CONTENT); + p->lex.ptok = p->lex.pcur; + RNODE(content)->nd_rval = yylval.val; + } + dispatch_scan_event(p, tSTRING_CONTENT); + if (yylval.val != content) + RNODE(content)->nd_rval = yylval.val; + yylval.val = content; +} +#else +#define flush_string_content(p, enc) ((void)(enc)) +#endif + +RUBY_FUNC_EXPORTED const unsigned int ruby_global_name_punct_bits[(0x7e - 0x20 + 31) / 32]; +/* this can be shared with ripper, since it's independent from struct + * parser_params. */ +#ifndef RIPPER +#define BIT(c, idx) (((c) / 32 - 1 == idx) ? (1U << ((c) % 32)) : 0) +#define SPECIAL_PUNCT(idx) ( \ + BIT('~', idx) | BIT('*', idx) | BIT('$', idx) | BIT('?', idx) | \ + BIT('!', idx) | BIT('@', idx) | BIT('/', idx) | BIT('\\', idx) | \ + BIT(';', idx) | BIT(',', idx) | BIT('.', idx) | BIT('=', idx) | \ + BIT(':', idx) | BIT('<', idx) | BIT('>', idx) | BIT('\"', idx) | \ + BIT('&', idx) | BIT('`', idx) | BIT('\'', idx) | BIT('+', idx) | \ + BIT('0', idx)) +const unsigned int ruby_global_name_punct_bits[] = { + SPECIAL_PUNCT(0), + SPECIAL_PUNCT(1), + SPECIAL_PUNCT(2), +}; +#undef BIT +#undef SPECIAL_PUNCT +#endif + +static enum yytokentype +parser_peek_variable_name(struct parser_params *p) +{ + int c; + const char *ptr = p->lex.pcur; + + if (ptr + 1 >= p->lex.pend) return 0; + c = *ptr++; + switch (c) { + case '$': + if ((c = *ptr) == '-') { + if (++ptr >= p->lex.pend) return 0; + c = *ptr; + } + else if (is_global_name_punct(c) || ISDIGIT(c)) { + return tSTRING_DVAR; + } + break; + case '@': + if ((c = *ptr) == '@') { + if (++ptr >= p->lex.pend) return 0; + c = *ptr; + } + break; + case '{': + p->lex.pcur = ptr; + p->command_start = TRUE; + return tSTRING_DBEG; + default: + return 0; + } + if (!ISASCII(c) || c == '_' || ISALPHA(c)) + return tSTRING_DVAR; + return 0; +} + +#define IS_ARG() IS_lex_state(EXPR_ARG_ANY) +#define IS_END() IS_lex_state(EXPR_END_ANY) +#define IS_BEG() (IS_lex_state(EXPR_BEG_ANY) || IS_lex_state_all(EXPR_ARG|EXPR_LABELED)) +#define IS_SPCARG(c) (IS_ARG() && space_seen && !ISSPACE(c)) +#define IS_LABEL_POSSIBLE() (\ + (IS_lex_state(EXPR_LABEL|EXPR_ENDFN) && !cmd_state) || \ + IS_ARG()) +#define IS_LABEL_SUFFIX(n) (peek_n(p, ':',(n)) && !peek_n(p, ':', (n)+1)) +#define IS_AFTER_OPERATOR() IS_lex_state(EXPR_FNAME | EXPR_DOT) + +static inline enum yytokentype +parser_string_term(struct parser_params *p, int func) +{ + p->lex.strterm = 0; + if (func & STR_FUNC_REGEXP) { + set_yylval_num(regx_options(p)); + dispatch_scan_event(p, tREGEXP_END); + SET_LEX_STATE(EXPR_END); + return tREGEXP_END; + } + if ((func & STR_FUNC_LABEL) && IS_LABEL_SUFFIX(0)) { + nextc(p); + SET_LEX_STATE(EXPR_BEG|EXPR_LABEL); + return tLABEL_END; + } + SET_LEX_STATE(EXPR_END); + return tSTRING_END; +} + +static enum yytokentype +parse_string(struct parser_params *p, rb_strterm_literal_t *quote) { - int func = quote->nd_func; - int term = nd_term(quote); - int paren = nd_paren(quote); + int func = (int)quote->u1.func; + int term = (int)quote->u3.term; + int paren = (int)quote->u2.paren; int c, space = 0; + rb_encoding *enc = p->enc; + rb_encoding *base_enc = 0; + VALUE lit; - if (func == -1) return tSTRING_END; - c = nextc(); + if (func & STR_FUNC_TERM) { + if (func & STR_FUNC_QWORDS) nextc(p); /* delayed term */ + SET_LEX_STATE(EXPR_END); + p->lex.strterm = 0; + return func & STR_FUNC_REGEXP ? tREGEXP_END : tSTRING_END; + } + c = nextc(p); if ((func & STR_FUNC_QWORDS) && ISSPACE(c)) { - do {c = nextc();} while (ISSPACE(c)); + do {c = nextc(p);} while (ISSPACE(c)); + space = 1; + } + if (func & STR_FUNC_LIST) { + quote->u1.func &= ~STR_FUNC_LIST; space = 1; } - if (c == term && !quote->nd_nest) { + if (c == term && !quote->u0.nest) { if (func & STR_FUNC_QWORDS) { - quote->nd_func = -1; + quote->u1.func |= STR_FUNC_TERM; + pushback(p, c); /* dispatch the term at tSTRING_END */ + add_delayed_token(p, p->lex.ptok, p->lex.pcur); return ' '; } - if (!(func & STR_FUNC_REGEXP)) return tSTRING_END; - yylval.num = regx_options(); - return tREGEXP_END; + return parser_string_term(p, func); } if (space) { - pushback(c); + pushback(p, c); + add_delayed_token(p, p->lex.ptok, p->lex.pcur); return ' '; } - newtok(); + newtok(p); if ((func & STR_FUNC_EXPAND) && c == '#') { - switch (c = nextc()) { - case '$': - case '@': - pushback(c); - return tSTRING_DVAR; - case '{': - return tSTRING_DBEG; + int t = parser_peek_variable_name(p); + if (t) return t; + tokadd(p, '#'); + c = nextc(p); + } + pushback(p, c); + if (tokadd_string(p, func, term, paren, "e->u0.nest, + &enc, &base_enc) == -1) { + if (p->eofp) { +#ifndef RIPPER +# define unterminated_literal(mesg) yyerror0(mesg) +#else +# define unterminated_literal(mesg) compile_error(p, mesg) +#endif + literal_flush(p, p->lex.pcur); + if (func & STR_FUNC_QWORDS) { + /* no content to add, bailing out here */ + unterminated_literal("unterminated list meets end of file"); + p->lex.strterm = 0; + return tSTRING_END; + } + if (func & STR_FUNC_REGEXP) { + unterminated_literal("unterminated regexp meets end of file"); + } + else { + unterminated_literal("unterminated string meets end of file"); + } + quote->u1.func |= STR_FUNC_TERM; } - tokadd('#'); - } - pushback(c); - if (tokadd_string(func, term, paren, "e->nd_nest) == -1) { - ruby_sourceline = nd_line(quote); - rb_compile_error("unterminated string meets end of file"); - return tSTRING_END; } - tokfix(); - yylval.node = NEW_STR(rb_str_new(tok(), toklen())); + tokfix(p); + lit = STR_NEW3(tok(p), toklen(p), enc, func); + set_yylval_str(lit); + flush_string_content(p, enc); + return tSTRING_CONTENT; } -static int -heredoc_identifier() +static enum yytokentype +heredoc_identifier(struct parser_params *p) { - int c = nextc(), term, func = 0, len; + /* + * term_len is length of `<<"END"` except `END`, + * in this case term_len is 4 (<, <, " and "). + */ + long len, offset = p->lex.pcur - p->lex.pbeg; + int c = nextc(p), term, func = 0, quote = 0; + enum yytokentype token = tSTRING_BEG; + int indent = 0; if (c == '-') { - c = nextc(); + c = nextc(p); + func = STR_FUNC_INDENT; + offset++; + } + else if (c == '~') { + c = nextc(p); func = STR_FUNC_INDENT; + offset++; + indent = INT_MAX; } switch (c) { case '\'': @@ -3181,206 +7556,1727 @@ heredoc_identifier() case '"': func |= str_dquote; goto quoted; case '`': - func |= str_xquote; + token = tXSTRING_BEG; + func |= str_xquote; goto quoted; + quoted: - newtok(); - tokadd(func); + quote++; + offset++; term = c; - while ((c = nextc()) != -1 && c != term) { - len = mbclen(c); - do {tokadd(c);} while (--len > 0 && (c = nextc()) != -1); - } - if (c == -1) { - rb_compile_error("unterminated here document identifier"); - return 0; + len = 0; + while ((c = nextc(p)) != term) { + if (c == -1 || c == '\r' || c == '\n') { + yyerror0("unterminated here document identifier"); + return -1; + } } break; default: - if (!is_identchar(c)) { - pushback(c); + if (!parser_is_identchar(p)) { + pushback(p, c); if (func & STR_FUNC_INDENT) { - pushback('-'); + pushback(p, indent > 0 ? '~' : '-'); } return 0; } - newtok(); - term = '"'; - tokadd(func |= str_dquote); + func |= str_dquote; do { - len = mbclen(c); - do {tokadd(c);} while (--len > 0 && (c = nextc()) != -1); - } while ((c = nextc()) != -1 && is_identchar(c)); - pushback(c); + int n = parser_precise_mbclen(p, p->lex.pcur-1); + if (n < 0) return 0; + p->lex.pcur += --n; + } while ((c = nextc(p)) != -1 && parser_is_identchar(p)); + pushback(p, c); break; } - tokfix(); - len = lex_p - lex_pbeg; - lex_p = lex_pend; - lex_strterm = rb_node_newnode(NODE_HEREDOC, - rb_str_new(tok(), toklen()), /* nd_lit */ - len, /* nd_nth */ - lex_lastline); /* nd_orig */ - return term == '`' ? tXSTRING_BEG : tSTRING_BEG; + len = p->lex.pcur - (p->lex.pbeg + offset) - quote; + if ((unsigned long)len >= HERETERM_LENGTH_MAX) + yyerror0("too long here document identifier"); + dispatch_scan_event(p, tHEREDOC_BEG); + lex_goto_eol(p); + + p->lex.strterm = new_strterm(0, 0, 0, p->lex.lastline); + p->lex.strterm->flags |= STRTERM_HEREDOC; + rb_strterm_heredoc_t *here = &p->lex.strterm->u.heredoc; + here->offset = offset; + here->sourceline = p->ruby_sourceline; + here->length = (int)len; + here->quote = quote; + here->func = func; + + token_flush(p); + p->heredoc_indent = indent; + p->heredoc_line_indent = 0; + return token; } static void -heredoc_restore(here) - NODE *here; +heredoc_restore(struct parser_params *p, rb_strterm_heredoc_t *here) { - VALUE line = here->nd_orig; - lex_lastline = line; - lex_pbeg = RSTRING(line)->ptr; - lex_pend = lex_pbeg + RSTRING(line)->len; - lex_p = lex_pbeg + here->nd_nth; - heredoc_end = ruby_sourceline; - ruby_sourceline = nd_line(here); - dispose_string(here->nd_lit); - rb_gc_force_recycle((VALUE)here); + VALUE line; + + p->lex.strterm = 0; + line = here->lastline; + p->lex.lastline = line; + p->lex.pbeg = RSTRING_PTR(line); + p->lex.pend = p->lex.pbeg + RSTRING_LEN(line); + p->lex.pcur = p->lex.pbeg + here->offset + here->length + here->quote; + p->lex.ptok = p->lex.pbeg + here->offset - here->quote; + p->heredoc_end = p->ruby_sourceline; + p->ruby_sourceline = (int)here->sourceline; + if (p->eofp) p->lex.nextline = Qnil; + p->eofp = 0; } static int -whole_match_p(eos, len, indent) - char *eos; - int len, indent; +dedent_string(VALUE string, int width) { - char *p = lex_pbeg; - int n; + char *str; + long len; + int i, col = 0; + + RSTRING_GETMEM(string, str, len); + for (i = 0; i < len && col < width; i++) { + if (str[i] == ' ') { + col++; + } + else if (str[i] == '\t') { + int n = TAB_WIDTH * (col / TAB_WIDTH + 1); + if (n > width) break; + col = n; + } + else { + break; + } + } + if (!i) return 0; + rb_str_modify(string); + str = RSTRING_PTR(string); + if (RSTRING_LEN(string) != len) + rb_fatal("literal string changed: %+"PRIsVALUE, string); + MEMMOVE(str, str + i, char, len - i); + rb_str_set_len(string, len - i); + return i; +} + +#ifndef RIPPER +static NODE * +heredoc_dedent(struct parser_params *p, NODE *root) +{ + NODE *node, *str_node, *prev_node; + int indent = p->heredoc_indent; + VALUE prev_lit = 0; + + if (indent <= 0) return root; + p->heredoc_indent = 0; + if (!root) return root; + + prev_node = node = str_node = root; + if (nd_type_p(root, NODE_LIST)) str_node = root->nd_head; + + while (str_node) { + VALUE lit = str_node->nd_lit; + if (str_node->flags & NODE_FL_NEWLINE) { + dedent_string(lit, indent); + } + if (!prev_lit) { + prev_lit = lit; + } + else if (!literal_concat0(p, prev_lit, lit)) { + return 0; + } + else { + NODE *end = node->nd_end; + node = prev_node->nd_next = node->nd_next; + if (!node) { + if (nd_type_p(prev_node, NODE_DSTR)) + nd_set_type(prev_node, NODE_STR); + break; + } + node->nd_end = end; + goto next_str; + } + + str_node = 0; + while ((node = (prev_node = node)->nd_next) != 0) { + next_str: + if (!nd_type_p(node, NODE_LIST)) break; + if ((str_node = node->nd_head) != 0) { + enum node_type type = nd_type(str_node); + if (type == NODE_STR || type == NODE_DSTR) break; + prev_lit = 0; + str_node = 0; + } + } + } + return root; +} +#else /* RIPPER */ +static VALUE +heredoc_dedent(struct parser_params *p, VALUE array) +{ + int indent = p->heredoc_indent; + + if (indent <= 0) return array; + p->heredoc_indent = 0; + dispatch2(heredoc_dedent, array, INT2NUM(indent)); + return array; +} + +/* + * call-seq: + * Ripper.dedent_string(input, width) -> Integer + * + * USE OF RIPPER LIBRARY ONLY. + * + * Strips up to +width+ leading whitespaces from +input+, + * and returns the stripped column width. + */ +static VALUE +parser_dedent_string(VALUE self, VALUE input, VALUE width) +{ + int wid, col; + + StringValue(input); + wid = NUM2UINT(width); + col = dedent_string(input, wid); + return INT2NUM(col); +} +#endif + +static int +whole_match_p(struct parser_params *p, const char *eos, long len, int indent) +{ + const char *ptr = p->lex.pbeg; + long n; if (indent) { - while (*p && ISSPACE(*p)) p++; + while (*ptr && ISSPACE(*ptr)) ptr++; + } + n = p->lex.pend - (ptr + len); + if (n < 0) return FALSE; + if (n > 0 && ptr[len] != '\n') { + if (ptr[len] != '\r') return FALSE; + if (n <= 1 || ptr[len+1] != '\n') return FALSE; + } + return strncmp(eos, ptr, len) == 0; +} + +static int +word_match_p(struct parser_params *p, const char *word, long len) +{ + if (strncmp(p->lex.pcur, word, len)) return 0; + if (p->lex.pcur + len == p->lex.pend) return 1; + int c = (unsigned char)p->lex.pcur[len]; + if (ISSPACE(c)) return 1; + switch (c) { + case '\0': case '\004': case '\032': return 1; } - n= lex_pend - (p + len); - if (n < 0 || (n > 0 && p[len] != '\n' && p[len] != '\r')) return Qfalse; - if (strncmp(eos, p, len) == 0) return Qtrue; - return Qfalse; + return 0; } +#define NUM_SUFFIX_R (1<<0) +#define NUM_SUFFIX_I (1<<1) +#define NUM_SUFFIX_ALL 3 + static int -here_document(here) - NODE *here; +number_literal_suffix(struct parser_params *p, int mask) +{ + int c, result = 0; + const char *lastp = p->lex.pcur; + + while ((c = nextc(p)) != -1) { + if ((mask & NUM_SUFFIX_I) && c == 'i') { + result |= (mask & NUM_SUFFIX_I); + mask &= ~NUM_SUFFIX_I; + /* r after i, rational of complex is disallowed */ + mask &= ~NUM_SUFFIX_R; + continue; + } + if ((mask & NUM_SUFFIX_R) && c == 'r') { + result |= (mask & NUM_SUFFIX_R); + mask &= ~NUM_SUFFIX_R; + continue; + } + if (!ISASCII(c) || ISALPHA(c) || c == '_') { + p->lex.pcur = lastp; + literal_flush(p, p->lex.pcur); + return 0; + } + pushback(p, c); + break; + } + return result; +} + +static enum yytokentype +set_number_literal(struct parser_params *p, VALUE v, + enum yytokentype type, int suffix) +{ + if (suffix & NUM_SUFFIX_I) { + v = rb_complex_raw(INT2FIX(0), v); + type = tIMAGINARY; + } + set_yylval_literal(v); + SET_LEX_STATE(EXPR_END); + return type; +} + +static enum yytokentype +set_integer_literal(struct parser_params *p, VALUE v, int suffix) +{ + enum yytokentype type = tINTEGER; + if (suffix & NUM_SUFFIX_R) { + v = rb_rational_raw1(v); + type = tRATIONAL; + } + return set_number_literal(p, v, type, suffix); +} + +#ifdef RIPPER +static void +dispatch_heredoc_end(struct parser_params *p) +{ + VALUE str; + if (has_delayed_token(p)) + dispatch_delayed_token(p, tSTRING_CONTENT); + str = STR_NEW(p->lex.ptok, p->lex.pend - p->lex.ptok); + ripper_dispatch1(p, ripper_token2eventid(tHEREDOC_END), str); + lex_goto_eol(p); + token_flush(p); +} + +#else +#define dispatch_heredoc_end(p) ((void)0) +#endif + +static enum yytokentype +here_document(struct parser_params *p, rb_strterm_heredoc_t *here) { int c, func, indent = 0; - char *eos, *p, *pend; + const char *eos, *ptr, *ptr_end; long len; VALUE str = 0; + rb_encoding *enc = p->enc; + rb_encoding *base_enc = 0; + int bol; - eos = RSTRING(here->nd_lit)->ptr; - len = RSTRING(here->nd_lit)->len - 1; - indent = (func = *eos++) & STR_FUNC_INDENT; + eos = RSTRING_PTR(here->lastline) + here->offset; + len = here->length; + indent = (func = here->func) & STR_FUNC_INDENT; - if ((c = nextc()) == -1) { + if ((c = nextc(p)) == -1) { error: - rb_compile_error("can't find string \"%s\" anywhere before EOF", eos); - heredoc_restore(lex_strterm); - lex_strterm = 0; - return 0; +#ifdef RIPPER + if (!has_delayed_token(p)) { + dispatch_scan_event(p, tSTRING_CONTENT); + } + else { + if ((len = p->lex.pcur - p->lex.ptok) > 0) { + if (!(func & STR_FUNC_REGEXP) && rb_enc_asciicompat(enc)) { + int cr = ENC_CODERANGE_UNKNOWN; + rb_str_coderange_scan_restartable(p->lex.ptok, p->lex.pcur, enc, &cr); + if (cr != ENC_CODERANGE_7BIT && + p->enc == rb_usascii_encoding() && + enc != rb_utf8_encoding()) { + enc = rb_ascii8bit_encoding(); + } + } + rb_enc_str_buf_cat(p->delayed.token, p->lex.ptok, len, enc); + } + dispatch_delayed_token(p, tSTRING_CONTENT); + } + lex_goto_eol(p); +#endif + heredoc_restore(p, &p->lex.strterm->u.heredoc); + compile_error(p, "can't find string \"%.*s\" anywhere before EOF", + (int)len, eos); + token_flush(p); + p->lex.strterm = 0; + SET_LEX_STATE(EXPR_END); + return tSTRING_END; } - if (was_bol() && whole_match_p(eos, len, indent)) { - heredoc_restore(lex_strterm); + bol = was_bol(p); + if (!bol) { + /* not beginning of line, cannot be the terminator */ + } + else if (p->heredoc_line_indent == -1) { + /* `heredoc_line_indent == -1` means + * - "after an interpolation in the same line", or + * - "in a continuing line" + */ + p->heredoc_line_indent = 0; + } + else if (whole_match_p(p, eos, len, indent)) { + dispatch_heredoc_end(p); + restore: + heredoc_restore(p, &p->lex.strterm->u.heredoc); + token_flush(p); + p->lex.strterm = 0; + SET_LEX_STATE(EXPR_END); return tSTRING_END; } if (!(func & STR_FUNC_EXPAND)) { do { - p = RSTRING(lex_lastline)->ptr; - pend = lex_pend; - if (pend > p) { - switch (pend[-1]) { + ptr = RSTRING_PTR(p->lex.lastline); + ptr_end = p->lex.pend; + if (ptr_end > ptr) { + switch (ptr_end[-1]) { case '\n': - if (--pend == p || pend[-1] != '\r') { - pend++; + if (--ptr_end == ptr || ptr_end[-1] != '\r') { + ptr_end++; break; } case '\r': - --pend; + --ptr_end; } } + + if (p->heredoc_indent > 0) { + long i = 0; + while (ptr + i < ptr_end && parser_update_heredoc_indent(p, ptr[i])) + i++; + p->heredoc_line_indent = 0; + } + if (str) - rb_str_cat(str, p, pend - p); + rb_str_cat(str, ptr, ptr_end - ptr); else - str = rb_str_new(p, pend - p); - if (pend < lex_pend) rb_str_cat(str, "\n", 1); - lex_p = lex_pend; - if (nextc() == -1) { - if (str) dispose_string(str); + str = STR_NEW(ptr, ptr_end - ptr); + if (ptr_end < p->lex.pend) rb_str_cat(str, "\n", 1); + lex_goto_eol(p); + if (p->heredoc_indent > 0) { + goto flush_str; + } + if (nextc(p) == -1) { + if (str) { + str = 0; + } goto error; } - } while (!whole_match_p(eos, len, indent)); + } while (!whole_match_p(p, eos, len, indent)); } else { - newtok(); + /* int mb = ENC_CODERANGE_7BIT, *mbp = &mb;*/ + newtok(p); if (c == '#') { - switch (c = nextc()) { - case '$': - case '@': - pushback(c); - return tSTRING_DVAR; - case '{': - return tSTRING_DBEG; + int t = parser_peek_variable_name(p); + if (p->heredoc_line_indent != -1) { + if (p->heredoc_indent > p->heredoc_line_indent) { + p->heredoc_indent = p->heredoc_line_indent; + } + p->heredoc_line_indent = -1; } - tokadd('#'); + if (t) return t; + tokadd(p, '#'); + c = nextc(p); } do { - pushback(c); - if ((c = tokadd_string(func, '\n', 0, NULL)) == -1) goto error; + pushback(p, c); + enc = p->enc; + if ((c = tokadd_string(p, func, '\n', 0, NULL, &enc, &base_enc)) == -1) { + if (p->eofp) goto error; + goto restore; + } if (c != '\n') { - yylval.node = NEW_STR(rb_str_new(tok(), toklen())); + if (c == '\\') p->heredoc_line_indent = -1; + flush: + str = STR_NEW3(tok(p), toklen(p), enc, func); + flush_str: + set_yylval_str(str); +#ifndef RIPPER + if (bol) yylval.node->flags |= NODE_FL_NEWLINE; +#endif + flush_string_content(p, enc); return tSTRING_CONTENT; } - tokadd(nextc()); - if ((c = nextc()) == -1) goto error; - } while (!whole_match_p(eos, len, indent)); - str = rb_str_new(tok(), toklen()); - } - heredoc_restore(lex_strterm); - lex_strterm = NEW_STRTERM(-1, 0, 0); - yylval.node = NEW_STR(str); + tokadd(p, nextc(p)); + if (p->heredoc_indent > 0) { + lex_goto_eol(p); + goto flush; + } + /* if (mbp && mb == ENC_CODERANGE_UNKNOWN) mbp = 0;*/ + if ((c = nextc(p)) == -1) goto error; + } while (!whole_match_p(p, eos, len, indent)); + str = STR_NEW3(tok(p), toklen(p), enc, func); + } + dispatch_heredoc_end(p); +#ifdef RIPPER + str = ripper_new_yylval(p, ripper_token2eventid(tSTRING_CONTENT), + yylval.val, str); +#endif + heredoc_restore(p, &p->lex.strterm->u.heredoc); + token_flush(p); + p->lex.strterm = NEW_STRTERM(func | STR_FUNC_TERM, 0, 0); + set_yylval_str(str); +#ifndef RIPPER + if (bol) yylval.node->flags |= NODE_FL_NEWLINE; +#endif return tSTRING_CONTENT; } #include "lex.c" +static int +arg_ambiguous(struct parser_params *p, char c) +{ +#ifndef RIPPER + if (c == '/') { + rb_warning1("ambiguity between regexp and two divisions: wrap regexp in parentheses or add a space after `%c' operator", WARN_I(c)); + } + else { + rb_warning1("ambiguous first argument; put parentheses or a space even after `%c' operator", WARN_I(c)); + } +#else + dispatch1(arg_ambiguous, rb_usascii_str_new(&c, 1)); +#endif + return TRUE; +} + +static ID +#ifndef RIPPER +formal_argument(struct parser_params *p, ID lhs) +#else +formal_argument(struct parser_params *p, VALUE lhs) +#endif +{ + ID id = get_id(lhs); + + switch (id_type(id)) { + case ID_LOCAL: + break; +#ifndef RIPPER +# define ERR(mesg) yyerror0(mesg) +#else +# define ERR(mesg) (dispatch2(param_error, WARN_S(mesg), lhs), ripper_error(p)) +#endif + case ID_CONST: + ERR("formal argument cannot be a constant"); + return 0; + case ID_INSTANCE: + ERR("formal argument cannot be an instance variable"); + return 0; + case ID_GLOBAL: + ERR("formal argument cannot be a global variable"); + return 0; + case ID_CLASS: + ERR("formal argument cannot be a class variable"); + return 0; + default: + ERR("formal argument must be local variable"); + return 0; +#undef ERR + } + shadowing_lvar(p, id); + return lhs; +} + +static int +lvar_defined(struct parser_params *p, ID id) +{ + return (dyna_in_block(p) && dvar_defined(p, id)) || local_id(p, id); +} + +/* emacsen -*- hack */ +static long +parser_encode_length(struct parser_params *p, const char *name, long len) +{ + long nlen; + + if (len > 5 && name[nlen = len - 5] == '-') { + if (rb_memcicmp(name + nlen + 1, "unix", 4) == 0) + return nlen; + } + if (len > 4 && name[nlen = len - 4] == '-') { + if (rb_memcicmp(name + nlen + 1, "dos", 3) == 0) + return nlen; + if (rb_memcicmp(name + nlen + 1, "mac", 3) == 0 && + !(len == 8 && rb_memcicmp(name, "utf8-mac", len) == 0)) + /* exclude UTF8-MAC because the encoding named "UTF8" doesn't exist in Ruby */ + return nlen; + } + return len; +} + static void -arg_ambiguous() +parser_set_encode(struct parser_params *p, const char *name) { - rb_warning("ambiguous first argument; put parentheses or even spaces"); + int idx = rb_enc_find_index(name); + rb_encoding *enc; + VALUE excargs[3]; + + if (idx < 0) { + excargs[1] = rb_sprintf("unknown encoding name: %s", name); + error: + excargs[0] = rb_eArgError; + excargs[2] = rb_make_backtrace(); + rb_ary_unshift(excargs[2], rb_sprintf("%"PRIsVALUE":%d", p->ruby_sourcefile_string, p->ruby_sourceline)); + rb_exc_raise(rb_make_exception(3, excargs)); + } + enc = rb_enc_from_index(idx); + if (!rb_enc_asciicompat(enc)) { + excargs[1] = rb_sprintf("%s is not ASCII compatible", rb_enc_name(enc)); + goto error; + } + p->enc = enc; +#ifndef RIPPER + if (p->debug_lines) { + VALUE lines = p->debug_lines; + long i, n = RARRAY_LEN(lines); + for (i = 0; i < n; ++i) { + rb_enc_associate_index(RARRAY_AREF(lines, i), idx); + } + } +#endif +} + +static int +comment_at_top(struct parser_params *p) +{ + const char *ptr = p->lex.pbeg, *ptr_end = p->lex.pcur - 1; + if (p->line_count != (p->has_shebang ? 2 : 1)) return 0; + while (ptr < ptr_end) { + if (!ISSPACE(*ptr)) return 0; + ptr++; + } + return 1; +} + +typedef long (*rb_magic_comment_length_t)(struct parser_params *p, const char *name, long len); +typedef void (*rb_magic_comment_setter_t)(struct parser_params *p, const char *name, const char *val); + +static int parser_invalid_pragma_value(struct parser_params *p, const char *name, const char *val); + +static void +magic_comment_encoding(struct parser_params *p, const char *name, const char *val) +{ + if (!comment_at_top(p)) { + return; + } + parser_set_encode(p, val); } -#define IS_ARG() (lex_state == EXPR_ARG || lex_state == EXPR_CMDARG) +static int +parser_get_bool(struct parser_params *p, const char *name, const char *val) +{ + switch (*val) { + case 't': case 'T': + if (STRCASECMP(val, "true") == 0) { + return TRUE; + } + break; + case 'f': case 'F': + if (STRCASECMP(val, "false") == 0) { + return FALSE; + } + break; + } + return parser_invalid_pragma_value(p, name, val); +} static int -yylex() +parser_invalid_pragma_value(struct parser_params *p, const char *name, const char *val) +{ + rb_warning2("invalid value for %s: %s", WARN_S(name), WARN_S(val)); + return -1; +} + +static void +parser_set_token_info(struct parser_params *p, const char *name, const char *val) { + int b = parser_get_bool(p, name, val); + if (b >= 0) p->token_info_enabled = b; +} + +static void +parser_set_compile_option_flag(struct parser_params *p, const char *name, const char *val) +{ + int b; + + if (p->token_seen) { + rb_warning1("`%s' is ignored after any tokens", WARN_S(name)); + return; + } + + b = parser_get_bool(p, name, val); + if (b < 0) return; + + if (!p->compile_option) + p->compile_option = rb_obj_hide(rb_ident_hash_new()); + rb_hash_aset(p->compile_option, ID2SYM(rb_intern(name)), + RBOOL(b)); +} + +static void +parser_set_shareable_constant_value(struct parser_params *p, const char *name, const char *val) +{ + for (const char *s = p->lex.pbeg, *e = p->lex.pcur; s < e; ++s) { + if (*s == ' ' || *s == '\t') continue; + if (*s == '#') break; + rb_warning1("`%s' is ignored unless in comment-only line", WARN_S(name)); + return; + } + + switch (*val) { + case 'n': case 'N': + if (STRCASECMP(val, "none") == 0) { + p->ctxt.shareable_constant_value = shareable_none; + return; + } + break; + case 'l': case 'L': + if (STRCASECMP(val, "literal") == 0) { + p->ctxt.shareable_constant_value = shareable_literal; + return; + } + break; + case 'e': case 'E': + if (STRCASECMP(val, "experimental_copy") == 0) { + p->ctxt.shareable_constant_value = shareable_copy; + return; + } + if (STRCASECMP(val, "experimental_everything") == 0) { + p->ctxt.shareable_constant_value = shareable_everything; + return; + } + break; + } + parser_invalid_pragma_value(p, name, val); +} + +# if WARN_PAST_SCOPE +static void +parser_set_past_scope(struct parser_params *p, const char *name, const char *val) +{ + int b = parser_get_bool(p, name, val); + if (b >= 0) p->past_scope_enabled = b; +} +# endif + +struct magic_comment { + const char *name; + rb_magic_comment_setter_t func; + rb_magic_comment_length_t length; +}; + +static const struct magic_comment magic_comments[] = { + {"coding", magic_comment_encoding, parser_encode_length}, + {"encoding", magic_comment_encoding, parser_encode_length}, + {"frozen_string_literal", parser_set_compile_option_flag}, + {"shareable_constant_value", parser_set_shareable_constant_value}, + {"warn_indent", parser_set_token_info}, +# if WARN_PAST_SCOPE + {"warn_past_scope", parser_set_past_scope}, +# endif +}; + +static const char * +magic_comment_marker(const char *str, long len) +{ + long i = 2; + + while (i < len) { + switch (str[i]) { + case '-': + if (str[i-1] == '*' && str[i-2] == '-') { + return str + i + 1; + } + i += 2; + break; + case '*': + if (i + 1 >= len) return 0; + if (str[i+1] != '-') { + i += 4; + } + else if (str[i-1] != '-') { + i += 2; + } + else { + return str + i + 2; + } + break; + default: + i += 3; + break; + } + } + return 0; +} + +static int +parser_magic_comment(struct parser_params *p, const char *str, long len) +{ + int indicator = 0; + VALUE name = 0, val = 0; + const char *beg, *end, *vbeg, *vend; +#define str_copy(_s, _p, _n) ((_s) \ + ? (void)(rb_str_resize((_s), (_n)), \ + MEMCPY(RSTRING_PTR(_s), (_p), char, (_n)), (_s)) \ + : (void)((_s) = STR_NEW((_p), (_n)))) + + if (len <= 7) return FALSE; + if (!!(beg = magic_comment_marker(str, len))) { + if (!(end = magic_comment_marker(beg, str + len - beg))) + return FALSE; + indicator = TRUE; + str = beg; + len = end - beg - 3; + } + + /* %r"([^\\s\'\":;]+)\\s*:\\s*(\"(?:\\\\.|[^\"])*\"|[^\"\\s;]+)[\\s;]*" */ + while (len > 0) { + const struct magic_comment *mc = magic_comments; + char *s; + int i; + long n = 0; + + for (; len > 0 && *str; str++, --len) { + switch (*str) { + case '\'': case '"': case ':': case ';': + continue; + } + if (!ISSPACE(*str)) break; + } + for (beg = str; len > 0; str++, --len) { + switch (*str) { + case '\'': case '"': case ':': case ';': + break; + default: + if (ISSPACE(*str)) break; + continue; + } + break; + } + for (end = str; len > 0 && ISSPACE(*str); str++, --len); + if (!len) break; + if (*str != ':') { + if (!indicator) return FALSE; + continue; + } + + do str++; while (--len > 0 && ISSPACE(*str)); + if (!len) break; + if (*str == '"') { + for (vbeg = ++str; --len > 0 && *str != '"'; str++) { + if (*str == '\\') { + --len; + ++str; + } + } + vend = str; + if (len) { + --len; + ++str; + } + } + else { + for (vbeg = str; len > 0 && *str != '"' && *str != ';' && !ISSPACE(*str); --len, str++); + vend = str; + } + if (indicator) { + while (len > 0 && (*str == ';' || ISSPACE(*str))) --len, str++; + } + else { + while (len > 0 && (ISSPACE(*str))) --len, str++; + if (len) return FALSE; + } + + n = end - beg; + str_copy(name, beg, n); + s = RSTRING_PTR(name); + for (i = 0; i < n; ++i) { + if (s[i] == '-') s[i] = '_'; + } + do { + if (STRNCASECMP(mc->name, s, n) == 0 && !mc->name[n]) { + n = vend - vbeg; + if (mc->length) { + n = (*mc->length)(p, vbeg, n); + } + str_copy(val, vbeg, n); + (*mc->func)(p, mc->name, RSTRING_PTR(val)); + break; + } + } while (++mc < magic_comments + numberof(magic_comments)); +#ifdef RIPPER + str_copy(val, vbeg, vend - vbeg); + dispatch2(magic_comment, name, val); +#endif + } + + return TRUE; +} + +static void +set_file_encoding(struct parser_params *p, const char *str, const char *send) +{ + int sep = 0; + const char *beg = str; + VALUE s; + + for (;;) { + if (send - str <= 6) return; + switch (str[6]) { + case 'C': case 'c': str += 6; continue; + case 'O': case 'o': str += 5; continue; + case 'D': case 'd': str += 4; continue; + case 'I': case 'i': str += 3; continue; + case 'N': case 'n': str += 2; continue; + case 'G': case 'g': str += 1; continue; + case '=': case ':': + sep = 1; + str += 6; + break; + default: + str += 6; + if (ISSPACE(*str)) break; + continue; + } + if (STRNCASECMP(str-6, "coding", 6) == 0) break; + sep = 0; + } + for (;;) { + do { + if (++str >= send) return; + } while (ISSPACE(*str)); + if (sep) break; + if (*str != '=' && *str != ':') return; + sep = 1; + str++; + } + beg = str; + while ((*str == '-' || *str == '_' || ISALNUM(*str)) && ++str < send); + s = rb_str_new(beg, parser_encode_length(p, beg, str - beg)); + parser_set_encode(p, RSTRING_PTR(s)); + rb_str_resize(s, 0); +} + +static void +parser_prepare(struct parser_params *p) +{ + int c = nextc0(p, FALSE); + p->token_info_enabled = !compile_for_eval && RTEST(ruby_verbose); + switch (c) { + case '#': + if (peek(p, '!')) p->has_shebang = 1; + break; + case 0xef: /* UTF-8 BOM marker */ + if (p->lex.pend - p->lex.pcur >= 2 && + (unsigned char)p->lex.pcur[0] == 0xbb && + (unsigned char)p->lex.pcur[1] == 0xbf) { + p->enc = rb_utf8_encoding(); + p->lex.pcur += 2; +#ifndef RIPPER + if (p->debug_lines) { + rb_enc_associate(p->lex.lastline, p->enc); + } +#endif + p->lex.pbeg = p->lex.pcur; + return; + } + break; + case EOF: + return; + } + pushback(p, c); + p->enc = rb_enc_get(p->lex.lastline); +} + +#ifndef RIPPER +#define ambiguous_operator(tok, op, syn) ( \ + rb_warning0("`"op"' after local variable or literal is interpreted as binary operator"), \ + rb_warning0("even though it seems like "syn"")) +#else +#define ambiguous_operator(tok, op, syn) \ + dispatch2(operator_ambiguous, TOKEN2VAL(tok), rb_str_new_cstr(syn)) +#endif +#define warn_balanced(tok, op, syn) ((void) \ + (!IS_lex_state_for(last_state, EXPR_CLASS|EXPR_DOT|EXPR_FNAME|EXPR_ENDFN) && \ + space_seen && !ISSPACE(c) && \ + (ambiguous_operator(tok, op, syn), 0)), \ + (enum yytokentype)(tok)) + +static VALUE +parse_rational(struct parser_params *p, char *str, int len, int seen_point) +{ + VALUE v; + char *point = &str[seen_point]; + size_t fraclen = len-seen_point-1; + memmove(point, point+1, fraclen+1); + v = rb_cstr_to_inum(str, 10, FALSE); + return rb_rational_new(v, rb_int_positive_pow(10, fraclen)); +} + +static enum yytokentype +no_digits(struct parser_params *p) +{ + yyerror0("numeric literal without digits"); + if (peek(p, '_')) nextc(p); + /* dummy 0, for tUMINUS_NUM at numeric */ + return set_integer_literal(p, INT2FIX(0), 0); +} + +static enum yytokentype +parse_numeric(struct parser_params *p, int c) +{ + int is_float, seen_point, seen_e, nondigit; + int suffix; + + is_float = seen_point = seen_e = nondigit = 0; + SET_LEX_STATE(EXPR_END); + newtok(p); + if (c == '-' || c == '+') { + tokadd(p, c); + c = nextc(p); + } + if (c == '0') { + int start = toklen(p); + c = nextc(p); + if (c == 'x' || c == 'X') { + /* hexadecimal */ + c = nextc(p); + if (c != -1 && ISXDIGIT(c)) { + do { + if (c == '_') { + if (nondigit) break; + nondigit = c; + continue; + } + if (!ISXDIGIT(c)) break; + nondigit = 0; + tokadd(p, c); + } while ((c = nextc(p)) != -1); + } + pushback(p, c); + tokfix(p); + if (toklen(p) == start) { + return no_digits(p); + } + else if (nondigit) goto trailing_uc; + suffix = number_literal_suffix(p, NUM_SUFFIX_ALL); + return set_integer_literal(p, rb_cstr_to_inum(tok(p), 16, FALSE), suffix); + } + if (c == 'b' || c == 'B') { + /* binary */ + c = nextc(p); + if (c == '0' || c == '1') { + do { + if (c == '_') { + if (nondigit) break; + nondigit = c; + continue; + } + if (c != '0' && c != '1') break; + nondigit = 0; + tokadd(p, c); + } while ((c = nextc(p)) != -1); + } + pushback(p, c); + tokfix(p); + if (toklen(p) == start) { + return no_digits(p); + } + else if (nondigit) goto trailing_uc; + suffix = number_literal_suffix(p, NUM_SUFFIX_ALL); + return set_integer_literal(p, rb_cstr_to_inum(tok(p), 2, FALSE), suffix); + } + if (c == 'd' || c == 'D') { + /* decimal */ + c = nextc(p); + if (c != -1 && ISDIGIT(c)) { + do { + if (c == '_') { + if (nondigit) break; + nondigit = c; + continue; + } + if (!ISDIGIT(c)) break; + nondigit = 0; + tokadd(p, c); + } while ((c = nextc(p)) != -1); + } + pushback(p, c); + tokfix(p); + if (toklen(p) == start) { + return no_digits(p); + } + else if (nondigit) goto trailing_uc; + suffix = number_literal_suffix(p, NUM_SUFFIX_ALL); + return set_integer_literal(p, rb_cstr_to_inum(tok(p), 10, FALSE), suffix); + } + if (c == '_') { + /* 0_0 */ + goto octal_number; + } + if (c == 'o' || c == 'O') { + /* prefixed octal */ + c = nextc(p); + if (c == -1 || c == '_' || !ISDIGIT(c)) { + return no_digits(p); + } + } + if (c >= '0' && c <= '7') { + /* octal */ + octal_number: + do { + if (c == '_') { + if (nondigit) break; + nondigit = c; + continue; + } + if (c < '0' || c > '9') break; + if (c > '7') goto invalid_octal; + nondigit = 0; + tokadd(p, c); + } while ((c = nextc(p)) != -1); + if (toklen(p) > start) { + pushback(p, c); + tokfix(p); + if (nondigit) goto trailing_uc; + suffix = number_literal_suffix(p, NUM_SUFFIX_ALL); + return set_integer_literal(p, rb_cstr_to_inum(tok(p), 8, FALSE), suffix); + } + if (nondigit) { + pushback(p, c); + goto trailing_uc; + } + } + if (c > '7' && c <= '9') { + invalid_octal: + yyerror0("Invalid octal digit"); + } + else if (c == '.' || c == 'e' || c == 'E') { + tokadd(p, '0'); + } + else { + pushback(p, c); + suffix = number_literal_suffix(p, NUM_SUFFIX_ALL); + return set_integer_literal(p, INT2FIX(0), suffix); + } + } + + for (;;) { + switch (c) { + case '0': case '1': case '2': case '3': case '4': + case '5': case '6': case '7': case '8': case '9': + nondigit = 0; + tokadd(p, c); + break; + + case '.': + if (nondigit) goto trailing_uc; + if (seen_point || seen_e) { + goto decode_num; + } + else { + int c0 = nextc(p); + if (c0 == -1 || !ISDIGIT(c0)) { + pushback(p, c0); + goto decode_num; + } + c = c0; + } + seen_point = toklen(p); + tokadd(p, '.'); + tokadd(p, c); + is_float++; + nondigit = 0; + break; + + case 'e': + case 'E': + if (nondigit) { + pushback(p, c); + c = nondigit; + goto decode_num; + } + if (seen_e) { + goto decode_num; + } + nondigit = c; + c = nextc(p); + if (c != '-' && c != '+' && !ISDIGIT(c)) { + pushback(p, c); + nondigit = 0; + goto decode_num; + } + tokadd(p, nondigit); + seen_e++; + is_float++; + tokadd(p, c); + nondigit = (c == '-' || c == '+') ? c : 0; + break; + + case '_': /* `_' in number just ignored */ + if (nondigit) goto decode_num; + nondigit = c; + break; + + default: + goto decode_num; + } + c = nextc(p); + } + + decode_num: + pushback(p, c); + if (nondigit) { + trailing_uc: + literal_flush(p, p->lex.pcur - 1); + YYLTYPE loc = RUBY_INIT_YYLLOC(); + compile_error(p, "trailing `%c' in number", nondigit); + parser_show_error_line(p, &loc); + } + tokfix(p); + if (is_float) { + enum yytokentype type = tFLOAT; + VALUE v; + + suffix = number_literal_suffix(p, seen_e ? NUM_SUFFIX_I : NUM_SUFFIX_ALL); + if (suffix & NUM_SUFFIX_R) { + type = tRATIONAL; + v = parse_rational(p, tok(p), toklen(p), seen_point); + } + else { + double d = strtod(tok(p), 0); + if (errno == ERANGE) { + rb_warning1("Float %s out of range", WARN_S(tok(p))); + errno = 0; + } + v = DBL2NUM(d); + } + return set_number_literal(p, v, type, suffix); + } + suffix = number_literal_suffix(p, NUM_SUFFIX_ALL); + return set_integer_literal(p, rb_cstr_to_inum(tok(p), 10, FALSE), suffix); +} + +static enum yytokentype +parse_qmark(struct parser_params *p, int space_seen) +{ + rb_encoding *enc; register int c; - int space_seen = 0; - int cmd_state; - enum lex_state last_state; + VALUE lit; - if (lex_strterm) { - int token; - if (nd_type(lex_strterm) == NODE_HEREDOC) { - token = here_document(lex_strterm); - if (token == tSTRING_END) { - lex_strterm = 0; - lex_state = EXPR_END; + if (IS_END()) { + SET_LEX_STATE(EXPR_VALUE); + return '?'; + } + c = nextc(p); + if (c == -1) { + compile_error(p, "incomplete character syntax"); + return 0; + } + if (rb_enc_isspace(c, p->enc)) { + if (!IS_ARG()) { + int c2 = escaped_control_code(c); + if (c2) { + WARN_SPACE_CHAR(c2, "?"); } } + ternary: + pushback(p, c); + SET_LEX_STATE(EXPR_VALUE); + return '?'; + } + newtok(p); + enc = p->enc; + if (!parser_isascii(p)) { + if (tokadd_mbchar(p, c) == -1) return 0; + } + else if ((rb_enc_isalnum(c, p->enc) || c == '_') && + p->lex.pcur < p->lex.pend && is_identchar(p->lex.pcur, p->lex.pend, p->enc)) { + if (space_seen) { + const char *start = p->lex.pcur - 1, *ptr = start; + do { + int n = parser_precise_mbclen(p, ptr); + if (n < 0) return -1; + ptr += n; + } while (ptr < p->lex.pend && is_identchar(ptr, p->lex.pend, p->enc)); + rb_warn2("`?' just followed by `%.*s' is interpreted as" \ + " a conditional operator, put a space after `?'", + WARN_I((int)(ptr - start)), WARN_S_L(start, (ptr - start))); + } + goto ternary; + } + else if (c == '\\') { + if (peek(p, 'u')) { + nextc(p); + enc = rb_utf8_encoding(); + tokadd_utf8(p, &enc, -1, 0, 0); + } + else if (!lex_eol_p(p) && !(c = *p->lex.pcur, ISASCII(c))) { + nextc(p); + if (tokadd_mbchar(p, c) == -1) return 0; + } + else { + c = read_escape(p, 0, &enc); + tokadd(p, c); + } + } + else { + tokadd(p, c); + } + tokfix(p); + lit = STR_NEW3(tok(p), toklen(p), enc, 0); + set_yylval_str(lit); + SET_LEX_STATE(EXPR_END); + return tCHAR; +} + +static enum yytokentype +parse_percent(struct parser_params *p, const int space_seen, const enum lex_state_e last_state) +{ + register int c; + const char *ptok = p->lex.pcur; + + if (IS_BEG()) { + int term; + int paren; + + c = nextc(p); + quotation: + if (c == -1) goto unterminated; + if (!ISALNUM(c)) { + term = c; + if (!ISASCII(c)) goto unknown; + c = 'Q'; + } + else { + term = nextc(p); + if (rb_enc_isalnum(term, p->enc) || !parser_isascii(p)) { + unknown: + pushback(p, term); + c = parser_precise_mbclen(p, p->lex.pcur); + if (c < 0) return 0; + p->lex.pcur += c; + yyerror0("unknown type of %string"); + return 0; + } + } + if (term == -1) { + unterminated: + compile_error(p, "unterminated quoted string meets end of file"); + return 0; + } + paren = term; + if (term == '(') term = ')'; + else if (term == '[') term = ']'; + else if (term == '{') term = '}'; + else if (term == '<') term = '>'; + else paren = 0; + + p->lex.ptok = ptok-1; + switch (c) { + case 'Q': + p->lex.strterm = NEW_STRTERM(str_dquote, term, paren); + return tSTRING_BEG; + + case 'q': + p->lex.strterm = NEW_STRTERM(str_squote, term, paren); + return tSTRING_BEG; + + case 'W': + p->lex.strterm = NEW_STRTERM(str_dword, term, paren); + return tWORDS_BEG; + + case 'w': + p->lex.strterm = NEW_STRTERM(str_sword, term, paren); + return tQWORDS_BEG; + + case 'I': + p->lex.strterm = NEW_STRTERM(str_dword, term, paren); + return tSYMBOLS_BEG; + + case 'i': + p->lex.strterm = NEW_STRTERM(str_sword, term, paren); + return tQSYMBOLS_BEG; + + case 'x': + p->lex.strterm = NEW_STRTERM(str_xquote, term, paren); + return tXSTRING_BEG; + + case 'r': + p->lex.strterm = NEW_STRTERM(str_regexp, term, paren); + return tREGEXP_BEG; + + case 's': + p->lex.strterm = NEW_STRTERM(str_ssym, term, paren); + SET_LEX_STATE(EXPR_FNAME|EXPR_FITEM); + return tSYMBEG; + + default: + yyerror0("unknown type of %string"); + return 0; + } + } + if ((c = nextc(p)) == '=') { + set_yylval_id('%'); + SET_LEX_STATE(EXPR_BEG); + return tOP_ASGN; + } + if (IS_SPCARG(c) || (IS_lex_state(EXPR_FITEM) && c == 's')) { + goto quotation; + } + SET_LEX_STATE(IS_AFTER_OPERATOR() ? EXPR_ARG : EXPR_BEG); + pushback(p, c); + return warn_balanced('%', "%%", "string literal"); +} + +static int +tokadd_ident(struct parser_params *p, int c) +{ + do { + if (tokadd_mbchar(p, c) == -1) return -1; + c = nextc(p); + } while (parser_is_identchar(p)); + pushback(p, c); + return 0; +} + +static ID +tokenize_ident(struct parser_params *p, const enum lex_state_e last_state) +{ + ID ident = TOK_INTERN(); + + set_yylval_name(ident); + + return ident; +} + +static int +parse_numvar(struct parser_params *p) +{ + size_t len; + int overflow; + unsigned long n = ruby_scan_digits(tok(p)+1, toklen(p)-1, 10, &len, &overflow); + const unsigned long nth_ref_max = + ((FIXNUM_MAX < INT_MAX) ? FIXNUM_MAX : INT_MAX) >> 1; + /* NTH_REF is left-shifted to be ORed with back-ref flag and + * turned into a Fixnum, in compile.c */ + + if (overflow || n > nth_ref_max) { + /* compile_error()? */ + rb_warn1("`%s' is too big for a number variable, always nil", WARN_S(tok(p))); + return 0; /* $0 is $PROGRAM_NAME, not NTH_REF */ + } + else { + return (int)n; + } +} + +static enum yytokentype +parse_gvar(struct parser_params *p, const enum lex_state_e last_state) +{ + const char *ptr = p->lex.pcur; + register int c; + + SET_LEX_STATE(EXPR_END); + p->lex.ptok = ptr - 1; /* from '$' */ + newtok(p); + c = nextc(p); + switch (c) { + case '_': /* $_: last read line string */ + c = nextc(p); + if (parser_is_identchar(p)) { + tokadd(p, '$'); + tokadd(p, '_'); + break; + } + pushback(p, c); + c = '_'; + /* fall through */ + case '~': /* $~: match-data */ + case '*': /* $*: argv */ + case '$': /* $$: pid */ + case '?': /* $?: last status */ + case '!': /* $!: error string */ + case '@': /* $@: error position */ + case '/': /* $/: input record separator */ + case '\\': /* $\: output record separator */ + case ';': /* $;: field separator */ + case ',': /* $,: output field separator */ + case '.': /* $.: last read line number */ + case '=': /* $=: ignorecase */ + case ':': /* $:: load path */ + case '<': /* $<: reading filename */ + case '>': /* $>: default output handle */ + case '\"': /* $": already loaded files */ + tokadd(p, '$'); + tokadd(p, c); + goto gvar; + + case '-': + tokadd(p, '$'); + tokadd(p, c); + c = nextc(p); + if (parser_is_identchar(p)) { + if (tokadd_mbchar(p, c) == -1) return 0; + } + else { + pushback(p, c); + pushback(p, '-'); + return '$'; + } + gvar: + set_yylval_name(TOK_INTERN()); + return tGVAR; + + case '&': /* $&: last match */ + case '`': /* $`: string before last match */ + case '\'': /* $': string after last match */ + case '+': /* $+: string matches last paren. */ + if (IS_lex_state_for(last_state, EXPR_FNAME)) { + tokadd(p, '$'); + tokadd(p, c); + goto gvar; + } + set_yylval_node(NEW_BACK_REF(c, &_cur_loc)); + return tBACK_REF; + + case '1': case '2': case '3': + case '4': case '5': case '6': + case '7': case '8': case '9': + tokadd(p, '$'); + do { + tokadd(p, c); + c = nextc(p); + } while (c != -1 && ISDIGIT(c)); + pushback(p, c); + if (IS_lex_state_for(last_state, EXPR_FNAME)) goto gvar; + tokfix(p); + c = parse_numvar(p); + set_yylval_node(NEW_NTH_REF(c, &_cur_loc)); + return tNTH_REF; + + default: + if (!parser_is_identchar(p)) { + YYLTYPE loc = RUBY_INIT_YYLLOC(); + if (c == -1 || ISSPACE(c)) { + compile_error(p, "`$' without identifiers is not allowed as a global variable name"); + } + else { + pushback(p, c); + compile_error(p, "`$%c' is not allowed as a global variable name", c); + } + parser_show_error_line(p, &loc); + set_yylval_noname(); + return tGVAR; + } + /* fall through */ + case '0': + tokadd(p, '$'); + } + + if (tokadd_ident(p, c)) return 0; + SET_LEX_STATE(EXPR_END); + tokenize_ident(p, last_state); + return tGVAR; +} + +#ifndef RIPPER +static bool +parser_numbered_param(struct parser_params *p, int n) +{ + if (n < 0) return false; + + if (DVARS_TERMINAL_P(p->lvtbl->args) || DVARS_TERMINAL_P(p->lvtbl->args->prev)) { + return false; + } + if (p->max_numparam == ORDINAL_PARAM) { + compile_error(p, "ordinary parameter is defined"); + return false; + } + struct vtable *args = p->lvtbl->args; + if (p->max_numparam < n) { + p->max_numparam = n; + } + while (n > args->pos) { + vtable_add(args, NUMPARAM_IDX_TO_ID(args->pos+1)); + } + return true; +} +#endif + +static enum yytokentype +parse_atmark(struct parser_params *p, const enum lex_state_e last_state) +{ + const char *ptr = p->lex.pcur; + enum yytokentype result = tIVAR; + register int c = nextc(p); + YYLTYPE loc; + + p->lex.ptok = ptr - 1; /* from '@' */ + newtok(p); + tokadd(p, '@'); + if (c == '@') { + result = tCVAR; + tokadd(p, '@'); + c = nextc(p); + } + SET_LEX_STATE(IS_lex_state_for(last_state, EXPR_FNAME) ? EXPR_ENDFN : EXPR_END); + if (c == -1 || !parser_is_identchar(p)) { + pushback(p, c); + RUBY_SET_YYLLOC(loc); + if (result == tIVAR) { + compile_error(p, "`@' without identifiers is not allowed as an instance variable name"); + } + else { + compile_error(p, "`@@' without identifiers is not allowed as a class variable name"); + } + parser_show_error_line(p, &loc); + set_yylval_noname(); + SET_LEX_STATE(EXPR_END); + return result; + } + else if (ISDIGIT(c)) { + pushback(p, c); + RUBY_SET_YYLLOC(loc); + if (result == tIVAR) { + compile_error(p, "`@%c' is not allowed as an instance variable name", c); + } else { - token = parse_string(lex_strterm); - if (token == tSTRING_END || token == tREGEXP_END) { - rb_gc_force_recycle((VALUE)lex_strterm); - lex_strterm = 0; - lex_state = EXPR_END; + compile_error(p, "`@@%c' is not allowed as a class variable name", c); + } + parser_show_error_line(p, &loc); + set_yylval_noname(); + SET_LEX_STATE(EXPR_END); + return result; + } + + if (tokadd_ident(p, c)) return 0; + tokenize_ident(p, last_state); + return result; +} + +static enum yytokentype +parse_ident(struct parser_params *p, int c, int cmd_state) +{ + enum yytokentype result; + int mb = ENC_CODERANGE_7BIT; + const enum lex_state_e last_state = p->lex.state; + ID ident; + + do { + if (!ISASCII(c)) mb = ENC_CODERANGE_UNKNOWN; + if (tokadd_mbchar(p, c) == -1) return 0; + c = nextc(p); + } while (parser_is_identchar(p)); + if ((c == '!' || c == '?') && !peek(p, '=')) { + result = tFID; + tokadd(p, c); + } + else if (c == '=' && IS_lex_state(EXPR_FNAME) && + (!peek(p, '~') && !peek(p, '>') && (!peek(p, '=') || (peek_n(p, '>', 1))))) { + result = tIDENTIFIER; + tokadd(p, c); + } + else { + result = tCONSTANT; /* assume provisionally */ + pushback(p, c); + } + tokfix(p); + + if (IS_LABEL_POSSIBLE()) { + if (IS_LABEL_SUFFIX(0)) { + SET_LEX_STATE(EXPR_ARG|EXPR_LABELED); + nextc(p); + set_yylval_name(TOK_INTERN()); + return tLABEL; + } + } + if (mb == ENC_CODERANGE_7BIT && !IS_lex_state(EXPR_DOT)) { + const struct kwtable *kw; + + /* See if it is a reserved word. */ + kw = rb_reserved_word(tok(p), toklen(p)); + if (kw) { + enum lex_state_e state = p->lex.state; + if (IS_lex_state_for(state, EXPR_FNAME)) { + SET_LEX_STATE(EXPR_ENDFN); + set_yylval_name(rb_intern2(tok(p), toklen(p))); + return kw->id[0]; + } + SET_LEX_STATE(kw->state); + if (IS_lex_state(EXPR_BEG)) { + p->command_start = TRUE; + } + if (kw->id[0] == keyword_do) { + if (lambda_beginning_p()) { + p->lex.lpar_beg = -1; /* make lambda_beginning_p() == FALSE in the body of "-> do ... end" */ + return keyword_do_LAMBDA; + } + if (COND_P()) return keyword_do_cond; + if (CMDARG_P() && !IS_lex_state_for(state, EXPR_CMDARG)) + return keyword_do_block; + return keyword_do; + } + if (IS_lex_state_for(state, (EXPR_BEG | EXPR_LABELED))) + return kw->id[0]; + else { + if (kw->id[0] != kw->id[1]) + SET_LEX_STATE(EXPR_BEG | EXPR_LABEL); + return kw->id[1]; } } - return token; } - cmd_state = command_start; - command_start = Qfalse; + + if (IS_lex_state(EXPR_BEG_ANY | EXPR_ARG_ANY | EXPR_DOT)) { + if (cmd_state) { + SET_LEX_STATE(EXPR_CMDARG); + } + else { + SET_LEX_STATE(EXPR_ARG); + } + } + else if (p->lex.state == EXPR_FNAME) { + SET_LEX_STATE(EXPR_ENDFN); + } + else { + SET_LEX_STATE(EXPR_END); + } + + ident = tokenize_ident(p, last_state); + if (result == tCONSTANT && is_local_id(ident)) result = tIDENTIFIER; + if (!IS_lex_state_for(last_state, EXPR_DOT|EXPR_FNAME) && + (result == tIDENTIFIER) && /* not EXPR_FNAME, not attrasgn */ + lvar_defined(p, ident)) { + SET_LEX_STATE(EXPR_END|EXPR_LABEL); + } + return result; +} + +static enum yytokentype +parser_yylex(struct parser_params *p) +{ + register int c; + int space_seen = 0; + int cmd_state; + int label; + enum lex_state_e last_state; + int fallthru = FALSE; + int token_seen = p->token_seen; + + if (p->lex.strterm) { + if (p->lex.strterm->flags & STRTERM_HEREDOC) { + return here_document(p, &p->lex.strterm->u.heredoc); + } + else { + token_flush(p); + return parse_string(p, &p->lex.strterm->u.literal); + } + } + cmd_state = p->command_start; + p->command_start = FALSE; + p->token_seen = TRUE; retry: - switch (c = nextc()) { + last_state = p->lex.state; +#ifndef RIPPER + token_flush(p); +#endif + switch (c = nextc(p)) { case '\0': /* NUL */ case '\004': /* ^D */ case '\032': /* ^Z */ @@ -3388,111 +9284,197 @@ yylex() return 0; /* white spaces */ - case ' ': case '\t': case '\f': case '\r': + case '\r': + if (!p->cr_seen) { + p->cr_seen = TRUE; + /* carried over with p->lex.nextline for nextc() */ + rb_warn0("encountered \\r in middle of line, treated as a mere space"); + } + /* fall through */ + case ' ': case '\t': case '\f': case '\13': /* '\v' */ - space_seen++; + space_seen = 1; +#ifdef RIPPER + while ((c = nextc(p))) { + switch (c) { + case ' ': case '\t': case '\f': case '\r': + case '\13': /* '\v' */ + break; + default: + goto outofloop; + } + } + outofloop: + pushback(p, c); + dispatch_scan_event(p, tSP); +#endif goto retry; case '#': /* it's a comment */ - while ((c = nextc()) != '\n') { - if (c == -1) - return 0; + p->token_seen = token_seen; + /* no magic_comment in shebang line */ + if (!parser_magic_comment(p, p->lex.pcur, p->lex.pend - p->lex.pcur)) { + if (comment_at_top(p)) { + set_file_encoding(p, p->lex.pcur, p->lex.pend); + } } + lex_goto_eol(p); + dispatch_scan_event(p, tCOMMENT); + fallthru = TRUE; /* fall through */ case '\n': - switch (lex_state) { - case EXPR_BEG: - case EXPR_FNAME: - case EXPR_DOT: - case EXPR_CLASS: + p->token_seen = token_seen; + c = (IS_lex_state(EXPR_BEG|EXPR_CLASS|EXPR_FNAME|EXPR_DOT) && + !IS_lex_state(EXPR_LABELED)); + if (c || IS_lex_state_all(EXPR_ARG|EXPR_LABELED)) { + if (!fallthru) { + dispatch_scan_event(p, tIGNORED_NL); + } + fallthru = FALSE; + if (!c && p->ctxt.in_kwarg) { + goto normal_newline; + } goto retry; - default: - break; } - command_start = Qtrue; - lex_state = EXPR_BEG; + while (1) { + switch (c = nextc(p)) { + case ' ': case '\t': case '\f': case '\r': + case '\13': /* '\v' */ + space_seen = 1; + break; + case '#': + pushback(p, c); + if (space_seen) dispatch_scan_event(p, tSP); + goto retry; + case '&': + case '.': { + dispatch_delayed_token(p, tIGNORED_NL); + if (peek(p, '.') == (c == '&')) { + pushback(p, c); + dispatch_scan_event(p, tSP); + goto retry; + } + } + default: + p->ruby_sourceline--; + p->lex.nextline = p->lex.lastline; + case -1: /* EOF no decrement*/ +#ifndef RIPPER + if (p->lex.prevline && !p->eofp) p->lex.lastline = p->lex.prevline; + p->lex.pbeg = RSTRING_PTR(p->lex.lastline); + p->lex.pend = p->lex.pcur = p->lex.pbeg + RSTRING_LEN(p->lex.lastline); + pushback(p, 1); /* always pushback */ + p->lex.ptok = p->lex.pcur; +#else + lex_goto_eol(p); + if (c != -1) { + p->lex.ptok = p->lex.pcur; + } +#endif + goto normal_newline; + } + } + normal_newline: + p->command_start = TRUE; + SET_LEX_STATE(EXPR_BEG); return '\n'; case '*': - if ((c = nextc()) == '*') { - if ((c = nextc()) == '=') { - yylval.id = tPOW; - lex_state = EXPR_BEG; + if ((c = nextc(p)) == '*') { + if ((c = nextc(p)) == '=') { + set_yylval_id(idPow); + SET_LEX_STATE(EXPR_BEG); return tOP_ASGN; } - pushback(c); - c = tPOW; + pushback(p, c); + if (IS_SPCARG(c)) { + rb_warning0("`**' interpreted as argument prefix"); + c = tDSTAR; + } + else if (IS_BEG()) { + c = tDSTAR; + } + else { + c = warn_balanced((enum ruby_method_ids)tPOW, "**", "argument prefix"); + } } else { if (c == '=') { - yylval.id = '*'; - lex_state = EXPR_BEG; + set_yylval_id('*'); + SET_LEX_STATE(EXPR_BEG); return tOP_ASGN; } - pushback(c); - if (IS_ARG() && space_seen && !ISSPACE(c)){ - rb_warning("`*' interpreted as argument prefix"); + pushback(p, c); + if (IS_SPCARG(c)) { + rb_warning0("`*' interpreted as argument prefix"); c = tSTAR; } - else if (lex_state == EXPR_BEG || lex_state == EXPR_MID) { + else if (IS_BEG()) { c = tSTAR; } else { - c = '*'; + c = warn_balanced('*', "*", "argument prefix"); } } - switch (lex_state) { - case EXPR_FNAME: case EXPR_DOT: - lex_state = EXPR_ARG; break; - default: - lex_state = EXPR_BEG; break; - } + SET_LEX_STATE(IS_AFTER_OPERATOR() ? EXPR_ARG : EXPR_BEG); return c; case '!': - lex_state = EXPR_BEG; - if ((c = nextc()) == '=') { + c = nextc(p); + if (IS_AFTER_OPERATOR()) { + SET_LEX_STATE(EXPR_ARG); + if (c == '@') { + return '!'; + } + } + else { + SET_LEX_STATE(EXPR_BEG); + } + if (c == '=') { return tNEQ; } if (c == '~') { return tNMATCH; } - pushback(c); + pushback(p, c); return '!'; case '=': - if (was_bol()) { + if (was_bol(p)) { /* skip embedded rd document */ - if (strncmp(lex_p, "begin", 5) == 0 && ISSPACE(lex_p[5])) { + if (word_match_p(p, "begin", 5)) { + int first_p = TRUE; + + lex_goto_eol(p); + dispatch_scan_event(p, tEMBDOC_BEG); for (;;) { - lex_p = lex_pend; - c = nextc(); + lex_goto_eol(p); + if (!first_p) { + dispatch_scan_event(p, tEMBDOC); + } + first_p = FALSE; + c = nextc(p); if (c == -1) { - rb_compile_error("embedded document meets end of file"); + compile_error(p, "embedded document meets end of file"); return 0; } - if (c != '=') continue; - if (strncmp(lex_p, "end", 3) == 0 && - (lex_p + 3 == lex_pend || ISSPACE(lex_p[3]))) { + if (c == '=' && word_match_p(p, "end", 3)) { break; } + pushback(p, c); } - lex_p = lex_pend; + lex_goto_eol(p); + dispatch_scan_event(p, tEMBDOC_END); goto retry; } } - switch (lex_state) { - case EXPR_FNAME: case EXPR_DOT: - lex_state = EXPR_ARG; break; - default: - lex_state = EXPR_BEG; break; - } - if ((c = nextc()) == '=') { - if ((c = nextc()) == '=') { + SET_LEX_STATE(IS_AFTER_OPERATOR() ? EXPR_ARG : EXPR_BEG); + if ((c = nextc(p)) == '=') { + if ((c = nextc(p)) == '=') { return tEQQ; } - pushback(c); + pushback(p, c); return tEQ; } if (c == '~') { @@ -3501,1101 +9483,591 @@ yylex() else if (c == '>') { return tASSOC; } - pushback(c); + pushback(p, c); return '='; case '<': - c = nextc(); + c = nextc(p); if (c == '<' && - lex_state != EXPR_END && - lex_state != EXPR_DOT && - lex_state != EXPR_ENDARG && - lex_state != EXPR_CLASS && - (!IS_ARG() || space_seen)) { - int token = heredoc_identifier(); - if (token) return token; - } - switch (lex_state) { - case EXPR_FNAME: case EXPR_DOT: - lex_state = EXPR_ARG; break; - default: - lex_state = EXPR_BEG; break; + !IS_lex_state(EXPR_DOT | EXPR_CLASS) && + !IS_END() && + (!IS_ARG() || IS_lex_state(EXPR_LABELED) || space_seen)) { + int token = heredoc_identifier(p); + if (token) return token < 0 ? 0 : token; + } + if (IS_AFTER_OPERATOR()) { + SET_LEX_STATE(EXPR_ARG); + } + else { + if (IS_lex_state(EXPR_CLASS)) + p->command_start = TRUE; + SET_LEX_STATE(EXPR_BEG); } if (c == '=') { - if ((c = nextc()) == '>') { + if ((c = nextc(p)) == '>') { return tCMP; } - pushback(c); + pushback(p, c); return tLEQ; } if (c == '<') { - if ((c = nextc()) == '=') { - yylval.id = tLSHFT; - lex_state = EXPR_BEG; + if ((c = nextc(p)) == '=') { + set_yylval_id(idLTLT); + SET_LEX_STATE(EXPR_BEG); return tOP_ASGN; } - pushback(c); - return tLSHFT; + pushback(p, c); + return warn_balanced((enum ruby_method_ids)tLSHFT, "<<", "here document"); } - pushback(c); + pushback(p, c); return '<'; case '>': - switch (lex_state) { - case EXPR_FNAME: case EXPR_DOT: - lex_state = EXPR_ARG; break; - default: - lex_state = EXPR_BEG; break; - } - if ((c = nextc()) == '=') { + SET_LEX_STATE(IS_AFTER_OPERATOR() ? EXPR_ARG : EXPR_BEG); + if ((c = nextc(p)) == '=') { return tGEQ; } if (c == '>') { - if ((c = nextc()) == '=') { - yylval.id = tRSHFT; - lex_state = EXPR_BEG; + if ((c = nextc(p)) == '=') { + set_yylval_id(idGTGT); + SET_LEX_STATE(EXPR_BEG); return tOP_ASGN; } - pushback(c); + pushback(p, c); return tRSHFT; } - pushback(c); + pushback(p, c); return '>'; case '"': - lex_strterm = NEW_STRTERM(str_dquote, '"', 0); + label = (IS_LABEL_POSSIBLE() ? str_label : 0); + p->lex.strterm = NEW_STRTERM(str_dquote | label, '"', 0); + p->lex.ptok = p->lex.pcur-1; return tSTRING_BEG; case '`': - if (lex_state == EXPR_FNAME) { - lex_state = EXPR_END; + if (IS_lex_state(EXPR_FNAME)) { + SET_LEX_STATE(EXPR_ENDFN); return c; } - if (lex_state == EXPR_DOT) { + if (IS_lex_state(EXPR_DOT)) { if (cmd_state) - lex_state = EXPR_CMDARG; + SET_LEX_STATE(EXPR_CMDARG); else - lex_state = EXPR_ARG; + SET_LEX_STATE(EXPR_ARG); return c; } - lex_strterm = NEW_STRTERM(str_xquote, '`', 0); + p->lex.strterm = NEW_STRTERM(str_xquote, '`', 0); return tXSTRING_BEG; case '\'': - lex_strterm = NEW_STRTERM(str_squote, '\'', 0); + label = (IS_LABEL_POSSIBLE() ? str_label : 0); + p->lex.strterm = NEW_STRTERM(str_squote | label, '\'', 0); + p->lex.ptok = p->lex.pcur-1; return tSTRING_BEG; case '?': - if (lex_state == EXPR_END || lex_state == EXPR_ENDARG) { - lex_state = EXPR_BEG; - return '?'; - } - c = nextc(); - if (c == -1) { - rb_compile_error("incomplete character syntax"); - return 0; - } - if (ISSPACE(c)){ - if (!IS_ARG()){ - int c2 = 0; - switch (c) { - case ' ': - c2 = 's'; - break; - case '\n': - c2 = 'n'; - break; - case '\t': - c2 = 't'; - break; - case '\v': - c2 = 'v'; - break; - case '\r': - c2 = 'r'; - break; - case '\f': - c2 = 'f'; - break; - } - if (c2) { - rb_warn("invalid character syntax; use ?\\%c", c2); - } - } - ternary: - pushback(c); - lex_state = EXPR_BEG; - return '?'; - } - else if (ismbchar(c)) { - rb_warn("multibyte character literal not supported yet; use ?\\%.3o", c); - goto ternary; - } - else if ((ISALNUM(c) || c == '_') && lex_p < lex_pend && is_identchar(*lex_p)) { - goto ternary; - } - else if (c == '\\') { - c = read_escape(); - } - c &= 0xff; - lex_state = EXPR_END; - yylval.node = NEW_LIT(INT2FIX(c)); - return tINTEGER; + return parse_qmark(p, space_seen); case '&': - if ((c = nextc()) == '&') { - lex_state = EXPR_BEG; - if ((c = nextc()) == '=') { - yylval.id = tANDOP; - lex_state = EXPR_BEG; + if ((c = nextc(p)) == '&') { + SET_LEX_STATE(EXPR_BEG); + if ((c = nextc(p)) == '=') { + set_yylval_id(idANDOP); + SET_LEX_STATE(EXPR_BEG); return tOP_ASGN; } - pushback(c); + pushback(p, c); return tANDOP; } else if (c == '=') { - yylval.id = '&'; - lex_state = EXPR_BEG; + set_yylval_id('&'); + SET_LEX_STATE(EXPR_BEG); return tOP_ASGN; } - pushback(c); - if (IS_ARG() && space_seen && !ISSPACE(c)){ - rb_warning("`&' interpreted as argument prefix"); + else if (c == '.') { + set_yylval_id(idANDDOT); + SET_LEX_STATE(EXPR_DOT); + return tANDDOT; + } + pushback(p, c); + if (IS_SPCARG(c)) { + if ((c != ':') || + (c = peekc_n(p, 1)) == -1 || + !(c == '\'' || c == '"' || + is_identchar((p->lex.pcur+1), p->lex.pend, p->enc))) { + rb_warning0("`&' interpreted as argument prefix"); + } c = tAMPER; } - else if (lex_state == EXPR_BEG || lex_state == EXPR_MID) { + else if (IS_BEG()) { c = tAMPER; } else { - c = '&'; - } - switch (lex_state) { - case EXPR_FNAME: case EXPR_DOT: - lex_state = EXPR_ARG; break; - default: - lex_state = EXPR_BEG; + c = warn_balanced('&', "&", "argument prefix"); } + SET_LEX_STATE(IS_AFTER_OPERATOR() ? EXPR_ARG : EXPR_BEG); return c; case '|': - if ((c = nextc()) == '|') { - lex_state = EXPR_BEG; - if ((c = nextc()) == '=') { - yylval.id = tOROP; - lex_state = EXPR_BEG; + if ((c = nextc(p)) == '|') { + SET_LEX_STATE(EXPR_BEG); + if ((c = nextc(p)) == '=') { + set_yylval_id(idOROP); + SET_LEX_STATE(EXPR_BEG); return tOP_ASGN; } - pushback(c); + pushback(p, c); + if (IS_lex_state_for(last_state, EXPR_BEG)) { + c = '|'; + pushback(p, '|'); + return c; + } return tOROP; } if (c == '=') { - yylval.id = '|'; - lex_state = EXPR_BEG; + set_yylval_id('|'); + SET_LEX_STATE(EXPR_BEG); return tOP_ASGN; } - if (lex_state == EXPR_FNAME || lex_state == EXPR_DOT) { - lex_state = EXPR_ARG; - } - else { - lex_state = EXPR_BEG; - } - pushback(c); + SET_LEX_STATE(IS_AFTER_OPERATOR() ? EXPR_ARG : EXPR_BEG|EXPR_LABEL); + pushback(p, c); return '|'; case '+': - c = nextc(); - if (lex_state == EXPR_FNAME || lex_state == EXPR_DOT) { - lex_state = EXPR_ARG; + c = nextc(p); + if (IS_AFTER_OPERATOR()) { + SET_LEX_STATE(EXPR_ARG); if (c == '@') { return tUPLUS; } - pushback(c); + pushback(p, c); return '+'; } if (c == '=') { - yylval.id = '+'; - lex_state = EXPR_BEG; + set_yylval_id('+'); + SET_LEX_STATE(EXPR_BEG); return tOP_ASGN; } - if (lex_state == EXPR_BEG || lex_state == EXPR_MID || - (IS_ARG() && space_seen && !ISSPACE(c))) { - if (IS_ARG()) arg_ambiguous(); - lex_state = EXPR_BEG; - pushback(c); - if (ISDIGIT(c)) { - c = '+'; - goto start_num; + if (IS_BEG() || (IS_SPCARG(c) && arg_ambiguous(p, '+'))) { + SET_LEX_STATE(EXPR_BEG); + pushback(p, c); + if (c != -1 && ISDIGIT(c)) { + return parse_numeric(p, '+'); } return tUPLUS; } - lex_state = EXPR_BEG; - pushback(c); - return '+'; + SET_LEX_STATE(EXPR_BEG); + pushback(p, c); + return warn_balanced('+', "+", "unary operator"); case '-': - c = nextc(); - if (lex_state == EXPR_FNAME || lex_state == EXPR_DOT) { - lex_state = EXPR_ARG; + c = nextc(p); + if (IS_AFTER_OPERATOR()) { + SET_LEX_STATE(EXPR_ARG); if (c == '@') { return tUMINUS; } - pushback(c); + pushback(p, c); return '-'; } if (c == '=') { - yylval.id = '-'; - lex_state = EXPR_BEG; + set_yylval_id('-'); + SET_LEX_STATE(EXPR_BEG); return tOP_ASGN; } - if (lex_state == EXPR_BEG || lex_state == EXPR_MID || - (IS_ARG() && space_seen && !ISSPACE(c))) { - if (IS_ARG()) arg_ambiguous(); - lex_state = EXPR_BEG; - pushback(c); - if (ISDIGIT(c)) { + if (c == '>') { + SET_LEX_STATE(EXPR_ENDFN); + return tLAMBDA; + } + if (IS_BEG() || (IS_SPCARG(c) && arg_ambiguous(p, '-'))) { + SET_LEX_STATE(EXPR_BEG); + pushback(p, c); + if (c != -1 && ISDIGIT(c)) { return tUMINUS_NUM; } return tUMINUS; } - lex_state = EXPR_BEG; - pushback(c); - return '-'; - - case '.': - lex_state = EXPR_BEG; - if ((c = nextc()) == '.') { - if ((c = nextc()) == '.') { - return tDOT3; - } - pushback(c); - return tDOT2; - } - pushback(c); - if (ISDIGIT(c)) { - yyerror("no .<digit> floating literal anymore; put 0 before dot"); - } - lex_state = EXPR_DOT; - return '.'; - - start_num: - case '0': case '1': case '2': case '3': case '4': - case '5': case '6': case '7': case '8': case '9': - { - int is_float, seen_point, seen_e, nondigit; - - is_float = seen_point = seen_e = nondigit = 0; - lex_state = EXPR_END; - newtok(); - if (c == '-' || c == '+') { - tokadd(c); - c = nextc(); - } - if (c == '0') { - int start = toklen(); - c = nextc(); - if (c == 'x' || c == 'X') { - /* hexadecimal */ - c = nextc(); - if (ISXDIGIT(c)) { - do { - if (c == '_') { - if (nondigit) break; - nondigit = c; - continue; - } - if (!ISXDIGIT(c)) break; - nondigit = 0; - tokadd(c); - } while ((c = nextc()) != -1); - } - pushback(c); - tokfix(); - if (toklen() == start) { - yyerror("numeric literal without digits"); - } - else if (nondigit) goto trailing_uc; - yylval.node = NEW_LIT(rb_cstr_to_inum(tok(), 16, Qfalse)); - return tINTEGER; - } - if (c == 'b' || c == 'B') { - /* binary */ - c = nextc(); - if (c == '0' || c == '1') { - do { - if (c == '_') { - if (nondigit) break; - nondigit = c; - continue; - } - if (c != '0' && c != '1') break; - nondigit = 0; - tokadd(c); - } while ((c = nextc()) != -1); - } - pushback(c); - tokfix(); - if (toklen() == start) { - yyerror("numeric literal without digits"); - } - else if (nondigit) goto trailing_uc; - yylval.node = NEW_LIT(rb_cstr_to_inum(tok(), 2, Qfalse)); - return tINTEGER; - } - if (c == 'd' || c == 'D') { - /* decimal */ - c = nextc(); - if (ISDIGIT(c)) { - do { - if (c == '_') { - if (nondigit) break; - nondigit = c; - continue; - } - if (!ISDIGIT(c)) break; - nondigit = 0; - tokadd(c); - } while ((c = nextc()) != -1); - } - pushback(c); - tokfix(); - if (toklen() == start) { - yyerror("numeric literal without digits"); - } - else if (nondigit) goto trailing_uc; - yylval.node = NEW_LIT(rb_cstr_to_inum(tok(), 10, Qfalse)); - return tINTEGER; - } - if (c == '_') { - /* 0_0 */ - goto octal_number; - } - if (c == 'o' || c == 'O') { - /* prefixed octal */ - c = nextc(); - if (c == '_') { - yyerror("numeric literal without digits"); - } - } - if (c >= '0' && c <= '7') { - /* octal */ - octal_number: - do { - if (c == '_') { - if (nondigit) break; - nondigit = c; - continue; - } - if (c < '0' || c > '7') break; - nondigit = 0; - tokadd(c); - } while ((c = nextc()) != -1); - if (toklen() > start) { - pushback(c); - tokfix(); - if (nondigit) goto trailing_uc; - yylval.node = NEW_LIT(rb_cstr_to_inum(tok(), 8, Qfalse)); - return tINTEGER; - } - if (nondigit) { - pushback(c); - goto trailing_uc; - } - } - if (c > '7' && c <= '9') { - yyerror("Illegal octal digit"); + SET_LEX_STATE(EXPR_BEG); + pushback(p, c); + return warn_balanced('-', "-", "unary operator"); + + case '.': { + int is_beg = IS_BEG(); + SET_LEX_STATE(EXPR_BEG); + if ((c = nextc(p)) == '.') { + if ((c = nextc(p)) == '.') { + if (p->ctxt.in_argdef) { + SET_LEX_STATE(EXPR_ENDARG); + return tBDOT3; } - else if (c == '.' || c == 'e' || c == 'E') { - tokadd('0'); + if (p->lex.paren_nest == 0 && looking_at_eol_p(p)) { + rb_warn0("... at EOL, should be parenthesized?"); } - else { - pushback(c); - yylval.node = NEW_LIT(INT2FIX(0)); - return tINTEGER; + else if (p->lex.lpar_beg >= 0 && p->lex.lpar_beg+1 == p->lex.paren_nest) { + if (IS_lex_state_for(last_state, EXPR_LABEL)) + return tDOT3; } + return is_beg ? tBDOT3 : tDOT3; } - - for (;;) { - switch (c) { - case '0': case '1': case '2': case '3': case '4': - case '5': case '6': case '7': case '8': case '9': - nondigit = 0; - tokadd(c); - break; - - case '.': - if (nondigit) goto trailing_uc; - if (seen_point || seen_e) { - goto decode_num; - } - else { - int c0 = nextc(); - if (!ISDIGIT(c0)) { - pushback(c0); - goto decode_num; - } - c = c0; - } - tokadd('.'); - tokadd(c); - is_float++; - seen_point++; - nondigit = 0; - break; - - case 'e': - case 'E': - if (nondigit) { - pushback(c); - c = nondigit; - goto decode_num; - } - if (seen_e) { - goto decode_num; - } - tokadd(c); - seen_e++; - is_float++; - nondigit = c; - c = nextc(); - if (c != '-' && c != '+') continue; - tokadd(c); - nondigit = c; - break; - - case '_': /* `_' in number just ignored */ - if (nondigit) goto decode_num; - nondigit = c; - break; - - default: - goto decode_num; - } - c = nextc(); + pushback(p, c); + return is_beg ? tBDOT2 : tDOT2; + } + pushback(p, c); + if (c != -1 && ISDIGIT(c)) { + char prev = p->lex.pcur-1 > p->lex.pbeg ? *(p->lex.pcur-2) : 0; + parse_numeric(p, '.'); + if (ISDIGIT(prev)) { + yyerror0("unexpected fraction part after numeric literal"); } - - decode_num: - pushback(c); - tokfix(); - if (nondigit) { - char tmp[30]; - trailing_uc: - sprintf(tmp, "trailing `%c' in number", nondigit); - yyerror(tmp); - } - if (is_float) { - double d = strtod(tok(), 0); - if (errno == ERANGE) { - rb_warn("Float %s out of range", tok()); - errno = 0; - } - yylval.node = NEW_LIT(rb_float_new(d)); - return tFLOAT; + else { + yyerror0("no .<digit> floating literal anymore; put 0 before dot"); } - yylval.node = NEW_LIT(rb_cstr_to_inum(tok(), 10, Qfalse)); - return tINTEGER; + SET_LEX_STATE(EXPR_END); + p->lex.ptok = p->lex.pcur; + goto retry; } + set_yylval_id('.'); + SET_LEX_STATE(EXPR_DOT); + return '.'; + } + + case '0': case '1': case '2': case '3': case '4': + case '5': case '6': case '7': case '8': case '9': + return parse_numeric(p, c); + + case ')': + COND_POP(); + CMDARG_POP(); + SET_LEX_STATE(EXPR_ENDFN); + p->lex.paren_nest--; + return c; case ']': + COND_POP(); + CMDARG_POP(); + SET_LEX_STATE(EXPR_END); + p->lex.paren_nest--; + return c; + case '}': - case ')': - COND_LEXPOP(); - CMDARG_LEXPOP(); - lex_state = EXPR_END; + /* tSTRING_DEND does COND_POP and CMDARG_POP in the yacc's rule */ + if (!p->lex.brace_nest--) return tSTRING_DEND; + COND_POP(); + CMDARG_POP(); + SET_LEX_STATE(EXPR_END); + p->lex.paren_nest--; return c; case ':': - c = nextc(); + c = nextc(p); if (c == ':') { - if (lex_state == EXPR_BEG || lex_state == EXPR_MID || - lex_state == EXPR_CLASS || (IS_ARG() && space_seen)) { - lex_state = EXPR_BEG; + if (IS_BEG() || IS_lex_state(EXPR_CLASS) || IS_SPCARG(-1)) { + SET_LEX_STATE(EXPR_BEG); return tCOLON3; } - lex_state = EXPR_DOT; + set_yylval_id(idCOLON2); + SET_LEX_STATE(EXPR_DOT); return tCOLON2; } - if (lex_state == EXPR_END || lex_state == EXPR_ENDARG || ISSPACE(c)) { - pushback(c); - lex_state = EXPR_BEG; - return ':'; + if (IS_END() || ISSPACE(c) || c == '#') { + pushback(p, c); + c = warn_balanced(':', ":", "symbol literal"); + SET_LEX_STATE(EXPR_BEG); + return c; } switch (c) { case '\'': - lex_strterm = NEW_STRTERM(str_ssym, c, 0); + p->lex.strterm = NEW_STRTERM(str_ssym, c, 0); break; case '"': - lex_strterm = NEW_STRTERM(str_dsym, c, 0); + p->lex.strterm = NEW_STRTERM(str_dsym, c, 0); break; default: - pushback(c); + pushback(p, c); break; } - lex_state = EXPR_FNAME; + SET_LEX_STATE(EXPR_FNAME); return tSYMBEG; case '/': - if (lex_state == EXPR_BEG || lex_state == EXPR_MID) { - lex_strterm = NEW_STRTERM(str_regexp, '/', 0); + if (IS_BEG()) { + p->lex.strterm = NEW_STRTERM(str_regexp, '/', 0); return tREGEXP_BEG; } - if ((c = nextc()) == '=') { - yylval.id = '/'; - lex_state = EXPR_BEG; + if ((c = nextc(p)) == '=') { + set_yylval_id('/'); + SET_LEX_STATE(EXPR_BEG); return tOP_ASGN; } - pushback(c); - if (IS_ARG() && space_seen) { - if (!ISSPACE(c)) { - arg_ambiguous(); - lex_strterm = NEW_STRTERM(str_regexp, '/', 0); - return tREGEXP_BEG; - } - } - switch (lex_state) { - case EXPR_FNAME: case EXPR_DOT: - lex_state = EXPR_ARG; break; - default: - lex_state = EXPR_BEG; break; + pushback(p, c); + if (IS_SPCARG(c)) { + arg_ambiguous(p, '/'); + p->lex.strterm = NEW_STRTERM(str_regexp, '/', 0); + return tREGEXP_BEG; } - return '/'; + SET_LEX_STATE(IS_AFTER_OPERATOR() ? EXPR_ARG : EXPR_BEG); + return warn_balanced('/', "/", "regexp literal"); case '^': - if ((c = nextc()) == '=') { - yylval.id = '^'; - lex_state = EXPR_BEG; + if ((c = nextc(p)) == '=') { + set_yylval_id('^'); + SET_LEX_STATE(EXPR_BEG); return tOP_ASGN; } - switch (lex_state) { - case EXPR_FNAME: case EXPR_DOT: - lex_state = EXPR_ARG; break; - default: - lex_state = EXPR_BEG; break; - } - pushback(c); + SET_LEX_STATE(IS_AFTER_OPERATOR() ? EXPR_ARG : EXPR_BEG); + pushback(p, c); return '^'; case ';': - command_start = Qtrue; + SET_LEX_STATE(EXPR_BEG); + p->command_start = TRUE; + return ';'; + case ',': - lex_state = EXPR_BEG; - return c; + SET_LEX_STATE(EXPR_BEG|EXPR_LABEL); + return ','; case '~': - if (lex_state == EXPR_FNAME || lex_state == EXPR_DOT) { - if ((c = nextc()) != '@') { - pushback(c); + if (IS_AFTER_OPERATOR()) { + if ((c = nextc(p)) != '@') { + pushback(p, c); } + SET_LEX_STATE(EXPR_ARG); } - switch (lex_state) { - case EXPR_FNAME: case EXPR_DOT: - lex_state = EXPR_ARG; break; - default: - lex_state = EXPR_BEG; break; + else { + SET_LEX_STATE(EXPR_BEG); } return '~'; case '(': - command_start = Qtrue; - if (lex_state == EXPR_BEG || lex_state == EXPR_MID) { + if (IS_BEG()) { c = tLPAREN; } - else if (space_seen) { - if (lex_state == EXPR_CMDARG) { - c = tLPAREN_ARG; - } - else if (lex_state == EXPR_ARG) { - rb_warn("don't put space before argument parentheses"); - c = '('; - } + else if (!space_seen) { + /* foo( ... ) => method call, no ambiguity */ + } + else if (IS_ARG() || IS_lex_state_all(EXPR_END|EXPR_LABEL)) { + c = tLPAREN_ARG; } + else if (IS_lex_state(EXPR_ENDFN) && !lambda_beginning_p()) { + rb_warning0("parentheses after method name is interpreted as " + "an argument list, not a decomposed argument"); + } + p->lex.paren_nest++; COND_PUSH(0); CMDARG_PUSH(0); - lex_state = EXPR_BEG; + SET_LEX_STATE(EXPR_BEG|EXPR_LABEL); return c; case '[': - if (lex_state == EXPR_FNAME || lex_state == EXPR_DOT) { - lex_state = EXPR_ARG; - if ((c = nextc()) == ']') { - if ((c = nextc()) == '=') { + p->lex.paren_nest++; + if (IS_AFTER_OPERATOR()) { + if ((c = nextc(p)) == ']') { + p->lex.paren_nest--; + SET_LEX_STATE(EXPR_ARG); + if ((c = nextc(p)) == '=') { return tASET; } - pushback(c); + pushback(p, c); return tAREF; } - pushback(c); + pushback(p, c); + SET_LEX_STATE(EXPR_ARG|EXPR_LABEL); return '['; } - else if (lex_state == EXPR_BEG || lex_state == EXPR_MID) { + else if (IS_BEG()) { c = tLBRACK; } - else if (IS_ARG() && space_seen) { + else if (IS_ARG() && (space_seen || IS_lex_state(EXPR_LABELED))) { c = tLBRACK; } - lex_state = EXPR_BEG; + SET_LEX_STATE(EXPR_BEG|EXPR_LABEL); COND_PUSH(0); CMDARG_PUSH(0); return c; case '{': - if (IS_ARG() || lex_state == EXPR_END) + ++p->lex.brace_nest; + if (lambda_beginning_p()) + c = tLAMBEG; + else if (IS_lex_state(EXPR_LABELED)) + c = tLBRACE; /* hash */ + else if (IS_lex_state(EXPR_ARG_ANY | EXPR_END | EXPR_ENDFN)) c = '{'; /* block (primary) */ - else if (lex_state == EXPR_ENDARG) + else if (IS_lex_state(EXPR_ENDARG)) c = tLBRACE_ARG; /* block (expr) */ else c = tLBRACE; /* hash */ + if (c != tLBRACE) { + p->command_start = TRUE; + SET_LEX_STATE(EXPR_BEG); + } + else { + SET_LEX_STATE(EXPR_BEG|EXPR_LABEL); + } + ++p->lex.paren_nest; /* after lambda_beginning_p() */ COND_PUSH(0); CMDARG_PUSH(0); - lex_state = EXPR_BEG; return c; case '\\': - c = nextc(); + c = nextc(p); if (c == '\n') { space_seen = 1; + dispatch_scan_event(p, tSP); goto retry; /* skip \\n */ } - pushback(c); + if (c == ' ') return tSP; + if (ISSPACE(c)) return c; + pushback(p, c); return '\\'; case '%': - if (lex_state == EXPR_BEG || lex_state == EXPR_MID) { - int term; - int paren; - - c = nextc(); - quotation: - if (!ISALNUM(c)) { - term = c; - c = 'Q'; - } - else { - term = nextc(); - if (ISALNUM(term) || ismbchar(term)) { - yyerror("unknown type of %string"); - return 0; - } - } - if (c == -1 || term == -1) { - rb_compile_error("unterminated quoted string meets end of file"); - return 0; - } - paren = term; - if (term == '(') term = ')'; - else if (term == '[') term = ']'; - else if (term == '{') term = '}'; - else if (term == '<') term = '>'; - else paren = 0; - - switch (c) { - case 'Q': - lex_strterm = NEW_STRTERM(str_dquote, term, paren); - return tSTRING_BEG; - - case 'q': - lex_strterm = NEW_STRTERM(str_squote, term, paren); - return tSTRING_BEG; - - case 'W': - lex_strterm = NEW_STRTERM(str_dword, term, paren); - do {c = nextc();} while (ISSPACE(c)); - pushback(c); - return tWORDS_BEG; - - case 'w': - lex_strterm = NEW_STRTERM(str_sword, term, paren); - do {c = nextc();} while (ISSPACE(c)); - pushback(c); - return tQWORDS_BEG; - - case 'x': - lex_strterm = NEW_STRTERM(str_xquote, term, paren); - return tXSTRING_BEG; - - case 'r': - lex_strterm = NEW_STRTERM(str_regexp, term, paren); - return tREGEXP_BEG; - - case 's': - lex_strterm = NEW_STRTERM(str_ssym, term, paren); - lex_state = EXPR_FNAME; - return tSYMBEG; - - default: - yyerror("unknown type of %string"); - return 0; - } - } - if ((c = nextc()) == '=') { - yylval.id = '%'; - lex_state = EXPR_BEG; - return tOP_ASGN; - } - if (IS_ARG() && space_seen && !ISSPACE(c)) { - goto quotation; - } - switch (lex_state) { - case EXPR_FNAME: case EXPR_DOT: - lex_state = EXPR_ARG; break; - default: - lex_state = EXPR_BEG; break; - } - pushback(c); - return '%'; + return parse_percent(p, space_seen, last_state); case '$': - last_state = lex_state; - lex_state = EXPR_END; - newtok(); - c = nextc(); - switch (c) { - case '_': /* $_: last read line string */ - c = nextc(); - if (is_identchar(c)) { - tokadd('$'); - tokadd('_'); - break; - } - pushback(c); - c = '_'; - /* fall through */ - case '~': /* $~: match-data */ - local_cnt(c); - /* fall through */ - case '*': /* $*: argv */ - case '$': /* $$: pid */ - case '?': /* $?: last status */ - case '!': /* $!: error string */ - case '@': /* $@: error position */ - case '/': /* $/: input record separator */ - case '\\': /* $\: output record separator */ - case ';': /* $;: field separator */ - case ',': /* $,: output field separator */ - case '.': /* $.: last read line number */ - case '=': /* $=: ignorecase */ - case ':': /* $:: load path */ - case '<': /* $<: reading filename */ - case '>': /* $>: default output handle */ - case '\"': /* $": already loaded files */ - tokadd('$'); - tokadd(c); - tokfix(); - yylval.id = rb_intern(tok()); - return tGVAR; - - case '-': - tokadd('$'); - tokadd(c); - c = nextc(); - if (is_identchar(c)) { - tokadd(c); - } - else { - pushback(c); - } - gvar: - tokfix(); - yylval.id = rb_intern(tok()); - /* xxx shouldn't check if valid option variable */ - return tGVAR; - - case '&': /* $&: last match */ - case '`': /* $`: string before last match */ - case '\'': /* $': string after last match */ - case '+': /* $+: string matches last paren. */ - if (last_state == EXPR_FNAME) { - tokadd('$'); - tokadd(c); - goto gvar; - } - yylval.node = NEW_BACK_REF(c); - return tBACK_REF; - - case '1': case '2': case '3': - case '4': case '5': case '6': - case '7': case '8': case '9': - tokadd('$'); - do { - tokadd(c); - c = nextc(); - } while (ISDIGIT(c)); - pushback(c); - if (last_state == EXPR_FNAME) goto gvar; - tokfix(); - yylval.node = NEW_NTH_REF(atoi(tok()+1)); - return tNTH_REF; - - default: - if (!is_identchar(c)) { - pushback(c); - return '$'; - } - case '0': - tokadd('$'); - } - break; + return parse_gvar(p, last_state); case '@': - c = nextc(); - newtok(); - tokadd('@'); - if (c == '@') { - tokadd('@'); - c = nextc(); - } - if (ISDIGIT(c)) { - if (tokidx == 1) { - rb_compile_error("`@%c' is not allowed as an instance variable name", c); - } - else { - rb_compile_error("`@@%c' is not allowed as a class variable name", c); - } - return 0; - } - if (!is_identchar(c)) { - pushback(c); - return '@'; - } - break; + return parse_atmark(p, last_state); case '_': - if (was_bol() && whole_match_p("__END__", 7, 0)) { - ruby__end__seen = 1; - lex_lastline = 0; + if (was_bol(p) && whole_match_p(p, "__END__", 7, 0)) { + p->ruby__end__seen = 1; + p->eofp = 1; +#ifndef RIPPER return -1; +#else + lex_goto_eol(p); + dispatch_scan_event(p, k__END__); + return 0; +#endif } - newtok(); + newtok(p); break; default: - if (!is_identchar(c)) { - rb_compile_error("Invalid char `\\%03o' in expression", c); + if (!parser_is_identchar(p)) { + compile_error(p, "Invalid char `\\x%02X' in expression", c); + token_flush(p); goto retry; } - newtok(); + newtok(p); break; } - do { - tokadd(c); - if (ismbchar(c)) { - int i, len = mbclen(c)-1; - - for (i = 0; i < len; i++) { - c = nextc(); - tokadd(c); - } - } - c = nextc(); - } while (is_identchar(c)); - if ((c == '!' || c == '?') && is_identchar(tok()[0]) && !peek('=')) { - tokadd(c); - } - else { - pushback(c); - } - tokfix(); + return parse_ident(p, c, cmd_state); +} - { - int result = 0; +static enum yytokentype +yylex(YYSTYPE *lval, YYLTYPE *yylloc, struct parser_params *p) +{ + enum yytokentype t; - last_state = lex_state; - switch (tok()[0]) { - case '$': - lex_state = EXPR_END; - result = tGVAR; - break; - case '@': - lex_state = EXPR_END; - if (tok()[1] == '@') - result = tCVAR; - else - result = tIVAR; - break; + p->lval = lval; + lval->val = Qundef; + t = parser_yylex(p); - default: - if (toklast() == '!' || toklast() == '?') { - result = tFID; - } - else { - if (lex_state == EXPR_FNAME) { - if ((c = nextc()) == '=' && !peek('~') && !peek('>') && - (!peek('=') || (lex_p + 1 < lex_pend && lex_p[1] == '>'))) { - result = tIDENTIFIER; - tokadd(c); - tokfix(); - } - else { - pushback(c); - } - } - if (result == 0 && ISUPPER(tok()[0])) { - result = tCONSTANT; - } - else { - result = tIDENTIFIER; - } - } + if (p->lex.strterm && (p->lex.strterm->flags & STRTERM_HEREDOC)) + RUBY_SET_YYLLOC_FROM_STRTERM_HEREDOC(*yylloc); + else + RUBY_SET_YYLLOC(*yylloc); - if (lex_state != EXPR_DOT) { - struct kwtable *kw; - - /* See if it is a reserved word. */ - kw = rb_reserved_word(tok(), toklen()); - if (kw) { - enum lex_state state = lex_state; - lex_state = kw->state; - if (state == EXPR_FNAME) { - yylval.id = rb_intern(kw->name); - return kw->id[0]; - } - if (kw->id[0] == kDO) { - if (COND_P()) return kDO_COND; - if (CMDARG_P() && state != EXPR_CMDARG) - return kDO_BLOCK; - if (state == EXPR_ENDARG) - return kDO_BLOCK; - return kDO; - } - if (state == EXPR_BEG) - return kw->id[0]; - else { - if (kw->id[0] != kw->id[1]) - lex_state = EXPR_BEG; - return kw->id[1]; - } - } - } + if (has_delayed_token(p)) + dispatch_delayed_token(p, t); + else if (t != 0) + dispatch_scan_event(p, t); - if (lex_state == EXPR_BEG || - lex_state == EXPR_MID || - lex_state == EXPR_DOT || - lex_state == EXPR_ARG || - lex_state == EXPR_CMDARG) { - if (cmd_state) { - lex_state = EXPR_CMDARG; - } - else { - lex_state = EXPR_ARG; - } - } - else { - lex_state = EXPR_END; - } - } - yylval.id = rb_intern(tok()); - if (is_local_id(yylval.id) && - last_state != EXPR_DOT && - ((dyna_in_block() && rb_dvar_defined(yylval.id)) || local_id(yylval.id))) { - lex_state = EXPR_END; - } - return result; - } + return t; } -NODE* -rb_node_newnode(type, a0, a1, a2) - enum node_type type; - VALUE a0, a1, a2; -{ - NODE *n = (NODE*)rb_newobj(); +#define LVAR_USED ((ID)1 << (sizeof(ID) * CHAR_BIT - 1)) - n->flags |= T_NODE; - nd_set_type(n, type); - nd_set_line(n, ruby_sourceline); - n->nd_file = ruby_sourcefile; +static NODE* +node_newnode(struct parser_params *p, enum node_type type, VALUE a0, VALUE a1, VALUE a2, const rb_code_location_t *loc) +{ + NODE *n = rb_ast_newnode(p->ast, type); - n->u1.value = a0; - n->u2.value = a1; - n->u3.value = a2; + rb_node_init(n, type, a0, a1, a2); + nd_set_loc(n, loc); + nd_set_node_id(n, parser_get_node_id(p)); return n; } +static NODE * +nd_set_loc(NODE *nd, const YYLTYPE *loc) +{ + nd->nd_loc = *loc; + nd_set_line(nd, loc->beg_pos.lineno); + return nd; +} + +#ifndef RIPPER static enum node_type -nodetype(node) /* for debug */ - NODE *node; +nodetype(NODE *node) /* for debug */ { return (enum node_type)nd_type(node); } static int -nodeline(node) - NODE *node; +nodeline(NODE *node) { return nd_line(node); } static NODE* -newline_node(node) - NODE *node; +newline_node(NODE *node) { - NODE *nl = 0; if (node) { - if (nd_type(node) == NODE_NEWLINE) return node; - nl = NEW_NEWLINE(node); - fixpos(nl, node); - nl->nd_nth = nd_line(node); + node = remove_begin(node); + node->flags |= NODE_FL_NEWLINE; } - return nl; + return node; } static void -fixpos(node, orig) - NODE *node, *orig; +fixpos(NODE *node, NODE *orig) { if (!node) return; if (!orig) return; - if (orig == (NODE*)1) return; - node->nd_file = orig->nd_file; nd_set_line(node, nd_line(orig)); } static void -parser_warning(node, mesg) - NODE *node; - const char *mesg; +parser_warning(struct parser_params *p, NODE *node, const char *mesg) { - int line = ruby_sourceline; - ruby_sourceline = nd_line(node); - rb_warning("%s", mesg); - ruby_sourceline = line; + rb_compile_warning(p->ruby_sourcefile, nd_line(node), "%s", mesg); } static void -parser_warn(node, mesg) - NODE *node; - const char *mesg; +parser_warn(struct parser_params *p, NODE *node, const char *mesg) { - int line = ruby_sourceline; - ruby_sourceline = nd_line(node); - rb_warn("%s", mesg); - ruby_sourceline = line; + rb_compile_warn(p->ruby_sourcefile, nd_line(node), "%s", mesg); } static NODE* -block_append(head, tail) - NODE *head, *tail; +block_append(struct parser_params *p, NODE *head, NODE *tail) { - NODE *end, *h = head; + NODE *end, *h = head, *nd; if (tail == 0) return head; - again: if (h == 0) return tail; switch (nd_type(h)) { - case NODE_NEWLINE: - h = h->nd_next; - goto again; case NODE_LIT: case NODE_STR: - parser_warning(h, "unused literal ignored"); + case NODE_SELF: + case NODE_TRUE: + case NODE_FALSE: + case NODE_NIL: + parser_warning(p, h, "unused literal ignored"); return tail; default: - h = end = NEW_BLOCK(head); + h = end = NEW_BLOCK(head, &head->nd_loc); end->nd_end = end; - fixpos(end, head); head = end; break; case NODE_BLOCK: @@ -4603,44 +10075,39 @@ block_append(head, tail) break; } - if (RTEST(ruby_verbose)) { - NODE *nd = end->nd_head; - newline: - switch (nd_type(nd)) { - case NODE_RETURN: - case NODE_BREAK: - case NODE_NEXT: - case NODE_REDO: - case NODE_RETRY: - parser_warning(nd, "statement not reached"); - break; - - case NODE_NEWLINE: - nd = nd->nd_next; - goto newline; - - default: - break; + nd = end->nd_head; + switch (nd_type(nd)) { + case NODE_RETURN: + case NODE_BREAK: + case NODE_NEXT: + case NODE_REDO: + case NODE_RETRY: + if (RTEST(ruby_verbose)) { + parser_warning(p, tail, "statement not reached"); } + break; + + default: + break; } - if (nd_type(tail) != NODE_BLOCK) { - tail = NEW_BLOCK(tail); + if (!nd_type_p(tail, NODE_BLOCK)) { + tail = NEW_BLOCK(tail, &tail->nd_loc); tail->nd_end = tail; } end->nd_next = tail; h->nd_end = tail->nd_end; + nd_set_last_loc(head, nd_last_loc(tail)); return head; } /* append item to the list */ static NODE* -list_append(list, item) - NODE *list, *item; +list_append(struct parser_params *p, NODE *list, NODE *item) { NODE *last; - if (list == 0) return NEW_LIST(item); + if (list == 0) return NEW_LIST(item, &item->nd_loc); if (list->nd_next) { last = list->nd_next->nd_end; } @@ -4649,15 +10116,17 @@ list_append(list, item) } list->nd_alen += 1; - last->nd_next = NEW_LIST(item); + last->nd_next = NEW_LIST(item, &item->nd_loc); list->nd_next->nd_end = last->nd_next; + + nd_set_last_loc(list, nd_last_loc(item)); + return list; } /* concat two lists */ static NODE* -list_concat(head, tail) - NODE *head, *tail; +list_concat(NODE *head, NODE *tail) { NODE *last; @@ -4677,46 +10146,115 @@ list_concat(head, tail) head->nd_next->nd_end = tail; } + nd_set_last_loc(head, nd_last_loc(tail)); + return head; } +static int +literal_concat0(struct parser_params *p, VALUE head, VALUE tail) +{ + if (NIL_P(tail)) return 1; + if (!rb_enc_compatible(head, tail)) { + compile_error(p, "string literal encodings differ (%s / %s)", + rb_enc_name(rb_enc_get(head)), + rb_enc_name(rb_enc_get(tail))); + rb_str_resize(head, 0); + rb_str_resize(tail, 0); + return 0; + } + rb_str_buf_append(head, tail); + return 1; +} + +static VALUE +string_literal_head(enum node_type htype, NODE *head) +{ + if (htype != NODE_DSTR) return Qfalse; + if (head->nd_next) { + head = head->nd_next->nd_end->nd_head; + if (!head || !nd_type_p(head, NODE_STR)) return Qfalse; + } + const VALUE lit = head->nd_lit; + ASSUME(lit != Qfalse); + return lit; +} + /* concat two string literals */ static NODE * -literal_concat(head, tail) - NODE *head, *tail; +literal_concat(struct parser_params *p, NODE *head, NODE *tail, const YYLTYPE *loc) { enum node_type htype; + VALUE lit; if (!head) return tail; if (!tail) return head; htype = nd_type(head); if (htype == NODE_EVSTR) { - NODE *node = NEW_DSTR(rb_str_new(0, 0)); - head = list_append(node, head); + head = new_dstr(p, head, loc); + htype = NODE_DSTR; + } + if (p->heredoc_indent > 0) { + switch (htype) { + case NODE_STR: + nd_set_type(head, NODE_DSTR); + case NODE_DSTR: + return list_append(p, head, tail); + default: + break; + } } switch (nd_type(tail)) { case NODE_STR: + if ((lit = string_literal_head(htype, head)) != Qfalse) { + htype = NODE_STR; + } + else { + lit = head->nd_lit; + } if (htype == NODE_STR) { - rb_str_concat(head->nd_lit, tail->nd_lit); - rb_gc_force_recycle((VALUE)tail); + if (!literal_concat0(p, lit, tail->nd_lit)) { + error: + rb_discard_node(p, head); + rb_discard_node(p, tail); + return 0; + } + rb_discard_node(p, tail); } else { - list_append(head, tail); + list_append(p, head, tail); } break; case NODE_DSTR: if (htype == NODE_STR) { - rb_str_concat(head->nd_lit, tail->nd_lit); + if (!literal_concat0(p, head->nd_lit, tail->nd_lit)) + goto error; tail->nd_lit = head->nd_lit; - rb_gc_force_recycle((VALUE)head); + rb_discard_node(p, head); head = tail; } + else if (NIL_P(tail->nd_lit)) { + append: + head->nd_alen += tail->nd_alen - 1; + if (!head->nd_next) { + head->nd_next = tail->nd_next; + } + else if (tail->nd_next) { + head->nd_next->nd_end->nd_next = tail->nd_next; + head->nd_next->nd_end = tail->nd_next->nd_end; + } + rb_discard_node(p, tail); + } + else if ((lit = string_literal_head(htype, head)) != Qfalse) { + if (!literal_concat0(p, lit, tail->nd_lit)) + goto error; + tail->nd_lit = Qnil; + goto append; + } else { - nd_set_type(tail, NODE_ARRAY); - tail->nd_head = NEW_STR(tail->nd_lit); - list_concat(head, tail); + list_concat(head, NEW_NODE(NODE_LIST, NEW_STR(tail->nd_lit, loc), tail->nd_alen, tail->nd_next, loc)); } break; @@ -4725,300 +10263,1193 @@ literal_concat(head, tail) nd_set_type(head, NODE_DSTR); head->nd_alen = 1; } - list_append(head, tail); + list_append(p, head, tail); break; } return head; } static NODE * -evstr2dstr(node) - NODE *node; +evstr2dstr(struct parser_params *p, NODE *node) { - if (nd_type(node) == NODE_EVSTR) { - node = list_append(NEW_DSTR(rb_str_new(0, 0)), node); + if (nd_type_p(node, NODE_EVSTR)) { + node = new_dstr(p, node, &node->nd_loc); } return node; } static NODE * -new_evstr(node) - NODE *node; +new_evstr(struct parser_params *p, NODE *node, const YYLTYPE *loc) { NODE *head = node; - again: if (node) { switch (nd_type(node)) { - case NODE_STR: case NODE_DSTR: case NODE_EVSTR: + case NODE_STR: + nd_set_type(node, NODE_DSTR); + return node; + case NODE_DSTR: + break; + case NODE_EVSTR: return node; - case NODE_NEWLINE: - node = node->nd_next; - goto again; } } - return NEW_EVSTR(head); + return NEW_EVSTR(head, loc); } static NODE * -call_op(recv, id, narg, arg1) - NODE *recv; - ID id; - int narg; - NODE *arg1; +new_dstr(struct parser_params *p, NODE *node, const YYLTYPE *loc) +{ + VALUE lit = STR_NEW0(); + NODE *dstr = NEW_DSTR(lit, loc); + RB_OBJ_WRITTEN(p->ast, Qnil, lit); + return list_append(p, dstr, node); +} + +static NODE * +call_bin_op(struct parser_params *p, NODE *recv, ID id, NODE *arg1, + const YYLTYPE *op_loc, const YYLTYPE *loc) { + NODE *expr; value_expr(recv); - if (narg == 1) { - value_expr(arg1); - arg1 = NEW_LIST(arg1); - } - else { - arg1 = 0; - } - return NEW_CALL(recv, id, arg1); + value_expr(arg1); + expr = NEW_OPCALL(recv, id, NEW_LIST(arg1, &arg1->nd_loc), loc); + nd_set_line(expr, op_loc->beg_pos.lineno); + return expr; +} + +static NODE * +call_uni_op(struct parser_params *p, NODE *recv, ID id, const YYLTYPE *op_loc, const YYLTYPE *loc) +{ + NODE *opcall; + value_expr(recv); + opcall = NEW_OPCALL(recv, id, 0, loc); + nd_set_line(opcall, op_loc->beg_pos.lineno); + return opcall; +} + +static NODE * +new_qcall(struct parser_params* p, ID atype, NODE *recv, ID mid, NODE *args, const YYLTYPE *op_loc, const YYLTYPE *loc) +{ + NODE *qcall = NEW_QCALL(atype, recv, mid, args, loc); + nd_set_line(qcall, op_loc->beg_pos.lineno); + return qcall; +} + +static NODE* +new_command_qcall(struct parser_params* p, ID atype, NODE *recv, ID mid, NODE *args, NODE *block, const YYLTYPE *op_loc, const YYLTYPE *loc) +{ + NODE *ret; + if (block) block_dup_check(p, args, block); + ret = new_qcall(p, atype, recv, mid, args, op_loc, loc); + if (block) ret = method_add_block(p, ret, block, loc); + fixpos(ret, recv); + return ret; } +#define nd_once_body(node) (nd_type_p((node), NODE_ONCE) ? (node)->nd_body : node) static NODE* -match_gen(node1, node2) - NODE *node1; - NODE *node2; +match_op(struct parser_params *p, NODE *node1, NODE *node2, const YYLTYPE *op_loc, const YYLTYPE *loc) { - local_cnt('~'); + NODE *n; + int line = op_loc->beg_pos.lineno; value_expr(node1); value_expr(node2); - if (node1) { - switch (nd_type(node1)) { + if (node1 && (n = nd_once_body(node1)) != 0) { + switch (nd_type(n)) { case NODE_DREGX: - case NODE_DREGX_ONCE: - return NEW_MATCH2(node1, node2); + { + NODE *match = NEW_MATCH2(node1, node2, loc); + nd_set_line(match, line); + return match; + } case NODE_LIT: - if (TYPE(node1->nd_lit) == T_REGEXP) { - return NEW_MATCH2(node1, node2); + if (RB_TYPE_P(n->nd_lit, T_REGEXP)) { + const VALUE lit = n->nd_lit; + NODE *match = NEW_MATCH2(node1, node2, loc); + match->nd_args = reg_named_capture_assign(p, lit, loc); + nd_set_line(match, line); + return match; } } } - if (node2) { - switch (nd_type(node2)) { - case NODE_DREGX: - case NODE_DREGX_ONCE: - return NEW_MATCH3(node2, node1); + if (node2 && (n = nd_once_body(node2)) != 0) { + NODE *match3; + switch (nd_type(n)) { case NODE_LIT: - if (TYPE(node2->nd_lit) == T_REGEXP) { - return NEW_MATCH3(node2, node1); - } + if (!RB_TYPE_P(n->nd_lit, T_REGEXP)) break; + /* fallthru */ + case NODE_DREGX: + match3 = NEW_MATCH3(node2, node1, loc); + return match3; } } - return NEW_CALL(node1, tMATCH, NEW_LIST(node2)); + n = NEW_CALL(node1, tMATCH, NEW_LIST(node2, &node2->nd_loc), loc); + nd_set_line(n, line); + return n; +} + +# if WARN_PAST_SCOPE +static int +past_dvar_p(struct parser_params *p, ID id) +{ + struct vtable *past = p->lvtbl->past; + while (past) { + if (vtable_included(past, id)) return 1; + past = past->prev; + } + return 0; +} +# endif + +static int +numparam_nested_p(struct parser_params *p) +{ + struct local_vars *local = p->lvtbl; + NODE *outer = local->numparam.outer; + NODE *inner = local->numparam.inner; + if (outer || inner) { + NODE *used = outer ? outer : inner; + compile_error(p, "numbered parameter is already used in\n" + "%s:%d: %s block here", + p->ruby_sourcefile, nd_line(used), + outer ? "outer" : "inner"); + parser_show_error_line(p, &used->nd_loc); + return 1; + } + return 0; } static NODE* -gettable(id) - ID id; +gettable(struct parser_params *p, ID id, const YYLTYPE *loc) +{ + ID *vidp = NULL; + NODE *node; + switch (id) { + case keyword_self: + return NEW_SELF(loc); + case keyword_nil: + return NEW_NIL(loc); + case keyword_true: + return NEW_TRUE(loc); + case keyword_false: + return NEW_FALSE(loc); + case keyword__FILE__: + { + VALUE file = p->ruby_sourcefile_string; + if (NIL_P(file)) + file = rb_str_new(0, 0); + else + file = rb_str_dup(file); + node = NEW_STR(file, loc); + RB_OBJ_WRITTEN(p->ast, Qnil, file); + } + return node; + case keyword__LINE__: + return NEW_LIT(INT2FIX(p->tokline), loc); + case keyword__ENCODING__: + node = NEW_LIT(rb_enc_from_encoding(p->enc), loc); + RB_OBJ_WRITTEN(p->ast, Qnil, node->nd_lit); + return node; + + } + switch (id_type(id)) { + case ID_LOCAL: + if (dyna_in_block(p) && dvar_defined_ref(p, id, &vidp)) { + if (NUMPARAM_ID_P(id) && numparam_nested_p(p)) return 0; + if (id == p->cur_arg) { + compile_error(p, "circular argument reference - %"PRIsWARN, rb_id2str(id)); + return 0; + } + if (vidp) *vidp |= LVAR_USED; + node = NEW_DVAR(id, loc); + return node; + } + if (local_id_ref(p, id, &vidp)) { + if (id == p->cur_arg) { + compile_error(p, "circular argument reference - %"PRIsWARN, rb_id2str(id)); + return 0; + } + if (vidp) *vidp |= LVAR_USED; + node = NEW_LVAR(id, loc); + return node; + } + if (dyna_in_block(p) && NUMPARAM_ID_P(id) && + parser_numbered_param(p, NUMPARAM_ID_TO_IDX(id))) { + if (numparam_nested_p(p)) return 0; + node = NEW_DVAR(id, loc); + struct local_vars *local = p->lvtbl; + if (!local->numparam.current) local->numparam.current = node; + return node; + } +# if WARN_PAST_SCOPE + if (!p->ctxt.in_defined && RTEST(ruby_verbose) && past_dvar_p(p, id)) { + rb_warning1("possible reference to past scope - %"PRIsWARN, rb_id2str(id)); + } +# endif + /* method call without arguments */ + return NEW_VCALL(id, loc); + case ID_GLOBAL: + return NEW_GVAR(id, loc); + case ID_INSTANCE: + return NEW_IVAR(id, loc); + case ID_CONST: + return NEW_CONST(id, loc); + case ID_CLASS: + return NEW_CVAR(id, loc); + } + compile_error(p, "identifier %"PRIsVALUE" is not valid to get", rb_id2str(id)); + return 0; +} + +static NODE * +opt_arg_append(NODE *opt_list, NODE *opt) { - if (id == kSELF) { - return NEW_SELF(); + NODE *opts = opt_list; + opts->nd_loc.end_pos = opt->nd_loc.end_pos; + + while (opts->nd_next) { + opts = opts->nd_next; + opts->nd_loc.end_pos = opt->nd_loc.end_pos; } - else if (id == kNIL) { - return NEW_NIL(); + opts->nd_next = opt; + + return opt_list; +} + +static NODE * +kwd_append(NODE *kwlist, NODE *kw) +{ + if (kwlist) { + NODE *kws = kwlist; + kws->nd_loc.end_pos = kw->nd_loc.end_pos; + while (kws->nd_next) { + kws = kws->nd_next; + kws->nd_loc.end_pos = kw->nd_loc.end_pos; + } + kws->nd_next = kw; } - else if (id == kTRUE) { - return NEW_TRUE(); + return kwlist; +} + +static NODE * +new_defined(struct parser_params *p, NODE *expr, const YYLTYPE *loc) +{ + return NEW_DEFINED(remove_begin_all(expr), loc); +} + +static NODE* +symbol_append(struct parser_params *p, NODE *symbols, NODE *symbol) +{ + enum node_type type = nd_type(symbol); + switch (type) { + case NODE_DSTR: + nd_set_type(symbol, NODE_DSYM); + break; + case NODE_STR: + nd_set_type(symbol, NODE_LIT); + RB_OBJ_WRITTEN(p->ast, Qnil, symbol->nd_lit = rb_str_intern(symbol->nd_lit)); + break; + default: + compile_error(p, "unexpected node as symbol: %s", ruby_node_name(type)); } - else if (id == kFALSE) { - return NEW_FALSE(); + return list_append(p, symbols, symbol); +} + +static NODE * +new_regexp(struct parser_params *p, NODE *node, int options, const YYLTYPE *loc) +{ + NODE *list, *prev; + VALUE lit; + + if (!node) { + node = NEW_LIT(reg_compile(p, STR_NEW0(), options), loc); + RB_OBJ_WRITTEN(p->ast, Qnil, node->nd_lit); + return node; } - else if (id == k__FILE__) { - return NEW_STR(rb_str_new2(ruby_sourcefile)); + switch (nd_type(node)) { + case NODE_STR: + { + VALUE src = node->nd_lit; + nd_set_type(node, NODE_LIT); + nd_set_loc(node, loc); + RB_OBJ_WRITTEN(p->ast, Qnil, node->nd_lit = reg_compile(p, src, options)); + } + break; + default: + lit = STR_NEW0(); + node = NEW_NODE(NODE_DSTR, lit, 1, NEW_LIST(node, loc), loc); + RB_OBJ_WRITTEN(p->ast, Qnil, lit); + /* fall through */ + case NODE_DSTR: + nd_set_type(node, NODE_DREGX); + nd_set_loc(node, loc); + node->nd_cflag = options & RE_OPTION_MASK; + if (!NIL_P(node->nd_lit)) reg_fragment_check(p, node->nd_lit, options); + for (list = (prev = node)->nd_next; list; list = list->nd_next) { + NODE *frag = list->nd_head; + enum node_type type = nd_type(frag); + if (type == NODE_STR || (type == NODE_DSTR && !frag->nd_next)) { + VALUE tail = frag->nd_lit; + if (reg_fragment_check(p, tail, options) && prev && !NIL_P(prev->nd_lit)) { + VALUE lit = prev == node ? prev->nd_lit : prev->nd_head->nd_lit; + if (!literal_concat0(p, lit, tail)) { + return NEW_NIL(loc); /* dummy node on error */ + } + rb_str_resize(tail, 0); + prev->nd_next = list->nd_next; + rb_discard_node(p, list->nd_head); + rb_discard_node(p, list); + list = prev; + } + else { + prev = list; + } + } + else { + prev = 0; + } + } + if (!node->nd_next) { + VALUE src = node->nd_lit; + nd_set_type(node, NODE_LIT); + RB_OBJ_WRITTEN(p->ast, Qnil, node->nd_lit = reg_compile(p, src, options)); + } + if (options & RE_OPTION_ONCE) { + node = NEW_NODE(NODE_ONCE, 0, node, 0, loc); + } + break; } - else if (id == k__LINE__) { - return NEW_LIT(INT2FIX(ruby_sourceline)); + return node; +} + +static NODE * +new_kw_arg(struct parser_params *p, NODE *k, const YYLTYPE *loc) +{ + if (!k) return 0; + return NEW_KW_ARG(0, (k), loc); +} + +static NODE * +new_xstring(struct parser_params *p, NODE *node, const YYLTYPE *loc) +{ + if (!node) { + VALUE lit = STR_NEW0(); + NODE *xstr = NEW_XSTR(lit, loc); + RB_OBJ_WRITTEN(p->ast, Qnil, lit); + return xstr; } - else if (is_local_id(id)) { - if (dyna_in_block() && rb_dvar_defined(id)) return NEW_DVAR(id); - if (local_id(id)) return NEW_LVAR(id); - /* method call without arguments */ -#if 0 - /* Rite will warn this */ - rb_warn("ambiguous identifier; %s() or self.%s is better for method call", - rb_id2name(id), rb_id2name(id)); -#endif - return NEW_VCALL(id); + switch (nd_type(node)) { + case NODE_STR: + nd_set_type(node, NODE_XSTR); + nd_set_loc(node, loc); + break; + case NODE_DSTR: + nd_set_type(node, NODE_DXSTR); + nd_set_loc(node, loc); + break; + default: + node = NEW_NODE(NODE_DXSTR, Qnil, 1, NEW_LIST(node, loc), loc); + break; } - else if (is_global_id(id)) { - return NEW_GVAR(id); + return node; +} + +static void +check_literal_when(struct parser_params *p, NODE *arg, const YYLTYPE *loc) +{ + VALUE lit; + + if (!arg || !p->case_labels) return; + + lit = rb_node_case_when_optimizable_literal(arg); + if (lit == Qundef) return; + if (nd_type_p(arg, NODE_STR)) { + RB_OBJ_WRITTEN(p->ast, Qnil, arg->nd_lit = lit); } - else if (is_instance_id(id)) { - return NEW_IVAR(id); + + if (NIL_P(p->case_labels)) { + p->case_labels = rb_obj_hide(rb_hash_new()); } - else if (is_const_id(id)) { - return NEW_CONST(id); + else { + VALUE line = rb_hash_lookup(p->case_labels, lit); + if (!NIL_P(line)) { + rb_warning1("duplicated `when' clause with line %d is ignored", + WARN_IVAL(line)); + return; + } } - else if (is_class_id(id)) { - return NEW_CVAR(id); + rb_hash_aset(p->case_labels, lit, INT2NUM(p->ruby_sourceline)); +} + +#else /* !RIPPER */ +static int +id_is_var(struct parser_params *p, ID id) +{ + if (is_notop_id(id)) { + switch (id & ID_SCOPE_MASK) { + case ID_GLOBAL: case ID_INSTANCE: case ID_CONST: case ID_CLASS: + return 1; + case ID_LOCAL: + if (dyna_in_block(p)) { + if (NUMPARAM_ID_P(id) || dvar_defined(p, id)) return 1; + } + if (local_id(p, id)) return 1; + /* method call without arguments */ + return 0; + } } - rb_compile_error("identifier %s is not valid", rb_id2name(id)); + compile_error(p, "identifier %"PRIsVALUE" is not valid to get", rb_id2str(id)); return 0; } -static VALUE dyna_var_lookup _((ID id)); - -static NODE* -assignable(id, val) - ID id; - NODE *val; +static VALUE +new_regexp(struct parser_params *p, VALUE re, VALUE opt, const YYLTYPE *loc) { - value_expr(val); - if (id == kSELF) { - yyerror("Can't change the value of self"); + VALUE src = 0, err; + int options = 0; + if (ripper_is_node_yylval(re)) { + src = RNODE(re)->nd_cval; + re = RNODE(re)->nd_rval; } - else if (id == kNIL) { - yyerror("Can't assign to nil"); + if (ripper_is_node_yylval(opt)) { + options = (int)RNODE(opt)->nd_tag; + opt = RNODE(opt)->nd_rval; } - else if (id == kTRUE) { - yyerror("Can't assign to true"); + if (src && NIL_P(parser_reg_compile(p, src, options, &err))) { + compile_error(p, "%"PRIsVALUE, err); } - else if (id == kFALSE) { - yyerror("Can't assign to false"); + return dispatch2(regexp_literal, re, opt); +} +#endif /* !RIPPER */ + +static inline enum lex_state_e +parser_set_lex_state(struct parser_params *p, enum lex_state_e ls, int line) +{ + if (p->debug) { + ls = rb_parser_trace_lex_state(p, p->lex.state, ls, line); } - else if (id == k__FILE__) { - yyerror("Can't assign to __FILE__"); + return p->lex.state = ls; +} + +#ifndef RIPPER +static const char rb_parser_lex_state_names[][8] = { + "BEG", "END", "ENDARG", "ENDFN", "ARG", + "CMDARG", "MID", "FNAME", "DOT", "CLASS", + "LABEL", "LABELED","FITEM", +}; + +static VALUE +append_lex_state_name(enum lex_state_e state, VALUE buf) +{ + int i, sep = 0; + unsigned int mask = 1; + static const char none[] = "NONE"; + + for (i = 0; i < EXPR_MAX_STATE; ++i, mask <<= 1) { + if ((unsigned)state & mask) { + if (sep) { + rb_str_cat(buf, "|", 1); + } + sep = 1; + rb_str_cat_cstr(buf, rb_parser_lex_state_names[i]); + } } - else if (id == k__LINE__) { - yyerror("Can't assign to __LINE__"); + if (!sep) { + rb_str_cat(buf, none, sizeof(none)-1); } - else if (is_local_id(id)) { - if (rb_dvar_curr(id)) { - return NEW_DASGN_CURR(id, val); - } - else if (dyna_var_lookup(id)) { - return NEW_DASGN(id, val); - } - else if (local_id(id) || !dyna_in_block()) { - return NEW_LASGN(id, val); - } - else{ - rb_dvar_push(id, Qnil); - return NEW_DASGN_CURR(id, val); - } + return buf; +} + +static void +flush_debug_buffer(struct parser_params *p, VALUE out, VALUE str) +{ + VALUE mesg = p->debug_buffer; + + if (!NIL_P(mesg) && RSTRING_LEN(mesg)) { + p->debug_buffer = Qnil; + rb_io_puts(1, &mesg, out); + } + if (!NIL_P(str) && RSTRING_LEN(str)) { + rb_io_write(p->debug_output, str); + } +} + +enum lex_state_e +rb_parser_trace_lex_state(struct parser_params *p, enum lex_state_e from, + enum lex_state_e to, int line) +{ + VALUE mesg; + mesg = rb_str_new_cstr("lex_state: "); + append_lex_state_name(from, mesg); + rb_str_cat_cstr(mesg, " -> "); + append_lex_state_name(to, mesg); + rb_str_catf(mesg, " at line %d\n", line); + flush_debug_buffer(p, p->debug_output, mesg); + return to; +} + +VALUE +rb_parser_lex_state_name(enum lex_state_e state) +{ + return rb_fstring(append_lex_state_name(state, rb_str_new(0, 0))); +} + +static void +append_bitstack_value(stack_type stack, VALUE mesg) +{ + if (stack == 0) { + rb_str_cat_cstr(mesg, "0"); } - else if (is_global_id(id)) { - return NEW_GASGN(id, val); + else { + stack_type mask = (stack_type)1U << (CHAR_BIT * sizeof(stack_type) - 1); + for (; mask && !(stack & mask); mask >>= 1) continue; + for (; mask; mask >>= 1) rb_str_cat(mesg, stack & mask ? "1" : "0", 1); } - else if (is_instance_id(id)) { - return NEW_IASGN(id, val); +} + +void +rb_parser_show_bitstack(struct parser_params *p, stack_type stack, + const char *name, int line) +{ + VALUE mesg = rb_sprintf("%s: ", name); + append_bitstack_value(stack, mesg); + rb_str_catf(mesg, " at line %d\n", line); + flush_debug_buffer(p, p->debug_output, mesg); +} + +void +rb_parser_fatal(struct parser_params *p, const char *fmt, ...) +{ + va_list ap; + VALUE mesg = rb_str_new_cstr("internal parser error: "); + + va_start(ap, fmt); + rb_str_vcatf(mesg, fmt, ap); + va_end(ap); + yyerror0(RSTRING_PTR(mesg)); + RB_GC_GUARD(mesg); + + mesg = rb_str_new(0, 0); + append_lex_state_name(p->lex.state, mesg); + compile_error(p, "lex.state: %"PRIsVALUE, mesg); + rb_str_resize(mesg, 0); + append_bitstack_value(p->cond_stack, mesg); + compile_error(p, "cond_stack: %"PRIsVALUE, mesg); + rb_str_resize(mesg, 0); + append_bitstack_value(p->cmdarg_stack, mesg); + compile_error(p, "cmdarg_stack: %"PRIsVALUE, mesg); + if (p->debug_output == rb_ractor_stdout()) + p->debug_output = rb_ractor_stderr(); + p->debug = TRUE; +} + +static YYLTYPE * +rb_parser_set_pos(YYLTYPE *yylloc, int sourceline, int beg_pos, int end_pos) +{ + yylloc->beg_pos.lineno = sourceline; + yylloc->beg_pos.column = beg_pos; + yylloc->end_pos.lineno = sourceline; + yylloc->end_pos.column = end_pos; + return yylloc; +} + +YYLTYPE * +rb_parser_set_location_from_strterm_heredoc(struct parser_params *p, rb_strterm_heredoc_t *here, YYLTYPE *yylloc) +{ + int sourceline = here->sourceline; + int beg_pos = (int)here->offset - here->quote + - (rb_strlen_lit("<<-") - !(here->func & STR_FUNC_INDENT)); + int end_pos = (int)here->offset + here->length + here->quote; + + return rb_parser_set_pos(yylloc, sourceline, beg_pos, end_pos); +} + +YYLTYPE * +rb_parser_set_location_of_none(struct parser_params *p, YYLTYPE *yylloc) +{ + int sourceline = p->ruby_sourceline; + int beg_pos = (int)(p->lex.ptok - p->lex.pbeg); + int end_pos = (int)(p->lex.ptok - p->lex.pbeg); + return rb_parser_set_pos(yylloc, sourceline, beg_pos, end_pos); +} + +YYLTYPE * +rb_parser_set_location(struct parser_params *p, YYLTYPE *yylloc) +{ + int sourceline = p->ruby_sourceline; + int beg_pos = (int)(p->lex.ptok - p->lex.pbeg); + int end_pos = (int)(p->lex.pcur - p->lex.pbeg); + return rb_parser_set_pos(yylloc, sourceline, beg_pos, end_pos); +} +#endif /* !RIPPER */ + +static int +assignable0(struct parser_params *p, ID id, const char **err) +{ + if (!id) return -1; + switch (id) { + case keyword_self: + *err = "Can't change the value of self"; + return -1; + case keyword_nil: + *err = "Can't assign to nil"; + return -1; + case keyword_true: + *err = "Can't assign to true"; + return -1; + case keyword_false: + *err = "Can't assign to false"; + return -1; + case keyword__FILE__: + *err = "Can't assign to __FILE__"; + return -1; + case keyword__LINE__: + *err = "Can't assign to __LINE__"; + return -1; + case keyword__ENCODING__: + *err = "Can't assign to __ENCODING__"; + return -1; } - else if (is_const_id(id)) { - if (in_def || in_single) - yyerror("dynamic constant assignment"); - return NEW_CDECL(id, val, 0); + switch (id_type(id)) { + case ID_LOCAL: + if (dyna_in_block(p)) { + if (p->max_numparam > NO_PARAM && NUMPARAM_ID_P(id)) { + compile_error(p, "Can't assign to numbered parameter _%d", + NUMPARAM_ID_TO_IDX(id)); + return -1; + } + if (dvar_curr(p, id)) return NODE_DASGN; + if (dvar_defined(p, id)) return NODE_DASGN; + if (local_id(p, id)) return NODE_LASGN; + dyna_var(p, id); + return NODE_DASGN; + } + else { + if (!local_id(p, id)) local_var(p, id); + return NODE_LASGN; + } + break; + case ID_GLOBAL: return NODE_GASGN; + case ID_INSTANCE: return NODE_IASGN; + case ID_CONST: + if (!p->ctxt.in_def) return NODE_CDECL; + *err = "dynamic constant assignment"; + return -1; + case ID_CLASS: return NODE_CVASGN; + default: + compile_error(p, "identifier %"PRIsVALUE" is not valid to set", rb_id2str(id)); } - else if (is_class_id(id)) { - if (in_def || in_single) return NEW_CVASGN(id, val); - return NEW_CVDECL(id, val); + return -1; +} + +#ifndef RIPPER +static NODE* +assignable(struct parser_params *p, ID id, NODE *val, const YYLTYPE *loc) +{ + const char *err = 0; + int node_type = assignable0(p, id, &err); + switch (node_type) { + case NODE_DASGN: return NEW_DASGN(id, val, loc); + case NODE_LASGN: return NEW_LASGN(id, val, loc); + case NODE_GASGN: return NEW_GASGN(id, val, loc); + case NODE_IASGN: return NEW_IASGN(id, val, loc); + case NODE_CDECL: return NEW_CDECL(id, val, 0, loc); + case NODE_CVASGN: return NEW_CVASGN(id, val, loc); + } + if (err) yyerror1(loc, err); + return NEW_BEGIN(0, loc); +} +#else +static VALUE +assignable(struct parser_params *p, VALUE lhs) +{ + const char *err = 0; + assignable0(p, get_id(lhs), &err); + if (err) lhs = assign_error(p, err, lhs); + return lhs; +} +#endif + +static int +is_private_local_id(ID name) +{ + VALUE s; + if (name == idUScore) return 1; + if (!is_local_id(name)) return 0; + s = rb_id2str(name); + if (!s) return 0; + return RSTRING_PTR(s)[0] == '_'; +} + +static int +shadowing_lvar_0(struct parser_params *p, ID name) +{ + if (dyna_in_block(p)) { + if (dvar_curr(p, name)) { + if (is_private_local_id(name)) return 1; + yyerror0("duplicated argument name"); + } + else if (dvar_defined(p, name) || local_id(p, name)) { + vtable_add(p->lvtbl->vars, name); + if (p->lvtbl->used) { + vtable_add(p->lvtbl->used, (ID)p->ruby_sourceline | LVAR_USED); + } + return 0; + } } else { - rb_compile_error("identifier %s is not valid", rb_id2name(id)); + if (local_id(p, name)) { + if (is_private_local_id(name)) return 1; + yyerror0("duplicated argument name"); + } } - return 0; + return 1; +} + +static ID +shadowing_lvar(struct parser_params *p, ID name) +{ + shadowing_lvar_0(p, name); + return name; } +static void +new_bv(struct parser_params *p, ID name) +{ + if (!name) return; + if (!is_local_id(name)) { + compile_error(p, "invalid local variable - %"PRIsVALUE, + rb_id2str(name)); + return; + } + if (!shadowing_lvar_0(p, name)) return; + dyna_var(p, name); +} + +#ifndef RIPPER static NODE * -aryset(recv, idx) - NODE *recv, *idx; +aryset(struct parser_params *p, NODE *recv, NODE *idx, const YYLTYPE *loc) { - if (recv && nd_type(recv) == NODE_SELF) - recv = (NODE *)1; - else - value_expr(recv); - return NEW_ATTRASGN(recv, tASET, idx); + return NEW_ATTRASGN(recv, tASET, idx, loc); } -ID -rb_id_attrset(id) - ID id; +static void +block_dup_check(struct parser_params *p, NODE *node1, NODE *node2) { - id &= ~ID_SCOPE_MASK; - id |= ID_ATTRSET; - return id; + if (node2 && node1 && nd_type_p(node1, NODE_BLOCK_PASS)) { + compile_error(p, "both block arg and actual block given"); + } } static NODE * -attrset(recv, id) - NODE *recv; - ID id; +attrset(struct parser_params *p, NODE *recv, ID atype, ID id, const YYLTYPE *loc) { - if (recv && nd_type(recv) == NODE_SELF) - recv = (NODE *)1; - else - value_expr(recv); - return NEW_ATTRASGN(recv, rb_id_attrset(id), 0); + if (!CALL_Q_P(atype)) id = rb_id_attrset(id); + return NEW_ATTRASGN(recv, id, 0, loc); } static void -rb_backref_error(node) - NODE *node; +rb_backref_error(struct parser_params *p, NODE *node) { switch (nd_type(node)) { case NODE_NTH_REF: - rb_compile_error("Can't set variable $%d", node->nd_nth); + compile_error(p, "Can't set variable $%ld", node->nd_nth); break; case NODE_BACK_REF: - rb_compile_error("Can't set variable $%c", (int)node->nd_nth); + compile_error(p, "Can't set variable $%c", (int)node->nd_nth); break; } } +#else +static VALUE +backref_error(struct parser_params *p, NODE *ref, VALUE expr) +{ + VALUE mesg = rb_str_new_cstr("Can't set variable "); + rb_str_append(mesg, ref->nd_cval); + return dispatch2(assign_error, mesg, expr); +} +#endif + +#ifndef RIPPER +static NODE * +arg_append(struct parser_params *p, NODE *node1, NODE *node2, const YYLTYPE *loc) +{ + if (!node1) return NEW_LIST(node2, &node2->nd_loc); + switch (nd_type(node1)) { + case NODE_LIST: + return list_append(p, node1, node2); + case NODE_BLOCK_PASS: + node1->nd_head = arg_append(p, node1->nd_head, node2, loc); + node1->nd_loc.end_pos = node1->nd_head->nd_loc.end_pos; + return node1; + case NODE_ARGSPUSH: + node1->nd_body = list_append(p, NEW_LIST(node1->nd_body, &node1->nd_body->nd_loc), node2); + node1->nd_loc.end_pos = node1->nd_body->nd_loc.end_pos; + nd_set_type(node1, NODE_ARGSCAT); + return node1; + case NODE_ARGSCAT: + if (!nd_type_p(node1->nd_body, NODE_LIST)) break; + node1->nd_body = list_append(p, node1->nd_body, node2); + node1->nd_loc.end_pos = node1->nd_body->nd_loc.end_pos; + return node1; + } + return NEW_ARGSPUSH(node1, node2, loc); +} static NODE * -arg_concat(node1, node2) - NODE *node1; - NODE *node2; +arg_concat(struct parser_params *p, NODE *node1, NODE *node2, const YYLTYPE *loc) { if (!node2) return node1; - return NEW_ARGSCAT(node1, node2); + switch (nd_type(node1)) { + case NODE_BLOCK_PASS: + if (node1->nd_head) + node1->nd_head = arg_concat(p, node1->nd_head, node2, loc); + else + node1->nd_head = NEW_LIST(node2, loc); + return node1; + case NODE_ARGSPUSH: + if (!nd_type_p(node2, NODE_LIST)) break; + node1->nd_body = list_concat(NEW_LIST(node1->nd_body, loc), node2); + nd_set_type(node1, NODE_ARGSCAT); + return node1; + case NODE_ARGSCAT: + if (!nd_type_p(node2, NODE_LIST) || + !nd_type_p(node1->nd_body, NODE_LIST)) break; + node1->nd_body = list_concat(node1->nd_body, node2); + return node1; + } + return NEW_ARGSCAT(node1, node2, loc); +} + +static NODE * +last_arg_append(struct parser_params *p, NODE *args, NODE *last_arg, const YYLTYPE *loc) +{ + NODE *n1; + if ((n1 = splat_array(args)) != 0) { + return list_append(p, n1, last_arg); + } + return arg_append(p, args, last_arg, loc); +} + +static NODE * +rest_arg_append(struct parser_params *p, NODE *args, NODE *rest_arg, const YYLTYPE *loc) +{ + NODE *n1; + if ((nd_type_p(rest_arg, NODE_LIST)) && (n1 = splat_array(args)) != 0) { + return list_concat(n1, rest_arg); + } + return arg_concat(p, args, rest_arg, loc); +} + +static NODE * +splat_array(NODE* node) +{ + if (nd_type_p(node, NODE_SPLAT)) node = node->nd_head; + if (nd_type_p(node, NODE_LIST)) return node; + return 0; +} + +static void +mark_lvar_used(struct parser_params *p, NODE *rhs) +{ + ID *vidp = NULL; + if (!rhs) return; + switch (nd_type(rhs)) { + case NODE_LASGN: + if (local_id_ref(p, rhs->nd_vid, &vidp)) { + if (vidp) *vidp |= LVAR_USED; + } + break; + case NODE_DASGN: + if (dvar_defined_ref(p, rhs->nd_vid, &vidp)) { + if (vidp) *vidp |= LVAR_USED; + } + break; +#if 0 + case NODE_MASGN: + for (rhs = rhs->nd_head; rhs; rhs = rhs->nd_next) { + mark_lvar_used(p, rhs->nd_head); + } + break; +#endif + } +} + +static NODE * +const_decl_path(struct parser_params *p, NODE **dest) +{ + NODE *n = *dest; + if (!nd_type_p(n, NODE_CALL)) { + const YYLTYPE *loc = &n->nd_loc; + VALUE path; + if (n->nd_vid) { + path = rb_id2str(n->nd_vid); + } + else { + n = n->nd_else; + path = rb_ary_new(); + for (; n && nd_type_p(n, NODE_COLON2); n = n->nd_head) { + rb_ary_push(path, rb_id2str(n->nd_mid)); + } + if (n && nd_type_p(n, NODE_CONST)) { + // Const::Name + rb_ary_push(path, rb_id2str(n->nd_vid)); + } + else if (n && nd_type_p(n, NODE_COLON3)) { + // ::Const::Name + rb_ary_push(path, rb_str_new(0, 0)); + } + else { + // expression::Name + rb_ary_push(path, rb_str_new_cstr("...")); + } + path = rb_ary_join(rb_ary_reverse(path), rb_str_new_cstr("::")); + path = rb_fstring(path); + } + *dest = n = NEW_LIT(path, loc); + RB_OBJ_WRITTEN(p->ast, Qnil, n->nd_lit); + } + return n; } +extern VALUE rb_mRubyVMFrozenCore; + static NODE * -arg_add(node1, node2) - NODE *node1; - NODE *node2; +make_shareable_node(struct parser_params *p, NODE *value, bool copy, const YYLTYPE *loc) { - if (!node1) return NEW_LIST(node2); - if (nd_type(node1) == NODE_ARRAY) { - return list_append(node1, node2); + NODE *fcore = NEW_LIT(rb_mRubyVMFrozenCore, loc); + + if (copy) { + return NEW_CALL(fcore, rb_intern("make_shareable_copy"), + NEW_LIST(value, loc), loc); } else { - return NEW_ARGSPUSH(node1, node2); + return NEW_CALL(fcore, rb_intern("make_shareable"), + NEW_LIST(value, loc), loc); } } -static NODE* -node_assign(lhs, rhs) - NODE *lhs, *rhs; +static NODE * +ensure_shareable_node(struct parser_params *p, NODE **dest, NODE *value, const YYLTYPE *loc) +{ + NODE *fcore = NEW_LIT(rb_mRubyVMFrozenCore, loc); + NODE *args = NEW_LIST(value, loc); + args = list_append(p, args, const_decl_path(p, dest)); + return NEW_CALL(fcore, rb_intern("ensure_shareable"), args, loc); +} + +static int is_static_content(NODE *node); + +static VALUE +shareable_literal_value(NODE *node) +{ + if (!node) return Qnil; + enum node_type type = nd_type(node); + switch (type) { + case NODE_TRUE: + return Qtrue; + case NODE_FALSE: + return Qfalse; + case NODE_NIL: + return Qnil; + case NODE_LIT: + return node->nd_lit; + default: + return Qundef; + } +} + +#ifndef SHAREABLE_BARE_EXPRESSION +#define SHAREABLE_BARE_EXPRESSION 1 +#endif + +static NODE * +shareable_literal_constant(struct parser_params *p, enum shareability shareable, + NODE **dest, NODE *value, const YYLTYPE *loc, size_t level) +{ +# define shareable_literal_constant_next(n) \ + shareable_literal_constant(p, shareable, dest, (n), &(n)->nd_loc, level+1) + VALUE lit = Qnil; + + if (!value) return 0; + enum node_type type = nd_type(value); + switch (type) { + case NODE_TRUE: + case NODE_FALSE: + case NODE_NIL: + case NODE_LIT: + return value; + + case NODE_DSTR: + if (shareable == shareable_literal) { + value = NEW_CALL(value, idUMinus, 0, loc); + } + return value; + + case NODE_STR: + lit = rb_fstring(value->nd_lit); + nd_set_type(value, NODE_LIT); + RB_OBJ_WRITE(p->ast, &value->nd_lit, lit); + return value; + + case NODE_ZLIST: + lit = rb_ary_new(); + OBJ_FREEZE_RAW(lit); + NODE *n = NEW_LIT(lit, loc); + RB_OBJ_WRITTEN(p->ast, Qnil, n->nd_lit); + return n; + + case NODE_LIST: + lit = rb_ary_new(); + for (NODE *n = value; n; n = n->nd_next) { + NODE *elt = n->nd_head; + if (elt) { + elt = shareable_literal_constant_next(elt); + if (elt) { + n->nd_head = elt; + } + else if (RTEST(lit)) { + rb_ary_clear(lit); + lit = Qfalse; + } + } + if (RTEST(lit)) { + VALUE e = shareable_literal_value(elt); + if (e != Qundef) { + rb_ary_push(lit, e); + } + else { + rb_ary_clear(lit); + lit = Qnil; /* make shareable at runtime */ + } + } + } + break; + + case NODE_HASH: + if (!value->nd_brace) return 0; + lit = rb_hash_new(); + for (NODE *n = value->nd_head; n; n = n->nd_next->nd_next) { + NODE *key = n->nd_head; + NODE *val = n->nd_next->nd_head; + if (key) { + key = shareable_literal_constant_next(key); + if (key) { + n->nd_head = key; + } + else if (RTEST(lit)) { + rb_hash_clear(lit); + lit = Qfalse; + } + } + if (val) { + val = shareable_literal_constant_next(val); + if (val) { + n->nd_next->nd_head = val; + } + else if (RTEST(lit)) { + rb_hash_clear(lit); + lit = Qfalse; + } + } + if (RTEST(lit)) { + VALUE k = shareable_literal_value(key); + VALUE v = shareable_literal_value(val); + if (k != Qundef && v != Qundef) { + rb_hash_aset(lit, k, v); + } + else { + rb_hash_clear(lit); + lit = Qnil; /* make shareable at runtime */ + } + } + } + break; + + default: + if (shareable == shareable_literal && + (SHAREABLE_BARE_EXPRESSION || level > 0)) { + return ensure_shareable_node(p, dest, value, loc); + } + return 0; + } + + /* Array or Hash */ + if (!lit) return 0; + if (NIL_P(lit)) { + // if shareable_literal, all elements should have been ensured + // as shareable + value = make_shareable_node(p, value, false, loc); + } + else { + value = NEW_LIT(rb_ractor_make_shareable(lit), loc); + RB_OBJ_WRITTEN(p->ast, Qnil, value->nd_lit); + } + + return value; +# undef shareable_literal_constant_next +} + +static NODE * +shareable_constant_value(struct parser_params *p, enum shareability shareable, + NODE *lhs, NODE *value, const YYLTYPE *loc) +{ + if (!value) return 0; + switch (shareable) { + case shareable_none: + return value; + + case shareable_literal: + { + NODE *lit = shareable_literal_constant(p, shareable, &lhs, value, loc, 0); + if (lit) return lit; + return value; + } + break; + + case shareable_copy: + case shareable_everything: + { + NODE *lit = shareable_literal_constant(p, shareable, &lhs, value, loc, 0); + if (lit) return lit; + return make_shareable_node(p, value, shareable == shareable_copy, loc); + } + break; + + default: + UNREACHABLE_RETURN(0); + } +} + +static NODE * +node_assign(struct parser_params *p, NODE *lhs, NODE *rhs, struct lex_context ctxt, const YYLTYPE *loc) { if (!lhs) return 0; - value_expr(rhs); switch (nd_type(lhs)) { + case NODE_CDECL: + rhs = shareable_constant_value(p, ctxt.shareable_constant_value, lhs, rhs, loc); + /* fallthru */ + case NODE_GASGN: case NODE_IASGN: case NODE_LASGN: case NODE_DASGN: - case NODE_DASGN_CURR: case NODE_MASGN: - case NODE_CDECL: - case NODE_CVDECL: case NODE_CVASGN: lhs->nd_value = rhs; + nd_set_loc(lhs, loc); break; case NODE_ATTRASGN: - case NODE_CALL: - lhs->nd_args = arg_add(lhs->nd_args, rhs); + lhs->nd_args = arg_append(p, lhs->nd_args, rhs, loc); + nd_set_loc(lhs, loc); break; default: @@ -5029,27 +11460,33 @@ node_assign(lhs, rhs) return lhs; } -static int -value_expr0(node) - NODE *node; +static NODE * +value_expr_check(struct parser_params *p, NODE *node) { - int cond = 0; + NODE *void_node = 0, *vn; + if (!node) { + rb_warning0("empty expression"); + } while (node) { switch (nd_type(node)) { - case NODE_DEFN: - case NODE_DEFS: - parser_warning(node, "void value expression"); - return Qfalse; - case NODE_RETURN: case NODE_BREAK: case NODE_NEXT: case NODE_REDO: case NODE_RETRY: - if (!cond) yyerror("void value expression"); - /* or "control never reach"? */ - return Qfalse; + return void_node ? void_node : node; + + case NODE_CASE3: + if (!node->nd_body || !nd_type_p(node->nd_body, NODE_IN)) { + compile_error(p, "unexpected node"); + return NULL; + } + if (node->nd_body->nd_body) { + return NULL; + } + /* single line pattern matching */ + return void_node ? void_node : node; case NODE_BLOCK: while (node->nd_next) { @@ -5063,44 +11500,59 @@ value_expr0(node) break; case NODE_IF: - if (!value_expr(node->nd_body)) return Qfalse; + case NODE_UNLESS: + if (!node->nd_body) { + return NULL; + } + else if (!node->nd_else) { + return NULL; + } + vn = value_expr_check(p, node->nd_body); + if (!vn) return NULL; + if (!void_node) void_node = vn; node = node->nd_else; break; case NODE_AND: case NODE_OR: - cond = 1; - node = node->nd_2nd; + node = node->nd_1st; break; - case NODE_NEWLINE: - node = node->nd_next; - break; + case NODE_LASGN: + case NODE_DASGN: + case NODE_MASGN: + mark_lvar_used(p, node); + return NULL; default: - return Qtrue; + return NULL; } } - return Qtrue; + return NULL; } +static int +value_expr_gen(struct parser_params *p, NODE *node) +{ + NODE *void_node = value_expr_check(p, node); + if (void_node) { + yyerror1(&void_node->nd_loc, "void value expression"); + /* or "control never reach"? */ + return FALSE; + } + return TRUE; +} static void -void_expr0(node) - NODE *node; +void_expr(struct parser_params *p, NODE *node) { - char *useless = 0; + const char *useless = 0; if (!RTEST(ruby_verbose)) return; - again: - if (!node) return; + if (!node || !(node = nd_once_body(node))) return; switch (nd_type(node)) { - case NODE_NEWLINE: - node = node->nd_next; - goto again; - - case NODE_CALL: + case NODE_OPCALL: switch (node->nd_mid) { case '+': case '-': @@ -5135,14 +11587,12 @@ void_expr0(node) useless = "a variable"; break; case NODE_CONST: - case NODE_CREF: useless = "a constant"; break; case NODE_LIT: case NODE_STR: case NODE_DSTR: case NODE_DREGX: - case NODE_DREGX_ONCE: useless = "a literal"; break; case NODE_COLON2: @@ -5173,221 +11623,238 @@ void_expr0(node) } if (useless) { - int line = ruby_sourceline; - - ruby_sourceline = nd_line(node); - rb_warn("useless use of %s in void context", useless); - ruby_sourceline = line; + rb_warn1L(nd_line(node), "possibly useless use of %s in void context", WARN_S(useless)); } } -static void -void_stmts(node) - NODE *node; +static NODE * +void_stmts(struct parser_params *p, NODE *node) { - if (!RTEST(ruby_verbose)) return; - if (!node) return; - if (nd_type(node) != NODE_BLOCK) return; + NODE *const n = node; + if (!RTEST(ruby_verbose)) return n; + if (!node) return n; + if (!nd_type_p(node, NODE_BLOCK)) return n; - for (;;) { - if (!node->nd_next) return; - void_expr(node->nd_head); + while (node->nd_next) { + void_expr(p, node->nd_head); node = node->nd_next; } + return n; } static NODE * -remove_begin(node) - NODE *node; +remove_begin(NODE *node) +{ + NODE **n = &node, *n1 = node; + while (n1 && nd_type_p(n1, NODE_BEGIN) && n1->nd_body) { + *n = n1 = n1->nd_body; + } + return node; +} + +static NODE * +remove_begin_all(NODE *node) +{ + NODE **n = &node, *n1 = node; + while (n1 && nd_type_p(n1, NODE_BEGIN)) { + *n = n1 = n1->nd_body; + } + return node; +} + +static void +reduce_nodes(struct parser_params *p, NODE **body) { - NODE **n = &node; - while (*n) { - switch (nd_type(*n)) { - case NODE_NEWLINE: - n = &(*n)->nd_next; + NODE *node = *body; + + if (!node) { + *body = NEW_NIL(&NULL_LOC); + return; + } +#define subnodes(n1, n2) \ + ((!node->n1) ? (node->n2 ? (body = &node->n2, 1) : 0) : \ + (!node->n2) ? (body = &node->n1, 1) : \ + (reduce_nodes(p, &node->n1), body = &node->n2, 1)) + + while (node) { + int newline = (int)(node->flags & NODE_FL_NEWLINE); + switch (nd_type(node)) { + end: + case NODE_NIL: + *body = 0; + return; + case NODE_RETURN: + *body = node = node->nd_stts; + if (newline && node) node->flags |= NODE_FL_NEWLINE; continue; case NODE_BEGIN: - *n = (*n)->nd_body; + *body = node = node->nd_body; + if (newline && node) node->flags |= NODE_FL_NEWLINE; + continue; + case NODE_BLOCK: + body = &node->nd_end->nd_head; + break; + case NODE_IF: + case NODE_UNLESS: + if (subnodes(nd_body, nd_else)) break; + return; + case NODE_CASE: + body = &node->nd_body; + break; + case NODE_WHEN: + if (!subnodes(nd_body, nd_next)) goto end; + break; + case NODE_ENSURE: + if (!subnodes(nd_head, nd_resq)) goto end; + break; + case NODE_RESCUE: + if (node->nd_else) { + body = &node->nd_resq; + break; + } + if (!subnodes(nd_head, nd_resq)) goto end; + break; default: - return node; + return; } + node = *body; + if (newline && node) node->flags |= NODE_FL_NEWLINE; } - return node; + +#undef subnodes } static int -assign_in_cond(node) - NODE *node; +is_static_content(NODE *node) { + if (!node) return 1; switch (nd_type(node)) { - case NODE_MASGN: - yyerror("multiple assignment in conditional"); - return 1; + case NODE_HASH: + if (!(node = node->nd_head)) break; + case NODE_LIST: + do { + if (!is_static_content(node->nd_head)) return 0; + } while ((node = node->nd_next) != 0); + case NODE_LIT: + case NODE_STR: + case NODE_NIL: + case NODE_TRUE: + case NODE_FALSE: + case NODE_ZLIST: + break; + default: + return 0; + } + return 1; +} +static int +assign_in_cond(struct parser_params *p, NODE *node) +{ + switch (nd_type(node)) { + case NODE_MASGN: case NODE_LASGN: case NODE_DASGN: case NODE_GASGN: case NODE_IASGN: break; - case NODE_NEWLINE: default: return 0; } - switch (nd_type(node->nd_value)) { - case NODE_LIT: - case NODE_STR: - case NODE_NIL: - case NODE_TRUE: - case NODE_FALSE: + if (!node->nd_value) return 1; + if (is_static_content(node->nd_value)) { /* reports always */ - parser_warn(node->nd_value, "found = in conditional, should be =="); - return 1; - - case NODE_DSTR: - case NODE_XSTR: - case NODE_DXSTR: - case NODE_EVSTR: - case NODE_DREGX: - default: - break; + parser_warn(p, node->nd_value, "found `= literal' in conditional, should be =="); } -#if 0 - if (assign_in_cond(node->nd_value) == 0) { - parser_warning(node->nd_value, "assignment in condition"); - } -#endif return 1; } -static int -e_option_supplied() -{ - if (strcmp(ruby_sourcefile, "-e") == 0) - return Qtrue; - return Qfalse; -} - -static void -warn_unless_e_option(node, str) - NODE *node; - const char *str; -{ - if (!e_option_supplied()) parser_warn(node, str); -} +enum cond_type { + COND_IN_OP, + COND_IN_COND, + COND_IN_FF +}; -static void -warning_unless_e_option(node, str) - NODE *node; - const char *str; -{ - if (!e_option_supplied()) parser_warning(node, str); -} +#define SWITCH_BY_COND_TYPE(t, w, arg) \ + switch (t) { \ + case COND_IN_OP: break; \ + case COND_IN_COND: rb_##w##0(arg "literal in condition"); break; \ + case COND_IN_FF: rb_##w##0(arg "literal in flip-flop"); break; \ + } -static NODE *cond0(); +static NODE *cond0(struct parser_params*,NODE*,enum cond_type,const YYLTYPE*); static NODE* -range_op(node) - NODE *node; +range_op(struct parser_params *p, NODE *node, const YYLTYPE *loc) { enum node_type type; - if (!e_option_supplied()) return node; if (node == 0) return 0; - value_expr(node); - node = cond0(node); type = nd_type(node); - if (type == NODE_NEWLINE) { - node = node->nd_next; - type = nd_type(node); - } + value_expr(node); if (type == NODE_LIT && FIXNUM_P(node->nd_lit)) { - warn_unless_e_option(node, "integer literal in conditional range"); - return call_op(node,tEQ,1,NEW_GVAR(rb_intern("$."))); - } - return node; -} - -static int -literal_node(node) - NODE *node; -{ - if (!node) return 1; /* same as NODE_NIL */ - switch (nd_type(node)) { - case NODE_LIT: - case NODE_STR: - case NODE_DSTR: - case NODE_EVSTR: - case NODE_DREGX: - case NODE_DREGX_ONCE: - case NODE_DSYM: - return 2; - case NODE_TRUE: - case NODE_FALSE: - case NODE_NIL: - return 1; + if (!e_option_supplied(p)) parser_warn(p, node, "integer literal in flip-flop"); + ID lineno = rb_intern("$."); + return NEW_CALL(node, tEQ, NEW_LIST(NEW_GVAR(lineno, loc), loc), loc); } - return 0; + return cond0(p, node, COND_IN_FF, loc); } static NODE* -cond0(node) - NODE *node; +cond0(struct parser_params *p, NODE *node, enum cond_type type, const YYLTYPE *loc) { if (node == 0) return 0; - assign_in_cond(node); + if (!(node = nd_once_body(node))) return 0; + assign_in_cond(p, node); switch (nd_type(node)) { case NODE_DSTR: case NODE_EVSTR: case NODE_STR: - rb_warn("string literal in condition"); + SWITCH_BY_COND_TYPE(type, warn, "string ") break; case NODE_DREGX: - case NODE_DREGX_ONCE: - warning_unless_e_option(node, "regex literal in condition"); - local_cnt('_'); - local_cnt('~'); - return NEW_MATCH2(node, NEW_GVAR(rb_intern("$_"))); + if (!e_option_supplied(p)) SWITCH_BY_COND_TYPE(type, warning, "regex ") + + return NEW_MATCH2(node, NEW_GVAR(idLASTLINE, loc), loc); case NODE_AND: case NODE_OR: - node->nd_1st = cond0(node->nd_1st); - node->nd_2nd = cond0(node->nd_2nd); + node->nd_1st = cond0(p, node->nd_1st, COND_IN_COND, loc); + node->nd_2nd = cond0(p, node->nd_2nd, COND_IN_COND, loc); break; case NODE_DOT2: case NODE_DOT3: - node->nd_beg = range_op(node->nd_beg); - node->nd_end = range_op(node->nd_end); - if (nd_type(node) == NODE_DOT2) nd_set_type(node,NODE_FLIP2); - else if (nd_type(node) == NODE_DOT3) nd_set_type(node, NODE_FLIP3); - node->nd_cnt = local_append(internal_id()); - if (!e_option_supplied()) { - int b = literal_node(node->nd_beg); - int e = literal_node(node->nd_end); - if ((b == 1 && e == 1) || (b + e >= 2 && RTEST(ruby_verbose))) { - parser_warn(node, "range literal in condition"); - } - } + node->nd_beg = range_op(p, node->nd_beg, loc); + node->nd_end = range_op(p, node->nd_end, loc); + if (nd_type_p(node, NODE_DOT2)) nd_set_type(node,NODE_FLIP2); + else if (nd_type_p(node, NODE_DOT3)) nd_set_type(node, NODE_FLIP3); break; case NODE_DSYM: - parser_warning(node, "literal in condition"); + warn_symbol: + SWITCH_BY_COND_TYPE(type, warning, "symbol ") break; case NODE_LIT: - if (TYPE(node->nd_lit) == T_REGEXP) { - warn_unless_e_option(node, "regex literal in condition"); + if (RB_TYPE_P(node->nd_lit, T_REGEXP)) { + if (!e_option_supplied(p)) SWITCH_BY_COND_TYPE(type, warn, "regex ") nd_set_type(node, NODE_MATCH); - local_cnt('_'); - local_cnt('~'); + } + else if (node->nd_lit == Qtrue || + node->nd_lit == Qfalse) { + /* booleans are OK, e.g., while true */ + } + else if (SYMBOL_P(node->nd_lit)) { + goto warn_symbol; } else { - parser_warning(node, "literal in condition"); + SWITCH_BY_COND_TYPE(type, warning, "") } default: break; @@ -5396,950 +11863,2206 @@ cond0(node) } static NODE* -cond(node) - NODE *node; +cond(struct parser_params *p, NODE *node, const YYLTYPE *loc) { if (node == 0) return 0; - value_expr(node); - if (nd_type(node) == NODE_NEWLINE){ - node->nd_next = cond0(node->nd_next); - return node; - } - return cond0(node); + return cond0(p, node, COND_IN_COND, loc); } static NODE* -logop(type, left, right) - enum node_type type; - NODE *left, *right; +method_cond(struct parser_params *p, NODE *node, const YYLTYPE *loc) { - value_expr(left); - if (left && nd_type(left) == type) { - NODE *node = left, *second; - while ((second = node->nd_2nd) != 0 && nd_type(second) == type) { - node = second; - } - node->nd_2nd = NEW_NODE(type, second, right, 0); - return left; - } - return NEW_NODE(type, left, right, 0); + if (node == 0) return 0; + return cond0(p, node, COND_IN_OP, loc); } -static int -cond_negative(nodep) - NODE **nodep; +static NODE* +new_nil_at(struct parser_params *p, const rb_code_position_t *pos) { - NODE *c = *nodep; + YYLTYPE loc = {*pos, *pos}; + return NEW_NIL(&loc); +} - if (!c) return 0; - switch (nd_type(c)) { - case NODE_NOT: - *nodep = c->nd_body; - return 1; - case NODE_NEWLINE: - if (c->nd_next && nd_type(c->nd_next) == NODE_NOT) { - c->nd_next = c->nd_next->nd_body; - return 1; +static NODE* +new_if(struct parser_params *p, NODE *cc, NODE *left, NODE *right, const YYLTYPE *loc) +{ + if (!cc) return right; + cc = cond0(p, cc, COND_IN_COND, loc); + return newline_node(NEW_IF(cc, left, right, loc)); +} + +static NODE* +new_unless(struct parser_params *p, NODE *cc, NODE *left, NODE *right, const YYLTYPE *loc) +{ + if (!cc) return right; + cc = cond0(p, cc, COND_IN_COND, loc); + return newline_node(NEW_UNLESS(cc, left, right, loc)); +} + +static NODE* +logop(struct parser_params *p, ID id, NODE *left, NODE *right, + const YYLTYPE *op_loc, const YYLTYPE *loc) +{ + enum node_type type = id == idAND || id == idANDOP ? NODE_AND : NODE_OR; + NODE *op; + value_expr(left); + if (left && nd_type_p(left, type)) { + NODE *node = left, *second; + while ((second = node->nd_2nd) != 0 && nd_type_p(second, type)) { + node = second; } + node->nd_2nd = NEW_NODE(type, second, right, 0, loc); + nd_set_line(node->nd_2nd, op_loc->beg_pos.lineno); + left->nd_loc.end_pos = loc->end_pos; + return left; } - return 0; + op = NEW_NODE(type, left, right, 0, loc); + nd_set_line(op, op_loc->beg_pos.lineno); + return op; } static void -no_blockarg(node) - NODE *node; +no_blockarg(struct parser_params *p, NODE *node) { - if (node && nd_type(node) == NODE_BLOCK_PASS) { - rb_compile_error("block argument should not be given"); + if (node && nd_type_p(node, NODE_BLOCK_PASS)) { + compile_error(p, "block argument should not be given"); } } static NODE * -ret_args(node) - NODE *node; +ret_args(struct parser_params *p, NODE *node) { if (node) { - no_blockarg(node); - if (nd_type(node) == NODE_ARRAY && node->nd_next == 0) { - node = node->nd_head; - } - if (node && nd_type(node) == NODE_SPLAT) { - node = NEW_SVALUE(node); + no_blockarg(p, node); + if (nd_type_p(node, NODE_LIST)) { + if (node->nd_next == 0) { + node = node->nd_head; + } + else { + nd_set_type(node, NODE_VALUES); + } } } return node; } static NODE * -new_yield(node) - NODE *node; +new_yield(struct parser_params *p, NODE *node, const YYLTYPE *loc) { - long state = Qtrue; + if (node) no_blockarg(p, node); - if (node) { - no_blockarg(node); - if (nd_type(node) == NODE_ARRAY && node->nd_next == 0) { - node = node->nd_head; - state = Qfalse; - } - if (node && nd_type(node) == NODE_SPLAT) { - state = Qtrue; - } - } - else { - state = Qfalse; - } - return NEW_YIELD(node, state); + return NEW_YIELD(node, loc); } -static NODE* -negate_lit(node) - NODE *node; +static VALUE +negate_lit(struct parser_params *p, VALUE lit) { - switch (TYPE(node->nd_lit)) { - case T_FIXNUM: - node->nd_lit = LONG2FIX(-FIX2LONG(node->nd_lit)); - break; + if (FIXNUM_P(lit)) { + return LONG2FIX(-FIX2LONG(lit)); + } + if (SPECIAL_CONST_P(lit)) { +#if USE_FLONUM + if (FLONUM_P(lit)) { + return DBL2NUM(-RFLOAT_VALUE(lit)); + } +#endif + goto unknown; + } + switch (BUILTIN_TYPE(lit)) { case T_BIGNUM: - node->nd_lit = rb_funcall(node->nd_lit,tUMINUS,0,0); + BIGNUM_NEGATE(lit); + lit = rb_big_norm(lit); + break; + case T_RATIONAL: + RATIONAL_SET_NUM(lit, negate_lit(p, RRATIONAL(lit)->num)); + break; + case T_COMPLEX: + RCOMPLEX_SET_REAL(lit, negate_lit(p, RCOMPLEX(lit)->real)); + RCOMPLEX_SET_IMAG(lit, negate_lit(p, RCOMPLEX(lit)->imag)); break; case T_FLOAT: - RFLOAT(node->nd_lit)->value = -RFLOAT(node->nd_lit)->value; + lit = DBL2NUM(-RFLOAT_VALUE(lit)); break; + unknown: default: + rb_parser_fatal(p, "unknown literal type (%s) passed to negate_lit", + rb_builtin_class_name(lit)); break; } - return node; + return lit; } static NODE * -arg_blk_pass(node1, node2) - NODE *node1; - NODE *node2; +arg_blk_pass(NODE *node1, NODE *node2) { if (node2) { + if (!node1) return node2; node2->nd_head = node1; + nd_set_first_lineno(node2, nd_first_lineno(node1)); + nd_set_first_column(node2, nd_first_column(node1)); return node2; } return node1; } +static bool +args_info_empty_p(struct rb_args_info *args) +{ + if (args->pre_args_num) return false; + if (args->post_args_num) return false; + if (args->rest_arg) return false; + if (args->opt_args) return false; + if (args->block_arg) return false; + if (args->kw_args) return false; + if (args->kw_rest_arg) return false; + return true; +} + static NODE* -arg_prepend(node1, node2) - NODE *node1, *node2; +new_args(struct parser_params *p, NODE *pre_args, NODE *opt_args, ID rest_arg, NODE *post_args, NODE *tail, const YYLTYPE *loc) { - switch (nd_type(node2)) { - case NODE_ARRAY: - return list_concat(NEW_LIST(node1), node2); + int saved_line = p->ruby_sourceline; + struct rb_args_info *args = tail->nd_ainfo; - case NODE_SPLAT: - return arg_concat(node1, node2->nd_head); + if (args->block_arg == idFWD_BLOCK) { + if (rest_arg) { + yyerror1(&tail->nd_loc, "... after rest argument"); + return tail; + } + rest_arg = idFWD_REST; + } - case NODE_BLOCK_PASS: - node2->nd_body = arg_prepend(node1, node2->nd_body); - return node2; + args->pre_args_num = pre_args ? rb_long2int(pre_args->nd_plen) : 0; + args->pre_init = pre_args ? pre_args->nd_next : 0; - default: - rb_bug("unknown nodetype(%d) for arg_prepend", nd_type(node2)); - } - return 0; /* not reached */ + args->post_args_num = post_args ? rb_long2int(post_args->nd_plen) : 0; + args->post_init = post_args ? post_args->nd_next : 0; + args->first_post_arg = post_args ? post_args->nd_pid : 0; + + args->rest_arg = rest_arg; + + args->opt_args = opt_args; + + args->ruby2_keywords = rest_arg == idFWD_REST; + + p->ruby_sourceline = saved_line; + nd_set_loc(tail, loc); + + return tail; } static NODE* -new_call(r,m,a) - NODE *r; - ID m; - NODE *a; +new_args_tail(struct parser_params *p, NODE *kw_args, ID kw_rest_arg, ID block, const YYLTYPE *kw_rest_loc) { - if (a && nd_type(a) == NODE_BLOCK_PASS) { - a->nd_iter = NEW_CALL(r,m,a->nd_head); - return a; + int saved_line = p->ruby_sourceline; + NODE *node; + VALUE tmpbuf = rb_imemo_tmpbuf_auto_free_pointer(); + struct rb_args_info *args = ZALLOC(struct rb_args_info); + rb_imemo_tmpbuf_set_ptr(tmpbuf, args); + args->imemo = tmpbuf; + node = NEW_NODE(NODE_ARGS, 0, 0, args, &NULL_LOC); + RB_OBJ_WRITTEN(p->ast, Qnil, tmpbuf); + if (p->error_p) return node; + + args->block_arg = block; + args->kw_args = kw_args; + + if (kw_args) { + /* + * def foo(k1: 1, kr1:, k2: 2, **krest, &b) + * variable order: k1, kr1, k2, &b, internal_id, krest + * #=> <reorder> + * variable order: kr1, k1, k2, internal_id, krest, &b + */ + ID kw_bits = internal_id(p), *required_kw_vars, *kw_vars; + struct vtable *vtargs = p->lvtbl->args; + NODE *kwn = kw_args; + + if (block) block = vtargs->tbl[vtargs->pos-1]; + vtable_pop(vtargs, !!block + !!kw_rest_arg); + required_kw_vars = kw_vars = &vtargs->tbl[vtargs->pos]; + while (kwn) { + if (!NODE_REQUIRED_KEYWORD_P(kwn->nd_body)) + --kw_vars; + --required_kw_vars; + kwn = kwn->nd_next; + } + + for (kwn = kw_args; kwn; kwn = kwn->nd_next) { + ID vid = kwn->nd_body->nd_vid; + if (NODE_REQUIRED_KEYWORD_P(kwn->nd_body)) { + *required_kw_vars++ = vid; + } + else { + *kw_vars++ = vid; + } + } + + arg_var(p, kw_bits); + if (kw_rest_arg) arg_var(p, kw_rest_arg); + if (block) arg_var(p, block); + + args->kw_rest_arg = NEW_DVAR(kw_rest_arg, kw_rest_loc); + args->kw_rest_arg->nd_cflag = kw_bits; + } + else if (kw_rest_arg == idNil) { + args->no_kwarg = 1; + } + else if (kw_rest_arg) { + args->kw_rest_arg = NEW_DVAR(kw_rest_arg, kw_rest_loc); } - return NEW_CALL(r,m,a); + + p->ruby_sourceline = saved_line; + return node; } -static NODE* -new_fcall(m,a) - ID m; - NODE *a; +static NODE * +args_with_numbered(struct parser_params *p, NODE *args, int max_numparam) { - if (a && nd_type(a) == NODE_BLOCK_PASS) { - a->nd_iter = NEW_FCALL(m,a->nd_head); - return a; + if (max_numparam > NO_PARAM) { + if (!args) { + YYLTYPE loc = RUBY_INIT_YYLLOC(); + args = new_args_tail(p, 0, 0, 0, 0); + nd_set_loc(args, &loc); + } + args->nd_ainfo->pre_args_num = max_numparam; } - return NEW_FCALL(m,a); + return args; } static NODE* -new_super(a) - NODE *a; +new_array_pattern(struct parser_params *p, NODE *constant, NODE *pre_arg, NODE *aryptn, const YYLTYPE *loc) { - if (a && nd_type(a) == NODE_BLOCK_PASS) { - a->nd_iter = NEW_SUPER(a->nd_head); - return a; + struct rb_ary_pattern_info *apinfo = aryptn->nd_apinfo; + + aryptn->nd_pconst = constant; + + if (pre_arg) { + NODE *pre_args = NEW_LIST(pre_arg, loc); + if (apinfo->pre_args) { + apinfo->pre_args = list_concat(pre_args, apinfo->pre_args); + } + else { + apinfo->pre_args = pre_args; + } } - return NEW_SUPER(a); + return aryptn; } -static struct local_vars { - ID *tbl; - int nofree; - int cnt; - int dlev; - struct RVarmap* dyna_vars; - struct local_vars *prev; -} *lvtbl; - -static void -local_push(top) - int top; +static NODE* +new_array_pattern_tail(struct parser_params *p, NODE *pre_args, int has_rest, ID rest_arg, NODE *post_args, const YYLTYPE *loc) { - struct local_vars *local; + int saved_line = p->ruby_sourceline; + NODE *node; + VALUE tmpbuf = rb_imemo_tmpbuf_auto_free_pointer(); + struct rb_ary_pattern_info *apinfo = ZALLOC(struct rb_ary_pattern_info); + rb_imemo_tmpbuf_set_ptr(tmpbuf, apinfo); + node = NEW_NODE(NODE_ARYPTN, 0, tmpbuf, apinfo, loc); + RB_OBJ_WRITTEN(p->ast, Qnil, tmpbuf); - local = ALLOC(struct local_vars); - local->prev = lvtbl; - local->nofree = 0; - local->cnt = 0; - local->tbl = 0; - local->dlev = 0; - local->dyna_vars = ruby_dyna_vars; - lvtbl = local; - if (!top) { - /* preserve reference for GC, but link should be cut. */ - rb_dvar_push(0, (VALUE)ruby_dyna_vars); - ruby_dyna_vars->next = 0; + apinfo->pre_args = pre_args; + + if (has_rest) { + if (rest_arg) { + apinfo->rest_arg = assignable(p, rest_arg, 0, loc); + } + else { + apinfo->rest_arg = NODE_SPECIAL_NO_NAME_REST; + } + } + else { + apinfo->rest_arg = NULL; } + + apinfo->post_args = post_args; + + p->ruby_sourceline = saved_line; + return node; } -static void -local_pop() +static NODE* +new_find_pattern(struct parser_params *p, NODE *constant, NODE *fndptn, const YYLTYPE *loc) { - struct local_vars *local = lvtbl->prev; + fndptn->nd_pconst = constant; - if (lvtbl->tbl) { - if (!lvtbl->nofree) xfree(lvtbl->tbl); - else lvtbl->tbl[0] = lvtbl->cnt; - } - ruby_dyna_vars = lvtbl->dyna_vars; - xfree(lvtbl); - lvtbl = local; + return fndptn; +} + +static NODE* +new_find_pattern_tail(struct parser_params *p, ID pre_rest_arg, NODE *args, ID post_rest_arg, const YYLTYPE *loc) +{ + int saved_line = p->ruby_sourceline; + NODE *node; + VALUE tmpbuf = rb_imemo_tmpbuf_auto_free_pointer(); + struct rb_fnd_pattern_info *fpinfo = ZALLOC(struct rb_fnd_pattern_info); + rb_imemo_tmpbuf_set_ptr(tmpbuf, fpinfo); + node = NEW_NODE(NODE_FNDPTN, 0, tmpbuf, fpinfo, loc); + RB_OBJ_WRITTEN(p->ast, Qnil, tmpbuf); + + fpinfo->pre_rest_arg = pre_rest_arg ? assignable(p, pre_rest_arg, 0, loc) : NODE_SPECIAL_NO_NAME_REST; + fpinfo->args = args; + fpinfo->post_rest_arg = post_rest_arg ? assignable(p, post_rest_arg, 0, loc) : NODE_SPECIAL_NO_NAME_REST; + + p->ruby_sourceline = saved_line; + return node; } -static ID* -local_tbl() +static NODE* +new_hash_pattern(struct parser_params *p, NODE *constant, NODE *hshptn, const YYLTYPE *loc) { - lvtbl->nofree = 1; - return lvtbl->tbl; + hshptn->nd_pconst = constant; + return hshptn; } -static int -local_append(id) - ID id; +static NODE* +new_hash_pattern_tail(struct parser_params *p, NODE *kw_args, ID kw_rest_arg, const YYLTYPE *loc) { - if (lvtbl->tbl == 0) { - lvtbl->tbl = ALLOC_N(ID, 4); - lvtbl->tbl[0] = 0; - lvtbl->tbl[1] = '_'; - lvtbl->tbl[2] = '~'; - lvtbl->cnt = 2; - if (id == '_') return 0; - if (id == '~') return 1; + int saved_line = p->ruby_sourceline; + NODE *node, *kw_rest_arg_node; + + if (kw_rest_arg == idNil) { + kw_rest_arg_node = NODE_SPECIAL_NO_REST_KEYWORD; + } + else if (kw_rest_arg) { + kw_rest_arg_node = assignable(p, kw_rest_arg, 0, loc); } else { - REALLOC_N(lvtbl->tbl, ID, lvtbl->cnt+2); + kw_rest_arg_node = NULL; } - lvtbl->tbl[lvtbl->cnt+1] = id; - return lvtbl->cnt++; + node = NEW_NODE(NODE_HSHPTN, 0, kw_args, kw_rest_arg_node, loc); + + p->ruby_sourceline = saved_line; + return node; } -static int -local_cnt(id) - ID id; +static NODE* +dsym_node(struct parser_params *p, NODE *node, const YYLTYPE *loc) { - int cnt, max; + VALUE lit; + + if (!node) { + return NEW_LIT(ID2SYM(idNULL), loc); + } - if (id == 0) return lvtbl->cnt; + switch (nd_type(node)) { + case NODE_DSTR: + nd_set_type(node, NODE_DSYM); + nd_set_loc(node, loc); + break; + case NODE_STR: + lit = node->nd_lit; + RB_OBJ_WRITTEN(p->ast, Qnil, node->nd_lit = ID2SYM(rb_intern_str(lit))); + nd_set_type(node, NODE_LIT); + nd_set_loc(node, loc); + break; + default: + node = NEW_NODE(NODE_DSYM, Qnil, 1, NEW_LIST(node, loc), loc); + break; + } + return node; +} + +static int +append_literal_keys(st_data_t k, st_data_t v, st_data_t h) +{ + NODE *node = (NODE *)v; + NODE **result = (NODE **)h; + node->nd_alen = 2; + node->nd_next->nd_end = node->nd_next; + node->nd_next->nd_next = 0; + if (*result) + list_concat(*result, node); + else + *result = node; + return ST_CONTINUE; +} - for (cnt=1, max=lvtbl->cnt+1; cnt<max;cnt++) { - if (lvtbl->tbl[cnt] == id) return cnt-1; +static bool +hash_literal_key_p(VALUE k) +{ + switch (OBJ_BUILTIN_TYPE(k)) { + case T_NODE: + return false; + default: + return true; } - return local_append(id); } static int -local_id(id) - ID id; +literal_cmp(VALUE val, VALUE lit) +{ + if (val == lit) return 0; + if (!hash_literal_key_p(val) || !hash_literal_key_p(lit)) return -1; + return rb_iseq_cdhash_cmp(val, lit); +} + +static st_index_t +literal_hash(VALUE a) { - int i, max; + if (!hash_literal_key_p(a)) return (st_index_t)a; + return rb_iseq_cdhash_hash(a); +} + +static const struct st_hash_type literal_type = { + literal_cmp, + literal_hash, +}; - if (lvtbl == 0) return Qfalse; - for (i=3, max=lvtbl->cnt+1; i<max; i++) { - if (lvtbl->tbl[i] == id) return Qtrue; +static NODE * +remove_duplicate_keys(struct parser_params *p, NODE *hash) +{ + st_table *literal_keys = st_init_table_with_size(&literal_type, hash->nd_alen / 2); + NODE *result = 0; + NODE *last_expr = 0; + rb_code_location_t loc = hash->nd_loc; + while (hash && hash->nd_head && hash->nd_next) { + NODE *head = hash->nd_head; + NODE *value = hash->nd_next; + NODE *next = value->nd_next; + st_data_t key = (st_data_t)head; + st_data_t data; + value->nd_next = 0; + if (nd_type_p(head, NODE_LIT) && + st_delete(literal_keys, (key = (st_data_t)head->nd_lit, &key), &data)) { + NODE *dup_value = ((NODE *)data)->nd_next; + rb_compile_warn(p->ruby_sourcefile, nd_line((NODE *)data), + "key %+"PRIsVALUE" is duplicated and overwritten on line %d", + head->nd_lit, nd_line(head)); + if (dup_value == last_expr) { + value->nd_head = block_append(p, dup_value->nd_head, value->nd_head); + } + else { + last_expr->nd_head = block_append(p, dup_value->nd_head, last_expr->nd_head); + } + } + st_insert(literal_keys, (st_data_t)key, (st_data_t)hash); + last_expr = nd_type_p(head, NODE_LIT) ? value : head; + hash = next; + } + st_foreach(literal_keys, append_literal_keys, (st_data_t)&result); + st_free_table(literal_keys); + if (hash) { + if (!result) result = hash; + else list_concat(result, hash); } - return Qfalse; + result->nd_loc = loc; + return result; } +static NODE * +new_hash(struct parser_params *p, NODE *hash, const YYLTYPE *loc) +{ + if (hash) hash = remove_duplicate_keys(p, hash); + return NEW_HASH(hash, loc); +} +#endif + static void -top_local_init() +error_duplicate_pattern_variable(struct parser_params *p, ID id, const YYLTYPE *loc) { - local_push(1); - lvtbl->cnt = ruby_scope->local_tbl?ruby_scope->local_tbl[0]:0; - if (lvtbl->cnt > 0) { - lvtbl->tbl = ALLOC_N(ID, lvtbl->cnt+3); - MEMCPY(lvtbl->tbl, ruby_scope->local_tbl, ID, lvtbl->cnt+1); + if (is_private_local_id(id)) { + return; + } + if (st_is_member(p->pvtbl, id)) { + yyerror1(loc, "duplicated variable name"); } else { - lvtbl->tbl = 0; + st_insert(p->pvtbl, (st_data_t)id, 0); } - if (ruby_dyna_vars) - lvtbl->dlev = 1; - else - lvtbl->dlev = 0; } static void -top_local_setup() +error_duplicate_pattern_key(struct parser_params *p, VALUE key, const YYLTYPE *loc) { - int len = lvtbl->cnt; - int i; + if (!p->pktbl) { + p->pktbl = st_init_numtable(); + } + else if (st_is_member(p->pktbl, key)) { + yyerror1(loc, "duplicated key name"); + return; + } + st_insert(p->pktbl, (st_data_t)key, 0); +} - if (len > 0) { - i = ruby_scope->local_tbl?ruby_scope->local_tbl[0]:0; +#ifndef RIPPER +static NODE * +new_unique_key_hash(struct parser_params *p, NODE *hash, const YYLTYPE *loc) +{ + return NEW_HASH(hash, loc); +} +#endif /* !RIPPER */ - if (i < len) { - if (i == 0 || (ruby_scope->flags & SCOPE_MALLOC) == 0) { - VALUE *vars = ALLOC_N(VALUE, len+1); - if (ruby_scope->local_vars) { - *vars++ = ruby_scope->local_vars[-1]; - MEMCPY(vars, ruby_scope->local_vars, VALUE, i); - rb_mem_clear(vars+i, len-i); - } - else { - *vars++ = (VALUE)ruby_scope; - rb_mem_clear(vars, len); +#ifndef RIPPER +static NODE * +new_op_assign(struct parser_params *p, NODE *lhs, ID op, NODE *rhs, struct lex_context ctxt, const YYLTYPE *loc) +{ + NODE *asgn; + + if (lhs) { + ID vid = lhs->nd_vid; + YYLTYPE lhs_loc = lhs->nd_loc; + int shareable = ctxt.shareable_constant_value; + if (shareable) { + switch (nd_type(lhs)) { + case NODE_CDECL: + case NODE_COLON2: + case NODE_COLON3: + break; + default: + shareable = 0; + break; + } + } + if (op == tOROP) { + rhs = shareable_constant_value(p, shareable, lhs, rhs, &rhs->nd_loc); + lhs->nd_value = rhs; + nd_set_loc(lhs, loc); + asgn = NEW_OP_ASGN_OR(gettable(p, vid, &lhs_loc), lhs, loc); + if (is_notop_id(vid)) { + switch (id_type(vid)) { + case ID_GLOBAL: + case ID_INSTANCE: + case ID_CLASS: + asgn->nd_aid = vid; } - ruby_scope->local_vars = vars; - ruby_scope->flags |= SCOPE_MALLOC; } - else { - VALUE *vars = ruby_scope->local_vars-1; - REALLOC_N(vars, VALUE, len+1); - ruby_scope->local_vars = vars+1; - rb_mem_clear(ruby_scope->local_vars+i, len-i); + } + else if (op == tANDOP) { + if (shareable) { + rhs = shareable_constant_value(p, shareable, lhs, rhs, &rhs->nd_loc); } - if (ruby_scope->local_tbl && ruby_scope->local_vars[-1] == 0) { - if (!(ruby_scope->flags & SCOPE_CLONE)) - xfree(ruby_scope->local_tbl); + lhs->nd_value = rhs; + nd_set_loc(lhs, loc); + asgn = NEW_OP_ASGN_AND(gettable(p, vid, &lhs_loc), lhs, loc); + } + else { + asgn = lhs; + rhs = NEW_CALL(gettable(p, vid, &lhs_loc), op, NEW_LIST(rhs, &rhs->nd_loc), loc); + if (shareable) { + rhs = shareable_constant_value(p, shareable, lhs, rhs, &rhs->nd_loc); } - ruby_scope->local_tbl = local_tbl(); + asgn->nd_value = rhs; + nd_set_loc(asgn, loc); } } - local_pop(); + else { + asgn = NEW_BEGIN(0, loc); + } + return asgn; } -#define DVAR_USED FL_USER6 - -static VALUE -dyna_var_lookup(id) - ID id; +static NODE * +new_ary_op_assign(struct parser_params *p, NODE *ary, + NODE *args, ID op, NODE *rhs, const YYLTYPE *args_loc, const YYLTYPE *loc) { - struct RVarmap *vars = ruby_dyna_vars; + NODE *asgn; - while (vars) { - if (vars->id == id) { - FL_SET(vars, DVAR_USED); - return Qtrue; - } - vars = vars->next; + args = make_list(args, args_loc); + if (nd_type_p(args, NODE_BLOCK_PASS)) { + args = NEW_ARGSCAT(args, rhs, loc); + } + else { + args = arg_concat(p, args, rhs, loc); } - return Qfalse; + asgn = NEW_OP_ASGN1(ary, op, args, loc); + fixpos(asgn, ary); + return asgn; } -static struct RVarmap* -dyna_push() +static NODE * +new_attr_op_assign(struct parser_params *p, NODE *lhs, + ID atype, ID attr, ID op, NODE *rhs, const YYLTYPE *loc) { - struct RVarmap* vars = ruby_dyna_vars; + NODE *asgn; - rb_dvar_push(0, 0); - lvtbl->dlev++; - return vars; + asgn = NEW_OP_ASGN2(lhs, CALL_Q_P(atype), attr, op, rhs, loc); + fixpos(asgn, lhs); + return asgn; } -static void -dyna_pop(vars) - struct RVarmap* vars; +static NODE * +new_const_op_assign(struct parser_params *p, NODE *lhs, ID op, NODE *rhs, struct lex_context ctxt, const YYLTYPE *loc) { - lvtbl->dlev--; - ruby_dyna_vars = vars; + NODE *asgn; + + if (lhs) { + rhs = shareable_constant_value(p, ctxt.shareable_constant_value, lhs, rhs, loc); + asgn = NEW_OP_CDECL(lhs, op, rhs, loc); + } + else { + asgn = NEW_BEGIN(0, loc); + } + fixpos(asgn, lhs); + return asgn; } -static int -dyna_in_block() +static NODE * +const_decl(struct parser_params *p, NODE *path, const YYLTYPE *loc) +{ + if (p->ctxt.in_def) { + yyerror1(loc, "dynamic constant assignment"); + } + return NEW_CDECL(0, 0, (path), loc); +} +#else +static VALUE +const_decl(struct parser_params *p, VALUE path) +{ + if (p->ctxt.in_def) { + path = assign_error(p, "dynamic constant assignment", path); + } + return path; +} + +static VALUE +assign_error(struct parser_params *p, const char *mesg, VALUE a) +{ + a = dispatch2(assign_error, ERR_MESG(), a); + ripper_error(p); + return a; +} + +static VALUE +var_field(struct parser_params *p, VALUE a) { - return (lvtbl->dlev > 0); + return ripper_new_yylval(p, get_id(a), dispatch1(var_field, a), 0); } +#endif +#ifndef RIPPER static NODE * -dyna_init(node, pre) - NODE *node; - struct RVarmap *pre; +new_bodystmt(struct parser_params *p, NODE *head, NODE *rescue, NODE *rescue_else, NODE *ensure, const YYLTYPE *loc) { - struct RVarmap *post = ruby_dyna_vars; - NODE *var; + NODE *result = head; + if (rescue) { + NODE *tmp = rescue_else ? rescue_else : rescue; + YYLTYPE rescue_loc = code_loc_gen(&head->nd_loc, &tmp->nd_loc); - if (!node || !post || pre == post) return node; - for (var = 0; post != pre && post->id; post = post->next) { - if (FL_TEST(post, DVAR_USED)) { - var = NEW_DASGN_CURR(post->id, var); - } + result = NEW_RESCUE(head, rescue, rescue_else, &rescue_loc); + nd_set_line(result, rescue->nd_loc.beg_pos.lineno); + } + else if (rescue_else) { + result = block_append(p, result, rescue_else); + } + if (ensure) { + result = NEW_ENSURE(result, ensure, loc); } - return block_append(var, node); + fixpos(result, head); + return result; } +#endif -int -ruby_parser_stack_on_heap() +static void +warn_unused_var(struct parser_params *p, struct local_vars *local) { -#if defined(YYMALLOC) - return Qfalse; -#else - return Qtrue; + int cnt; + + if (!local->used) return; + cnt = local->used->pos; + if (cnt != local->vars->pos) { + rb_parser_fatal(p, "local->used->pos != local->vars->pos"); + } +#ifndef RIPPER + ID *v = local->vars->tbl; + ID *u = local->used->tbl; + for (int i = 0; i < cnt; ++i) { + if (!v[i] || (u[i] & LVAR_USED)) continue; + if (is_private_local_id(v[i])) continue; + rb_warn1L((int)u[i], "assigned but unused variable - %"PRIsWARN, rb_id2str(v[i])); + } #endif } -void -rb_gc_mark_parser() +static void +local_push(struct parser_params *p, int toplevel_scope) { -#if defined YYMALLOC - rb_gc_mark((VALUE)parser_heap); -#elif defined yystacksize - if (yyvsp) rb_gc_mark_locations((VALUE *)yyvs, (VALUE *)yyvsp); -#endif + struct local_vars *local; + int inherits_dvars = toplevel_scope && compile_for_eval; + int warn_unused_vars = RTEST(ruby_verbose); - if (!ruby_in_compile) return; + local = ALLOC(struct local_vars); + local->prev = p->lvtbl; + local->args = vtable_alloc(0); + local->vars = vtable_alloc(inherits_dvars ? DVARS_INHERIT : DVARS_TOPSCOPE); +#ifndef RIPPER + if (toplevel_scope && compile_for_eval) warn_unused_vars = 0; + if (toplevel_scope && e_option_supplied(p)) warn_unused_vars = 0; + local->numparam.outer = 0; + local->numparam.inner = 0; + local->numparam.current = 0; +#endif + local->used = warn_unused_vars ? vtable_alloc(0) : 0; + +# if WARN_PAST_SCOPE + local->past = 0; +# endif + CMDARG_PUSH(0); + COND_PUSH(0); + p->lvtbl = local; +} - rb_gc_mark_maybe((VALUE)yylval.node); - rb_gc_mark(ruby_debug_lines); - rb_gc_mark(lex_lastline); - rb_gc_mark(lex_input); - rb_gc_mark((VALUE)lex_strterm); +static void +local_pop(struct parser_params *p) +{ + struct local_vars *local = p->lvtbl->prev; + if (p->lvtbl->used) { + warn_unused_var(p, p->lvtbl); + vtable_free(p->lvtbl->used); + } +# if WARN_PAST_SCOPE + while (p->lvtbl->past) { + struct vtable *past = p->lvtbl->past; + p->lvtbl->past = past->prev; + vtable_free(past); + } +# endif + vtable_free(p->lvtbl->args); + vtable_free(p->lvtbl->vars); + CMDARG_POP(); + COND_POP(); + ruby_sized_xfree(p->lvtbl, sizeof(*p->lvtbl)); + p->lvtbl = local; } -void -rb_parser_append_print() +#ifndef RIPPER +static rb_ast_id_table_t * +local_tbl(struct parser_params *p) { - ruby_eval_tree = - block_append(ruby_eval_tree, - NEW_FCALL(rb_intern("print"), - NEW_ARRAY(NEW_GVAR(rb_intern("$_"))))); + int cnt_args = vtable_size(p->lvtbl->args); + int cnt_vars = vtable_size(p->lvtbl->vars); + int cnt = cnt_args + cnt_vars; + int i, j; + rb_ast_id_table_t *tbl; + + if (cnt <= 0) return 0; + tbl = rb_ast_new_local_table(p->ast, cnt); + MEMCPY(tbl->ids, p->lvtbl->args->tbl, ID, cnt_args); + /* remove IDs duplicated to warn shadowing */ + for (i = 0, j = cnt_args; i < cnt_vars; ++i) { + ID id = p->lvtbl->vars->tbl[i]; + if (!vtable_included(p->lvtbl->args, id)) { + tbl->ids[j++] = id; + } + } + if (j < cnt) { + tbl = rb_ast_resize_latest_local_table(p->ast, j); + } + + return tbl; } -void -rb_parser_while_loop(chop, split) - int chop, split; -{ - if (split) { - ruby_eval_tree = - block_append(NEW_GASGN(rb_intern("$F"), - NEW_CALL(NEW_GVAR(rb_intern("$_")), - rb_intern("split"), 0)), - ruby_eval_tree); - } - if (chop) { - ruby_eval_tree = - block_append(NEW_CALL(NEW_GVAR(rb_intern("$_")), - rb_intern("chop!"), 0), ruby_eval_tree); - } - ruby_eval_tree = NEW_OPT_N(ruby_eval_tree); -} - -static struct { - ID token; - char *name; -} op_tbl[] = { - {tDOT2, ".."}, - {tDOT3, "..."}, - {'+', "+"}, - {'-', "-"}, - {'+', "+(binary)"}, - {'-', "-(binary)"}, - {'*', "*"}, - {'/', "/"}, - {'%', "%"}, - {tPOW, "**"}, - {tUPLUS, "+@"}, - {tUMINUS, "-@"}, - {tUPLUS, "+(unary)"}, - {tUMINUS, "-(unary)"}, - {'|', "|"}, - {'^', "^"}, - {'&', "&"}, - {tCMP, "<=>"}, - {'>', ">"}, - {tGEQ, ">="}, - {'<', "<"}, - {tLEQ, "<="}, - {tEQ, "=="}, - {tEQQ, "==="}, - {tNEQ, "!="}, - {tMATCH, "=~"}, - {tNMATCH, "!~"}, - {'!', "!"}, - {'~', "~"}, - {'!', "!(unary)"}, - {'~', "~(unary)"}, - {'!', "!@"}, - {'~', "~@"}, - {tAREF, "[]"}, - {tASET, "[]="}, - {tLSHFT, "<<"}, - {tRSHFT, ">>"}, - {tCOLON2, "::"}, - {'`', "`"}, - {0, 0} -}; +static NODE* +node_newnode_with_locals(struct parser_params *p, enum node_type type, VALUE a1, VALUE a2, const rb_code_location_t *loc) +{ + rb_ast_id_table_t *a0; + NODE *n; -static st_table *sym_tbl; -static st_table *sym_rev_tbl; + a0 = local_tbl(p); + n = NEW_NODE(type, a0, a1, a2, loc); + return n; +} -void -Init_sym() +#endif + +static void +numparam_name(struct parser_params *p, ID id) { - sym_tbl = st_init_strtable_with_size(200); - sym_rev_tbl = st_init_numtable_with_size(200); + if (!NUMPARAM_ID_P(id)) return; + compile_error(p, "_%d is reserved for numbered parameter", + NUMPARAM_ID_TO_IDX(id)); } -static ID last_id = tLAST_TOKEN; +static void +arg_var(struct parser_params *p, ID id) +{ + numparam_name(p, id); + vtable_add(p->lvtbl->args, id); +} -static ID -internal_id() +static void +local_var(struct parser_params *p, ID id) { - return ID_INTERNAL | (++last_id << ID_SCOPE_SHIFT); + numparam_name(p, id); + vtable_add(p->lvtbl->vars, id); + if (p->lvtbl->used) { + vtable_add(p->lvtbl->used, (ID)p->ruby_sourceline); + } } static int -is_special_global_name(m) - const char *m; -{ - switch (*m) { - case '~': case '*': case '$': case '?': case '!': case '@': - case '/': case '\\': case ';': case ',': case '.': case '=': - case ':': case '<': case '>': case '\"': - case '&': case '`': case '\'': case '+': - case '0': - ++m; - break; - case '-': - ++m; - if (is_identchar(*m)) m += mbclen(*m); - break; - default: - if (!ISDIGIT(*m)) return 0; - do ++m; while (ISDIGIT(*m)); +local_id_ref(struct parser_params *p, ID id, ID **vidrefp) +{ + struct vtable *vars, *args, *used; + + vars = p->lvtbl->vars; + args = p->lvtbl->args; + used = p->lvtbl->used; + + while (vars && !DVARS_TERMINAL_P(vars->prev)) { + vars = vars->prev; + args = args->prev; + if (used) used = used->prev; + } + + if (vars && vars->prev == DVARS_INHERIT) { + return rb_local_defined(id, p->parent_iseq); + } + else if (vtable_included(args, id)) { + return 1; + } + else { + int i = vtable_included(vars, id); + if (i && used && vidrefp) *vidrefp = &used->tbl[i-1]; + return i != 0; } - return !*m; } -int -rb_symname_p(name) - const char *name; +static int +local_id(struct parser_params *p, ID id) { - const char *m = name; - int localid = Qfalse; - - if (!m) return Qfalse; - switch (*m) { - case '\0': - return Qfalse; + return local_id_ref(p, id, NULL); +} - case '$': - if (is_special_global_name(++m)) return Qtrue; - goto id; +static int +check_forwarding_args(struct parser_params *p) +{ + if (local_id(p, idFWD_REST) && +#if idFWD_KWREST + local_id(p, idFWD_KWREST) && +#endif + local_id(p, idFWD_BLOCK)) return TRUE; + compile_error(p, "unexpected ..."); + return FALSE; +} - case '@': - if (*++m == '@') ++m; - goto id; +static void +add_forwarding_args(struct parser_params *p) +{ + arg_var(p, idFWD_REST); +#if idFWD_KWREST + arg_var(p, idFWD_KWREST); +#endif + arg_var(p, idFWD_BLOCK); +} - case '<': - switch (*++m) { - case '<': ++m; break; - case '=': if (*++m == '>') ++m; break; - default: break; - } - break; +#ifndef RIPPER +static NODE * +new_args_forward_call(struct parser_params *p, NODE *leading, const YYLTYPE *loc, const YYLTYPE *argsloc) +{ + NODE *splat = NEW_SPLAT(NEW_LVAR(idFWD_REST, loc), loc); +#if idFWD_KWREST + NODE *kwrest = list_append(p, NEW_LIST(0, loc), NEW_LVAR(idFWD_KWREST, loc)); +#endif + NODE *block = NEW_BLOCK_PASS(NEW_LVAR(idFWD_BLOCK, loc), loc); + NODE *args = leading ? rest_arg_append(p, leading, splat, argsloc) : splat; +#if idFWD_KWREST + args = arg_append(p, splat, new_hash(p, kwrest, loc), loc); +#endif + return arg_blk_pass(args, block); +} +#endif - case '>': - switch (*++m) { - case '>': case '=': ++m; break; - } - break; +static NODE * +numparam_push(struct parser_params *p) +{ +#ifndef RIPPER + struct local_vars *local = p->lvtbl; + NODE *inner = local->numparam.inner; + if (!local->numparam.outer) { + local->numparam.outer = local->numparam.current; + } + local->numparam.inner = 0; + local->numparam.current = 0; + return inner; +#else + return 0; +#endif +} - case '=': - switch (*++m) { - case '~': ++m; break; - case '=': if (*++m == '=') ++m; break; - default: return Qfalse; - } - break; +static void +numparam_pop(struct parser_params *p, NODE *prev_inner) +{ +#ifndef RIPPER + struct local_vars *local = p->lvtbl; + if (prev_inner) { + /* prefer first one */ + local->numparam.inner = prev_inner; + } + else if (local->numparam.current) { + /* current and inner are exclusive */ + local->numparam.inner = local->numparam.current; + } + if (p->max_numparam > NO_PARAM) { + /* current and outer are exclusive */ + local->numparam.current = local->numparam.outer; + local->numparam.outer = 0; + } + else { + /* no numbered parameter */ + local->numparam.current = 0; + } +#endif +} - case '*': - if (*++m == '*') ++m; - break; +static const struct vtable * +dyna_push(struct parser_params *p) +{ + p->lvtbl->args = vtable_alloc(p->lvtbl->args); + p->lvtbl->vars = vtable_alloc(p->lvtbl->vars); + if (p->lvtbl->used) { + p->lvtbl->used = vtable_alloc(p->lvtbl->used); + } + return p->lvtbl->args; +} - case '+': case '-': - if (*++m == '@') ++m; - break; +static void +dyna_pop_vtable(struct parser_params *p, struct vtable **vtblp) +{ + struct vtable *tmp = *vtblp; + *vtblp = tmp->prev; +# if WARN_PAST_SCOPE + if (p->past_scope_enabled) { + tmp->prev = p->lvtbl->past; + p->lvtbl->past = tmp; + return; + } +# endif + vtable_free(tmp); +} - case '|': case '^': case '&': case '/': case '%': case '~': case '`': - ++m; - break; +static void +dyna_pop_1(struct parser_params *p) +{ + struct vtable *tmp; - case '[': - if (*++m != ']') return Qfalse; - if (*++m == '=') ++m; - break; + if ((tmp = p->lvtbl->used) != 0) { + warn_unused_var(p, p->lvtbl); + p->lvtbl->used = p->lvtbl->used->prev; + vtable_free(tmp); + } + dyna_pop_vtable(p, &p->lvtbl->args); + dyna_pop_vtable(p, &p->lvtbl->vars); +} - default: - localid = !ISUPPER(*m); - id: - if (*m != '_' && !ISALPHA(*m) && !ismbchar(*m)) return Qfalse; - while (is_identchar(*m)) m += mbclen(*m); - if (localid) { - switch (*m) { - case '!': case '?': case '=': ++m; - } +static void +dyna_pop(struct parser_params *p, const struct vtable *lvargs) +{ + while (p->lvtbl->args != lvargs) { + dyna_pop_1(p); + if (!p->lvtbl->args) { + struct local_vars *local = p->lvtbl->prev; + ruby_sized_xfree(p->lvtbl, sizeof(*p->lvtbl)); + p->lvtbl = local; } - break; } - return *m ? Qfalse : Qtrue; + dyna_pop_1(p); } -int -rb_sym_interned_p(str) - VALUE str; +static int +dyna_in_block(struct parser_params *p) { - ID id; - - if (st_lookup(sym_tbl, (st_data_t)RSTRING(str)->ptr, (st_data_t *)&id)) - return Qtrue; - return Qfalse; + return !DVARS_TERMINAL_P(p->lvtbl->vars) && p->lvtbl->vars->prev != DVARS_TOPSCOPE; } -ID -rb_intern(name) - const char *name; +static int +dvar_defined_ref(struct parser_params *p, ID id, ID **vidrefp) { - const char *m = name; - ID id; - int last; + struct vtable *vars, *args, *used; + int i; - if (st_lookup(sym_tbl, (st_data_t)name, (st_data_t *)&id)) - return id; + args = p->lvtbl->args; + vars = p->lvtbl->vars; + used = p->lvtbl->used; - last = strlen(name)-1; - id = 0; - switch (*name) { - case '$': - id |= ID_GLOBAL; - if (is_special_global_name(++m)) goto new_id; - break; - case '@': - if (name[1] == '@') { - m++; - id |= ID_CLASS; - } - else { - id |= ID_INSTANCE; + while (!DVARS_TERMINAL_P(vars)) { + if (vtable_included(args, id)) { + return 1; } - m++; - break; - default: - if (name[0] != '_' && ISASCII(name[0]) && !ISALNUM(name[0])) { - /* operators */ - int i; - - for (i=0; op_tbl[i].token; i++) { - if (*op_tbl[i].name == *name && - strcmp(op_tbl[i].name, name) == 0) { - id = op_tbl[i].token; - goto id_regist; - } - } + if ((i = vtable_included(vars, id)) != 0) { + if (used && vidrefp) *vidrefp = &used->tbl[i-1]; + return 1; } + args = args->prev; + vars = vars->prev; + if (!vidrefp) used = 0; + if (used) used = used->prev; + } + + if (vars == DVARS_INHERIT && !NUMPARAM_ID_P(id)) { + return rb_dvar_defined(id, p->parent_iseq); + } - if (name[last] == '=') { - /* attribute assignment */ - char *buf = ALLOCA_N(char,last+1); + return 0; +} - strncpy(buf, name, last); - buf[last] = '\0'; - id = rb_intern(buf); - if (id > tLAST_TOKEN && !is_attrset_id(id)) { - id = rb_id_attrset(id); - goto id_regist; - } - id = ID_ATTRSET; - } - else if (ISUPPER(name[0])) { - id = ID_CONST; +static int +dvar_defined(struct parser_params *p, ID id) +{ + return dvar_defined_ref(p, id, NULL); +} + +static int +dvar_curr(struct parser_params *p, ID id) +{ + return (vtable_included(p->lvtbl->args, id) || + vtable_included(p->lvtbl->vars, id)); +} + +static void +reg_fragment_enc_error(struct parser_params* p, VALUE str, int c) +{ + compile_error(p, + "regexp encoding option '%c' differs from source encoding '%s'", + c, rb_enc_name(rb_enc_get(str))); +} + +#ifndef RIPPER +int +rb_reg_fragment_setenc(struct parser_params* p, VALUE str, int options) +{ + int c = RE_OPTION_ENCODING_IDX(options); + + if (c) { + int opt, idx; + rb_char_to_option_kcode(c, &opt, &idx); + if (idx != ENCODING_GET(str) && + rb_enc_str_coderange(str) != ENC_CODERANGE_7BIT) { + goto error; + } + ENCODING_SET(str, idx); + } + else if (RE_OPTION_ENCODING_NONE(options)) { + if (!ENCODING_IS_ASCII8BIT(str) && + rb_enc_str_coderange(str) != ENC_CODERANGE_7BIT) { + c = 'n'; + goto error; } + rb_enc_associate(str, rb_ascii8bit_encoding()); + } + else if (p->enc == rb_usascii_encoding()) { + if (rb_enc_str_coderange(str) != ENC_CODERANGE_7BIT) { + /* raise in re.c */ + rb_enc_associate(str, rb_usascii_encoding()); + } else { - id = ID_LOCAL; + rb_enc_associate(str, rb_ascii8bit_encoding()); } - break; } - if (!ISDIGIT(*m)) { - while (m <= name + last && is_identchar(*m)) { - m += mbclen(*m); - } + return 0; + + error: + return c; +} + +static void +reg_fragment_setenc(struct parser_params* p, VALUE str, int options) +{ + int c = rb_reg_fragment_setenc(p, str, options); + if (c) reg_fragment_enc_error(p, str, c); +} + +static int +reg_fragment_check(struct parser_params* p, VALUE str, int options) +{ + VALUE err; + reg_fragment_setenc(p, str, options); + err = rb_reg_check_preprocess(str); + if (err != Qnil) { + err = rb_obj_as_string(err); + compile_error(p, "%"PRIsVALUE, err); + return 0; } - if (*m) id = ID_JUNK; - new_id: - id |= ++last_id << ID_SCOPE_SHIFT; - id_regist: - name = strdup(name); - st_add_direct(sym_tbl, (st_data_t)name, id); - st_add_direct(sym_rev_tbl, id, (st_data_t)name); - return id; + return 1; } -char * -rb_id2name(id) - ID id; +typedef struct { + struct parser_params* parser; + rb_encoding *enc; + NODE *succ_block; + const YYLTYPE *loc; +} reg_named_capture_assign_t; + +static int +reg_named_capture_assign_iter(const OnigUChar *name, const OnigUChar *name_end, + int back_num, int *back_refs, OnigRegex regex, void *arg0) +{ + reg_named_capture_assign_t *arg = (reg_named_capture_assign_t*)arg0; + struct parser_params* p = arg->parser; + rb_encoding *enc = arg->enc; + long len = name_end - name; + const char *s = (const char *)name; + ID var; + NODE *node, *succ; + + if (!len) return ST_CONTINUE; + if (rb_enc_symname_type(s, len, enc, (1U<<ID_LOCAL)) != ID_LOCAL) + return ST_CONTINUE; + + var = intern_cstr(s, len, enc); + if (len < MAX_WORD_LENGTH && rb_reserved_word(s, (int)len)) { + if (!lvar_defined(p, var)) return ST_CONTINUE; + } + node = node_assign(p, assignable(p, var, 0, arg->loc), NEW_LIT(ID2SYM(var), arg->loc), NO_LEX_CTXT, arg->loc); + succ = arg->succ_block; + if (!succ) succ = NEW_BEGIN(0, arg->loc); + succ = block_append(p, succ, node); + arg->succ_block = succ; + return ST_CONTINUE; +} + +static NODE * +reg_named_capture_assign(struct parser_params* p, VALUE regexp, const YYLTYPE *loc) { - char *name; + reg_named_capture_assign_t arg; - if (id < tLAST_TOKEN) { - int i; + arg.parser = p; + arg.enc = rb_enc_get(regexp); + arg.succ_block = 0; + arg.loc = loc; + onig_foreach_name(RREGEXP_PTR(regexp), reg_named_capture_assign_iter, &arg); - for (i=0; op_tbl[i].token; i++) { - if (op_tbl[i].token == id) - return op_tbl[i].name; - } - } + if (!arg.succ_block) return 0; + return arg.succ_block->nd_next; +} - if (st_lookup(sym_rev_tbl, id, (st_data_t *)&name)) - return name; +static VALUE +parser_reg_compile(struct parser_params* p, VALUE str, int options) +{ + reg_fragment_setenc(p, str, options); + return rb_parser_reg_compile(p, str, options); +} + +VALUE +rb_parser_reg_compile(struct parser_params* p, VALUE str, int options) +{ + return rb_reg_compile(str, options & RE_OPTION_MASK, p->ruby_sourcefile, p->ruby_sourceline); +} + +static VALUE +reg_compile(struct parser_params* p, VALUE str, int options) +{ + VALUE re; + VALUE err; + + err = rb_errinfo(); + re = parser_reg_compile(p, str, options); + if (NIL_P(re)) { + VALUE m = rb_attr_get(rb_errinfo(), idMesg); + rb_set_errinfo(err); + compile_error(p, "%"PRIsVALUE, m); + return Qnil; + } + return re; +} +#else +static VALUE +parser_reg_compile(struct parser_params* p, VALUE str, int options, VALUE *errmsg) +{ + VALUE err = rb_errinfo(); + VALUE re; + str = ripper_is_node_yylval(str) ? RNODE(str)->nd_cval : str; + int c = rb_reg_fragment_setenc(p, str, options); + if (c) reg_fragment_enc_error(p, str, c); + re = rb_parser_reg_compile(p, str, options); + if (NIL_P(re)) { + *errmsg = rb_attr_get(rb_errinfo(), idMesg); + rb_set_errinfo(err); + } + return re; +} +#endif - if (is_attrset_id(id)) { - ID id2 = (id & ~ID_SCOPE_MASK) | ID_LOCAL; +#ifndef RIPPER +void +rb_parser_set_options(VALUE vparser, int print, int loop, int chomp, int split) +{ + struct parser_params *p; + TypedData_Get_Struct(vparser, struct parser_params, &parser_data_type, p); + p->do_print = print; + p->do_loop = loop; + p->do_chomp = chomp; + p->do_split = split; +} + +static NODE * +parser_append_options(struct parser_params *p, NODE *node) +{ + static const YYLTYPE default_location = {{1, 0}, {1, 0}}; + const YYLTYPE *const LOC = &default_location; - again: - name = rb_id2name(id2); - if (name) { - char *buf = ALLOCA_N(char, strlen(name)+2); + if (p->do_print) { + NODE *print = NEW_FCALL(rb_intern("print"), + NEW_LIST(NEW_GVAR(idLASTLINE, LOC), LOC), + LOC); + node = block_append(p, node, print); + } - strcpy(buf, name); - strcat(buf, "="); - rb_intern(buf); - return rb_id2name(id); + if (p->do_loop) { + if (p->do_split) { + ID ifs = rb_intern("$;"); + ID fields = rb_intern("$F"); + NODE *args = NEW_LIST(NEW_GVAR(ifs, LOC), LOC); + NODE *split = NEW_GASGN(fields, + NEW_CALL(NEW_GVAR(idLASTLINE, LOC), + rb_intern("split"), args, LOC), + LOC); + node = block_append(p, split, node); } - if (is_local_id(id2)) { - id2 = (id & ~ID_SCOPE_MASK) | ID_CONST; - goto again; + if (p->do_chomp) { + NODE *chomp = NEW_CALL(NEW_GVAR(idLASTLINE, LOC), + rb_intern("chomp!"), 0, LOC); + node = block_append(p, chomp, node); } + + node = NEW_WHILE(NEW_VCALL(idGets, LOC), node, 1, LOC); } - return 0; + + return node; } -static int -symbols_i(key, value, ary) - char *key; - ID value; - VALUE ary; +void +rb_init_parse(void) { - rb_ary_push(ary, ID2SYM(value)); - return ST_CONTINUE; + /* just to suppress unused-function warnings */ + (void)nodetype; + (void)nodeline; } -/* - * call-seq: - * Symbol.all_symbols => array - * - * Returns an array of all the symbols currently in Ruby's symbol - * table. - * - * Symbol.all_symbols.size #=> 903 - * Symbol.all_symbols[1,20] #=> [:floor, :ARGV, :Binding, :symlink, - * :chown, :EOFError, :$;, :String, - * :LOCK_SH, :"setuid?", :$<, - * :default_proc, :compact, :extend, - * :Tms, :getwd, :$=, :ThreadGroup, - * :wait2, :$>] - */ +static ID +internal_id(struct parser_params *p) +{ + return rb_make_temporary_id(vtable_size(p->lvtbl->args) + vtable_size(p->lvtbl->vars)); +} +#endif /* !RIPPER */ -VALUE -rb_sym_all_symbols() +static void +parser_initialize(struct parser_params *p) { - VALUE ary = rb_ary_new2(sym_tbl->num_entries); + /* note: we rely on TypedData_Make_Struct to set most fields to 0 */ + p->command_start = TRUE; + p->ruby_sourcefile_string = Qnil; + p->lex.lpar_beg = -1; /* make lambda_beginning_p() == FALSE at first */ + p->node_id = 0; +#ifdef RIPPER + p->delayed.token = Qnil; + p->result = Qnil; + p->parsing_thread = Qnil; +#else + p->error_buffer = Qfalse; +#endif + p->debug_buffer = Qnil; + p->debug_output = rb_ractor_stdout(); + p->enc = rb_utf8_encoding(); +} - st_foreach(sym_tbl, symbols_i, ary); - return ary; +#ifdef RIPPER +#define parser_mark ripper_parser_mark +#define parser_free ripper_parser_free +#endif + +static void +parser_mark(void *ptr) +{ + struct parser_params *p = (struct parser_params*)ptr; + + rb_gc_mark(p->lex.input); + rb_gc_mark(p->lex.prevline); + rb_gc_mark(p->lex.lastline); + rb_gc_mark(p->lex.nextline); + rb_gc_mark(p->ruby_sourcefile_string); + rb_gc_mark((VALUE)p->lex.strterm); + rb_gc_mark((VALUE)p->ast); + rb_gc_mark(p->case_labels); +#ifndef RIPPER + rb_gc_mark(p->debug_lines); + rb_gc_mark(p->compile_option); + rb_gc_mark(p->error_buffer); +#else + rb_gc_mark(p->delayed.token); + rb_gc_mark(p->value); + rb_gc_mark(p->result); + rb_gc_mark(p->parsing_thread); +#endif + rb_gc_mark(p->debug_buffer); + rb_gc_mark(p->debug_output); +#ifdef YYMALLOC + rb_gc_mark((VALUE)p->heap); +#endif } -int -rb_is_const_id(id) - ID id; +static void +parser_free(void *ptr) { - if (is_const_id(id)) return Qtrue; - return Qfalse; + struct parser_params *p = (struct parser_params*)ptr; + struct local_vars *local, *prev; + + if (p->tokenbuf) { + ruby_sized_xfree(p->tokenbuf, p->toksiz); + } + for (local = p->lvtbl; local; local = prev) { + if (local->vars) xfree(local->vars); + prev = local->prev; + xfree(local); + } + { + token_info *ptinfo; + while ((ptinfo = p->token_info) != 0) { + p->token_info = ptinfo->next; + xfree(ptinfo); + } + } + xfree(ptr); } -int -rb_is_class_id(id) - ID id; +static size_t +parser_memsize(const void *ptr) +{ + struct parser_params *p = (struct parser_params*)ptr; + struct local_vars *local; + size_t size = sizeof(*p); + + size += p->toksiz; + for (local = p->lvtbl; local; local = local->prev) { + size += sizeof(*local); + if (local->vars) size += local->vars->capa * sizeof(ID); + } + return size; +} + +static const rb_data_type_t parser_data_type = { +#ifndef RIPPER + "parser", +#else + "ripper", +#endif + { + parser_mark, + parser_free, + parser_memsize, + }, + 0, 0, RUBY_TYPED_FREE_IMMEDIATELY +}; + +#ifndef RIPPER +#undef rb_reserved_word + +const struct kwtable * +rb_reserved_word(const char *str, unsigned int len) { - if (is_class_id(id)) return Qtrue; - return Qfalse; + return reserved_word(str, len); } -int -rb_is_instance_id(id) - ID id; +VALUE +rb_parser_new(void) { - if (is_instance_id(id)) return Qtrue; - return Qfalse; + struct parser_params *p; + VALUE parser = TypedData_Make_Struct(0, struct parser_params, + &parser_data_type, p); + parser_initialize(p); + return parser; } -int -rb_is_local_id(id) - ID id; +VALUE +rb_parser_set_context(VALUE vparser, const struct rb_iseq_struct *base, int main) { - if (is_local_id(id)) return Qtrue; - return Qfalse; + struct parser_params *p; + + TypedData_Get_Struct(vparser, struct parser_params, &parser_data_type, p); + p->error_buffer = main ? Qfalse : Qnil; + p->parent_iseq = base; + return vparser; } -int -rb_is_junk_id(id) - ID id; +void +rb_parser_keep_script_lines(VALUE vparser) { - if (is_junk_id(id)) return Qtrue; - return Qfalse; + struct parser_params *p; + + TypedData_Get_Struct(vparser, struct parser_params, &parser_data_type, p); + p->keep_script_lines = 1; } +#endif -static void -special_local_set(c, val) - char c; - VALUE val; +#ifdef RIPPER +#define rb_parser_end_seen_p ripper_parser_end_seen_p +#define rb_parser_encoding ripper_parser_encoding +#define rb_parser_get_yydebug ripper_parser_get_yydebug +#define rb_parser_set_yydebug ripper_parser_set_yydebug +#define rb_parser_get_debug_output ripper_parser_get_debug_output +#define rb_parser_set_debug_output ripper_parser_set_debug_output +static VALUE ripper_parser_end_seen_p(VALUE vparser); +static VALUE ripper_parser_encoding(VALUE vparser); +static VALUE ripper_parser_get_yydebug(VALUE self); +static VALUE ripper_parser_set_yydebug(VALUE self, VALUE flag); +static VALUE ripper_parser_get_debug_output(VALUE self); +static VALUE ripper_parser_set_debug_output(VALUE self, VALUE output); + +/* + * call-seq: + * ripper.error? -> Boolean + * + * Return true if parsed source has errors. + */ +static VALUE +ripper_error_p(VALUE vparser) { - int cnt; + struct parser_params *p; - top_local_init(); - cnt = local_cnt(c); - top_local_setup(); - ruby_scope->local_vars[cnt] = val; + TypedData_Get_Struct(vparser, struct parser_params, &parser_data_type, p); + return RBOOL(p->error_p); } +#endif +/* + * call-seq: + * ripper.end_seen? -> Boolean + * + * Return true if parsed source ended by +\_\_END\_\_+. + */ VALUE -rb_backref_get() +rb_parser_end_seen_p(VALUE vparser) { - VALUE *var = rb_svar(1); - if (var) { - return *var; - } - return Qnil; + struct parser_params *p; + + TypedData_Get_Struct(vparser, struct parser_params, &parser_data_type, p); + return RBOOL(p->ruby__end__seen); } -void -rb_backref_set(val) - VALUE val; +/* + * call-seq: + * ripper.encoding -> encoding + * + * Return encoding of the source. + */ +VALUE +rb_parser_encoding(VALUE vparser) { - VALUE *var = rb_svar(1); - if (var) { - *var = val; - } - else { - special_local_set('~', val); - } + struct parser_params *p; + + TypedData_Get_Struct(vparser, struct parser_params, &parser_data_type, p); + return rb_enc_from_encoding(p->enc); } +#ifdef RIPPER +/* + * call-seq: + * ripper.yydebug -> true or false + * + * Get yydebug. + */ VALUE -rb_lastline_get() +rb_parser_get_yydebug(VALUE self) { - VALUE *var = rb_svar(0); - if (var) { - return *var; - } - return Qnil; + struct parser_params *p; + + TypedData_Get_Struct(self, struct parser_params, &parser_data_type, p); + return RBOOL(p->debug); } +#endif -void -rb_lastline_set(val) - VALUE val; +/* + * call-seq: + * ripper.yydebug = flag + * + * Set yydebug. + */ +VALUE +rb_parser_set_yydebug(VALUE self, VALUE flag) { - VALUE *var = rb_svar(0); - if (var) { - *var = val; - } - else { - special_local_set('_', val); - } + struct parser_params *p; + + TypedData_Get_Struct(self, struct parser_params, &parser_data_type, p); + p->debug = RTEST(flag); + return flag; } +/* + * call-seq: + * ripper.debug_output -> obj + * + * Get debug output. + */ +VALUE +rb_parser_get_debug_output(VALUE self) +{ + struct parser_params *p; + + TypedData_Get_Struct(self, struct parser_params, &parser_data_type, p); + return p->debug_output; +} + +/* + * call-seq: + * ripper.debug_output = obj + * + * Set debug output. + */ +VALUE +rb_parser_set_debug_output(VALUE self, VALUE output) +{ + struct parser_params *p; + + TypedData_Get_Struct(self, struct parser_params, &parser_data_type, p); + return p->debug_output = output; +} + +#ifndef RIPPER #ifdef YYMALLOC #define HEAPCNT(n, size) ((n) * (size) / sizeof(YYSTYPE)) -#define NEWHEAP() rb_node_newnode(NODE_ALLOCA, 0, (VALUE)parser_heap, 0) -#define ADD2HEAP(n, c, p) ((parser_heap = (n))->u1.node = (p), \ - (n)->u3.cnt = (c), (p)) - -static void * -rb_parser_malloc(size) - size_t size; +/* Keep the order; NEWHEAP then xmalloc and ADD2HEAP to get rid of + * potential memory leak */ +#define NEWHEAP() rb_imemo_tmpbuf_parser_heap(0, p->heap, 0) +#define ADD2HEAP(new, cnt, ptr) ((p->heap = (new))->ptr = (ptr), \ + (new)->cnt = (cnt), (ptr)) + +void * +rb_parser_malloc(struct parser_params *p, size_t size) { size_t cnt = HEAPCNT(1, size); - NODE *n = NEWHEAP(); + rb_imemo_tmpbuf_t *n = NEWHEAP(); void *ptr = xmalloc(size); return ADD2HEAP(n, cnt, ptr); } -static void * -rb_parser_calloc(nelem, size) - size_t nelem, size; +void * +rb_parser_calloc(struct parser_params *p, size_t nelem, size_t size) { size_t cnt = HEAPCNT(nelem, size); - NODE *n = NEWHEAP(); + rb_imemo_tmpbuf_t *n = NEWHEAP(); void *ptr = xcalloc(nelem, size); return ADD2HEAP(n, cnt, ptr); } -static void * -rb_parser_realloc(ptr, size) - void *ptr; - size_t size; +void * +rb_parser_realloc(struct parser_params *p, void *ptr, size_t size) { - NODE *n; + rb_imemo_tmpbuf_t *n; size_t cnt = HEAPCNT(1, size); - if (ptr && (n = parser_heap) != NULL) { + if (ptr && (n = p->heap) != NULL) { do { - if (n->u1.node == ptr) { - n->u1.node = ptr = xrealloc(ptr, size); - if (n->u3.cnt) n->u3.cnt = cnt; + if (n->ptr == ptr) { + n->ptr = ptr = xrealloc(ptr, size); + if (n->cnt) n->cnt = cnt; return ptr; } - } while ((n = n->u2.node) != NULL); + } while ((n = n->next) != NULL); } n = NEWHEAP(); ptr = xrealloc(ptr, size); return ADD2HEAP(n, cnt, ptr); } -static void -rb_parser_free(ptr) - void *ptr; +void +rb_parser_free(struct parser_params *p, void *ptr) { - NODE **prev = &parser_heap, *n; + rb_imemo_tmpbuf_t **prev = &p->heap, *n; - while (n = *prev) { - if (n->u1.node == ptr) { - *prev = n->u2.node; - rb_gc_force_recycle((VALUE)n); + while ((n = *prev) != NULL) { + if (n->ptr == ptr) { + *prev = n->next; break; } - prev = &n->u2.node; + prev = &n->next; } - xfree(ptr); } #endif + +void +rb_parser_printf(struct parser_params *p, const char *fmt, ...) +{ + va_list ap; + VALUE mesg = p->debug_buffer; + + if (NIL_P(mesg)) p->debug_buffer = mesg = rb_str_new(0, 0); + va_start(ap, fmt); + rb_str_vcatf(mesg, fmt, ap); + va_end(ap); + if (RSTRING_END(mesg)[-1] == '\n') { + rb_io_write(p->debug_output, mesg); + p->debug_buffer = Qnil; + } +} + +static void +parser_compile_error(struct parser_params *p, const char *fmt, ...) +{ + va_list ap; + + rb_io_flush(p->debug_output); + p->error_p = 1; + va_start(ap, fmt); + p->error_buffer = + rb_syntax_error_append(p->error_buffer, + p->ruby_sourcefile_string, + p->ruby_sourceline, + rb_long2int(p->lex.pcur - p->lex.pbeg), + p->enc, fmt, ap); + va_end(ap); +} + +static size_t +count_char(const char *str, int c) +{ + int n = 0; + while (str[n] == c) ++n; + return n; +} + +/* + * strip enclosing double-quotes, same as the default yytnamerr except + * for that single-quotes matching back-quotes do not stop stripping. + * + * "\"`class' keyword\"" => "`class' keyword" + */ +RUBY_FUNC_EXPORTED size_t +rb_yytnamerr(struct parser_params *p, char *yyres, const char *yystr) +{ + if (*yystr == '"') { + size_t yyn = 0, bquote = 0; + const char *yyp = yystr; + + while (*++yyp) { + switch (*yyp) { + case '`': + if (!bquote) { + bquote = count_char(yyp+1, '`') + 1; + if (yyres) memcpy(&yyres[yyn], yyp, bquote); + yyn += bquote; + yyp += bquote - 1; + break; + } + goto default_char; + + case '\'': + if (bquote && count_char(yyp+1, '\'') + 1 == bquote) { + if (yyres) memcpy(yyres + yyn, yyp, bquote); + yyn += bquote; + yyp += bquote - 1; + bquote = 0; + break; + } + if (yyp[1] && yyp[1] != '\'' && yyp[2] == '\'') { + if (yyres) memcpy(yyres + yyn, yyp, 3); + yyn += 3; + yyp += 2; + break; + } + goto do_not_strip_quotes; + + case ',': + goto do_not_strip_quotes; + + case '\\': + if (*++yyp != '\\') + goto do_not_strip_quotes; + /* Fall through. */ + default_char: + default: + if (yyres) + yyres[yyn] = *yyp; + yyn++; + break; + + case '"': + case '\0': + if (yyres) + yyres[yyn] = '\0'; + return yyn; + } + } + do_not_strip_quotes: ; + } + + if (!yyres) return strlen(yystr); + + return (YYSIZE_T)(yystpcpy(yyres, yystr) - yyres); +} +#endif + +#ifdef RIPPER +#ifdef RIPPER_DEBUG +/* :nodoc: */ +static VALUE +ripper_validate_object(VALUE self, VALUE x) +{ + if (x == Qfalse) return x; + if (x == Qtrue) return x; + if (x == Qnil) return x; + if (x == Qundef) + rb_raise(rb_eArgError, "Qundef given"); + if (FIXNUM_P(x)) return x; + if (SYMBOL_P(x)) return x; + switch (BUILTIN_TYPE(x)) { + case T_STRING: + case T_OBJECT: + case T_ARRAY: + case T_BIGNUM: + case T_FLOAT: + case T_COMPLEX: + case T_RATIONAL: + break; + case T_NODE: + if (!nd_type_p((NODE *)x, NODE_RIPPER)) { + rb_raise(rb_eArgError, "NODE given: %p", (void *)x); + } + x = ((NODE *)x)->nd_rval; + break; + default: + rb_raise(rb_eArgError, "wrong type of ruby object: %p (%s)", + (void *)x, rb_obj_classname(x)); + } + if (!RBASIC_CLASS(x)) { + rb_raise(rb_eArgError, "hidden ruby object: %p (%s)", + (void *)x, rb_builtin_type_name(TYPE(x))); + } + return x; +} +#endif + +#define validate(x) ((x) = get_value(x)) + +static VALUE +ripper_dispatch0(struct parser_params *p, ID mid) +{ + return rb_funcall(p->value, mid, 0); +} + +static VALUE +ripper_dispatch1(struct parser_params *p, ID mid, VALUE a) +{ + validate(a); + return rb_funcall(p->value, mid, 1, a); +} + +static VALUE +ripper_dispatch2(struct parser_params *p, ID mid, VALUE a, VALUE b) +{ + validate(a); + validate(b); + return rb_funcall(p->value, mid, 2, a, b); +} + +static VALUE +ripper_dispatch3(struct parser_params *p, ID mid, VALUE a, VALUE b, VALUE c) +{ + validate(a); + validate(b); + validate(c); + return rb_funcall(p->value, mid, 3, a, b, c); +} + +static VALUE +ripper_dispatch4(struct parser_params *p, ID mid, VALUE a, VALUE b, VALUE c, VALUE d) +{ + validate(a); + validate(b); + validate(c); + validate(d); + return rb_funcall(p->value, mid, 4, a, b, c, d); +} + +static VALUE +ripper_dispatch5(struct parser_params *p, ID mid, VALUE a, VALUE b, VALUE c, VALUE d, VALUE e) +{ + validate(a); + validate(b); + validate(c); + validate(d); + validate(e); + return rb_funcall(p->value, mid, 5, a, b, c, d, e); +} + +static VALUE +ripper_dispatch7(struct parser_params *p, ID mid, VALUE a, VALUE b, VALUE c, VALUE d, VALUE e, VALUE f, VALUE g) +{ + validate(a); + validate(b); + validate(c); + validate(d); + validate(e); + validate(f); + validate(g); + return rb_funcall(p->value, mid, 7, a, b, c, d, e, f, g); +} + +static ID +ripper_get_id(VALUE v) +{ + NODE *nd; + if (!RB_TYPE_P(v, T_NODE)) return 0; + nd = (NODE *)v; + if (!nd_type_p(nd, NODE_RIPPER)) return 0; + return nd->nd_vid; +} + +static VALUE +ripper_get_value(VALUE v) +{ + NODE *nd; + if (v == Qundef) return Qnil; + if (!RB_TYPE_P(v, T_NODE)) return v; + nd = (NODE *)v; + if (!nd_type_p(nd, NODE_RIPPER)) return Qnil; + return nd->nd_rval; +} + +static void +ripper_error(struct parser_params *p) +{ + p->error_p = TRUE; +} + +static void +ripper_compile_error(struct parser_params *p, const char *fmt, ...) +{ + VALUE str; + va_list args; + + va_start(args, fmt); + str = rb_vsprintf(fmt, args); + va_end(args); + rb_funcall(p->value, rb_intern("compile_error"), 1, str); + ripper_error(p); +} + +static VALUE +ripper_lex_get_generic(struct parser_params *p, VALUE src) +{ + VALUE line = rb_funcallv_public(src, id_gets, 0, 0); + if (!NIL_P(line) && !RB_TYPE_P(line, T_STRING)) { + rb_raise(rb_eTypeError, + "gets returned %"PRIsVALUE" (expected String or nil)", + rb_obj_class(line)); + } + return line; +} + +static VALUE +ripper_lex_io_get(struct parser_params *p, VALUE src) +{ + return rb_io_gets(src); +} + +static VALUE +ripper_s_allocate(VALUE klass) +{ + struct parser_params *p; + VALUE self = TypedData_Make_Struct(klass, struct parser_params, + &parser_data_type, p); + p->value = self; + return self; +} + +#define ripper_initialized_p(r) ((r)->lex.input != 0) + +/* + * call-seq: + * Ripper.new(src, filename="(ripper)", lineno=1) -> ripper + * + * Create a new Ripper object. + * _src_ must be a String, an IO, or an Object which has #gets method. + * + * This method does not starts parsing. + * See also Ripper#parse and Ripper.parse. + */ +static VALUE +ripper_initialize(int argc, VALUE *argv, VALUE self) +{ + struct parser_params *p; + VALUE src, fname, lineno; + + TypedData_Get_Struct(self, struct parser_params, &parser_data_type, p); + rb_scan_args(argc, argv, "12", &src, &fname, &lineno); + if (RB_TYPE_P(src, T_FILE)) { + p->lex.gets = ripper_lex_io_get; + } + else if (rb_respond_to(src, id_gets)) { + p->lex.gets = ripper_lex_get_generic; + } + else { + StringValue(src); + p->lex.gets = lex_get_str; + } + p->lex.input = src; + p->eofp = 0; + if (NIL_P(fname)) { + fname = STR_NEW2("(ripper)"); + OBJ_FREEZE(fname); + } + else { + StringValueCStr(fname); + fname = rb_str_new_frozen(fname); + } + parser_initialize(p); + + p->ruby_sourcefile_string = fname; + p->ruby_sourcefile = RSTRING_PTR(fname); + p->ruby_sourceline = NIL_P(lineno) ? 0 : NUM2INT(lineno) - 1; + + return Qnil; +} + +static VALUE +ripper_parse0(VALUE parser_v) +{ + struct parser_params *p; + + TypedData_Get_Struct(parser_v, struct parser_params, &parser_data_type, p); + parser_prepare(p); + p->ast = rb_ast_new(); + ripper_yyparse((void*)p); + rb_ast_dispose(p->ast); + p->ast = 0; + return p->result; +} + +static VALUE +ripper_ensure(VALUE parser_v) +{ + struct parser_params *p; + + TypedData_Get_Struct(parser_v, struct parser_params, &parser_data_type, p); + p->parsing_thread = Qnil; + return Qnil; +} + +/* + * call-seq: + * ripper.parse + * + * Start parsing and returns the value of the root action. + */ +static VALUE +ripper_parse(VALUE self) +{ + struct parser_params *p; + + TypedData_Get_Struct(self, struct parser_params, &parser_data_type, p); + if (!ripper_initialized_p(p)) { + rb_raise(rb_eArgError, "method called for uninitialized object"); + } + if (!NIL_P(p->parsing_thread)) { + if (p->parsing_thread == rb_thread_current()) + rb_raise(rb_eArgError, "Ripper#parse is not reentrant"); + else + rb_raise(rb_eArgError, "Ripper#parse is not multithread-safe"); + } + p->parsing_thread = rb_thread_current(); + rb_ensure(ripper_parse0, self, ripper_ensure, self); + + return p->result; +} + +/* + * call-seq: + * ripper.column -> Integer + * + * Return column number of current parsing line. + * This number starts from 0. + */ +static VALUE +ripper_column(VALUE self) +{ + struct parser_params *p; + long col; + + TypedData_Get_Struct(self, struct parser_params, &parser_data_type, p); + if (!ripper_initialized_p(p)) { + rb_raise(rb_eArgError, "method called for uninitialized object"); + } + if (NIL_P(p->parsing_thread)) return Qnil; + col = p->lex.ptok - p->lex.pbeg; + return LONG2NUM(col); +} + +/* + * call-seq: + * ripper.filename -> String + * + * Return current parsing filename. + */ +static VALUE +ripper_filename(VALUE self) +{ + struct parser_params *p; + + TypedData_Get_Struct(self, struct parser_params, &parser_data_type, p); + if (!ripper_initialized_p(p)) { + rb_raise(rb_eArgError, "method called for uninitialized object"); + } + return p->ruby_sourcefile_string; +} + +/* + * call-seq: + * ripper.lineno -> Integer + * + * Return line number of current parsing line. + * This number starts from 1. + */ +static VALUE +ripper_lineno(VALUE self) +{ + struct parser_params *p; + + TypedData_Get_Struct(self, struct parser_params, &parser_data_type, p); + if (!ripper_initialized_p(p)) { + rb_raise(rb_eArgError, "method called for uninitialized object"); + } + if (NIL_P(p->parsing_thread)) return Qnil; + return INT2NUM(p->ruby_sourceline); +} + +/* + * call-seq: + * ripper.state -> Integer + * + * Return scanner state of current token. + */ +static VALUE +ripper_state(VALUE self) +{ + struct parser_params *p; + + TypedData_Get_Struct(self, struct parser_params, &parser_data_type, p); + if (!ripper_initialized_p(p)) { + rb_raise(rb_eArgError, "method called for uninitialized object"); + } + if (NIL_P(p->parsing_thread)) return Qnil; + return INT2NUM(p->lex.state); +} + +/* + * call-seq: + * ripper.token -> String + * + * Return the current token string. + */ +static VALUE +ripper_token(VALUE self) +{ + struct parser_params *p; + long pos, len; + + TypedData_Get_Struct(self, struct parser_params, &parser_data_type, p); + if (!ripper_initialized_p(p)) { + rb_raise(rb_eArgError, "method called for uninitialized object"); + } + if (NIL_P(p->parsing_thread)) return Qnil; + pos = p->lex.ptok - p->lex.pbeg; + len = p->lex.pcur - p->lex.ptok; + return rb_str_subseq(p->lex.lastline, pos, len); +} + +#ifdef RIPPER_DEBUG +/* :nodoc: */ +static VALUE +ripper_assert_Qundef(VALUE self, VALUE obj, VALUE msg) +{ + StringValue(msg); + if (obj == Qundef) { + rb_raise(rb_eArgError, "%"PRIsVALUE, msg); + } + return Qnil; +} + +/* :nodoc: */ +static VALUE +ripper_value(VALUE self, VALUE obj) +{ + return ULONG2NUM(obj); +} +#endif + +/* + * call-seq: + * Ripper.lex_state_name(integer) -> string + * + * Returns a string representation of lex_state. + */ +static VALUE +ripper_lex_state_name(VALUE self, VALUE state) +{ + return rb_parser_lex_state_name(NUM2INT(state)); +} + +void +Init_ripper(void) +{ + ripper_init_eventids1(); + ripper_init_eventids2(); + id_warn = rb_intern_const("warn"); + id_warning = rb_intern_const("warning"); + id_gets = rb_intern_const("gets"); + id_assoc = rb_intern_const("=>"); + + (void)yystpcpy; /* may not used in newer bison */ + + InitVM(ripper); +} + +void +InitVM_ripper(void) +{ + VALUE Ripper; + + Ripper = rb_define_class("Ripper", rb_cObject); + /* version of Ripper */ + rb_define_const(Ripper, "Version", rb_usascii_str_new2(RIPPER_VERSION)); + rb_define_alloc_func(Ripper, ripper_s_allocate); + rb_define_method(Ripper, "initialize", ripper_initialize, -1); + rb_define_method(Ripper, "parse", ripper_parse, 0); + rb_define_method(Ripper, "column", ripper_column, 0); + rb_define_method(Ripper, "filename", ripper_filename, 0); + rb_define_method(Ripper, "lineno", ripper_lineno, 0); + rb_define_method(Ripper, "state", ripper_state, 0); + rb_define_method(Ripper, "token", ripper_token, 0); + rb_define_method(Ripper, "end_seen?", rb_parser_end_seen_p, 0); + rb_define_method(Ripper, "encoding", rb_parser_encoding, 0); + rb_define_method(Ripper, "yydebug", rb_parser_get_yydebug, 0); + rb_define_method(Ripper, "yydebug=", rb_parser_set_yydebug, 1); + rb_define_method(Ripper, "debug_output", rb_parser_get_debug_output, 0); + rb_define_method(Ripper, "debug_output=", rb_parser_set_debug_output, 1); + rb_define_method(Ripper, "error?", ripper_error_p, 0); +#ifdef RIPPER_DEBUG + rb_define_method(Ripper, "assert_Qundef", ripper_assert_Qundef, 2); + rb_define_method(Ripper, "rawVALUE", ripper_value, 1); + rb_define_method(Ripper, "validate_object", ripper_validate_object, 1); +#endif + + rb_define_singleton_method(Ripper, "dedent_string", parser_dedent_string, 2); + rb_define_private_method(Ripper, "dedent_string", parser_dedent_string, 2); + + rb_define_singleton_method(Ripper, "lex_state_name", ripper_lex_state_name, 1); + +<% @exprs.each do |expr, desc| -%> + /* <%=desc%> */ + rb_define_const(Ripper, "<%=expr%>", INT2NUM(<%=expr%>)); +<% end %> + ripper_init_eventids1_table(Ripper); + ripper_init_eventids2_table(Ripper); + +# if 0 + /* Hack to let RDoc document SCRIPT_LINES__ */ + + /* + * When a Hash is assigned to +SCRIPT_LINES__+ the contents of files loaded + * after the assignment will be added as an Array of lines with the file + * name as the key. + */ + rb_define_global_const("SCRIPT_LINES__", Qnil); +#endif + +} +#endif /* RIPPER */ + +/* + * Local variables: + * mode: c + * c-file-style: "ruby" + * End: + */ |