/********************************************************************** parse.y - $Author$ created at: Fri May 28 18:02:42 JST 1993 Copyright (C) 1993-2007 Yukihiro Matsumoto **********************************************************************/ %{ #if !YYPURE # error needs pure parser #endif #define YYDEBUG 1 #define YYERROR_VERBOSE 1 #define YYSTACK_USE_ALLOCA 0 #define YYLTYPE rb_code_location_t #define YYLTYPE_IS_DECLARED 1 #include "ruby/ruby.h" #include "ruby/st.h" #include "ruby/encoding.h" #include "internal.h" #include "node.h" #include "parse.h" #include "symbol.h" #include "regenc.h" #include #include #include #include "probes.h" #ifndef WARN_PAST_SCOPE # define WARN_PAST_SCOPE 0 #endif #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 YYPRINT(out, tok, val) parser_token_value_print(p, (tok), &(val)) #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 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)) 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) \ (p->lex.state = \ (p->debug ? \ rb_parser_trace_lex_state(p, p->lex.state, (ls), __LINE__) : \ (enum lex_state_e)(ls))) typedef VALUE stack_type; 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_P() BITSTACK_SET_P(cond_stack) #define COND_SET(n) BITSTACK_SET(cond_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_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; }; #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; int linenum; int column; 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; 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 compile_option; VALUE debug_buffer; VALUE debug_output; ID cur_arg; rb_ast_t *ast; int node_id; 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 in_defined: 1; unsigned int in_main: 1; unsigned int in_kwarg: 1; unsigned int in_def: 1; unsigned int in_class: 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 warn_location: 1; NODE *eval_tree_begin; NODE *eval_tree; VALUE error_buffer; VALUE debug_lines; const struct rb_block *base_block; #else /* Ripper only */ VALUE delayed; int delayed_line; int delayed_col; VALUE value; VALUE result; VALUE parsing_thread; #endif }; #define new_tmpbuf() \ (rb_imemo_tmpbuf_t *)add_mark_object(p, rb_imemo_tmpbuf_auto_free_pointer(NULL)) #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 int parser_yyerror(struct parser_params*, const YYLTYPE *yylloc, const char*); #define yyerror0(msg) parser_yyerror(p, NULL, (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) #ifdef RIPPER #define compile_for_eval (0) #else #define compile_for_eval (p->base_block != 0 && !p->in_main) #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) 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 static inline VALUE add_mark_object(struct parser_params *p, VALUE obj) { if (!SPECIAL_CONST_P(obj) #ifdef RIPPER && !RB_TYPE_P(obj, T_NODE) /* Ripper jumbles NODE objects and other objects... */ #endif ) { rb_ast_add_mark_object(p->ast, obj); } return obj; } 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_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) = remove_begin(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 *evstr2dstr(struct parser_params*,NODE*); static NODE *splat_array(NODE*); 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 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_kw_arg(struct parser_params *p, NODE *k, const YYLTYPE *loc); 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*,const YYLTYPE*); static NODE *new_op_assign(struct parser_params *p, NODE *lhs, ID op, NODE *rhs, 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, 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_defined(struct parser_params *p, NODE *expr, const YYLTYPE *loc); static NODE *new_regexp(struct parser_params *, NODE *, int, const YYLTYPE *); #define make_array(ary, loc) ((ary) ? (nd_set_loc(ary, loc), ary) : NEW_ZARRAY(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 ID *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*); #define get_id(id) (id) #define get_value(val) (val) #else /* RIPPER */ #define NODE_RIPPER NODE_CDECL static inline VALUE ripper_new_yylval(struct parser_params *p, ID a, VALUE b, VALUE c) { add_mark_object(p, b); add_mark_object(p, c); return (VALUE)NEW_CDECL(a, b, c, &NULL_LOC); } static inline int ripper_is_node_yylval(VALUE n) { return RB_TYPE_P(n, T_NODE) && nd_type(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) 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, VALUE a); static VALUE parser_reg_compile(struct parser_params*, VALUE, int, 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); void rb_parser_set_location_from_strterm_heredoc(struct parser_params *p, rb_strterm_heredoc_t *here, YYLTYPE *yylloc); void rb_parser_set_location_of_none(struct parser_params *p, YYLTYPE *yylloc); void rb_parser_set_location(struct parser_params *p, YYLTYPE *yylloc); RUBY_SYMBOL_EXPORT_END static void parser_token_value_print(struct parser_params *p, enum yytokentype type, const YYSTYPE *valp); static ID formal_argument(struct parser_params*, ID); 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 **); static ID internal_id(struct parser_params*); 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); #ifdef RIPPER # define METHOD_NOT idNOT #else # define METHOD_NOT '!' #endif #define RE_OPTION_ONCE (1<<16) #define RE_OPTION_ENCODING_SHIFT 8 #define RE_OPTION_ENCODING(e) (((e)&0xff)<>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; struct rb_strterm_heredoc_struct { SIGNED_VALUE sourceline; /* lineno of the line that contains `<<"END"` */ VALUE term; /* `"END"` of `<<"END"` */ VALUE lastline; /* the string of line that contains `<<"END"` */ union { VALUE dummy; long lastidx; /* the column of `<<"END"` */ } u3; }; #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->term); rb_gc_mark(heredoc->lastline); } } #endif #define TOKEN2ID(tok) ( \ tTOKEN_LOCAL_BEGIN<(tok)&&(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; } #define new_defined(p,expr,loc) dispatch1(defined, (expr)) static VALUE heredoc_dedent(struct parser_params*,VALUE); #else #define ID2VAL(id) ((VALUE)(id)) #define TOKEN2VAL(t) ID2VAL(t) #define KWD2EID(t, v) keyword_##t #endif /* RIPPER */ #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; # 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 PRIsWARN "s" # 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 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 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 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); %} %pure-parser %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; const struct vtable *vars; struct rb_strterm_struct *strterm; } %token 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 tIDENTIFIER tFID tGVAR tIVAR tCONSTANT tCVAR tLABEL %token tINTEGER tFLOAT tRATIONAL tIMAGINARY tSTRING_CONTENT tCHAR %token tNTH_REF tBACK_REF %token tREGEXP_END %type singleton strings string string1 xstring regexp %type string_contents xstring_contents regexp_contents string_content %type words symbols symbol_list qwords qsymbols word_list qword_list qsym_list word %type literal numeric simple_numeric dsym cpath %type top_compstmt top_stmts top_stmt begin_block %type bodystmt compstmt stmts stmt_or_begin stmt expr arg primary command command_call method_call %type expr_value expr_value_do arg_value primary_value fcall rel_expr %type if_tail opt_else case_body cases opt_rescue exc_list exc_var opt_ensure %type args call_args opt_call_args %type paren_args opt_paren_args args_tail opt_args_tail block_args_tail opt_block_args_tail %type command_args aref_args opt_block_arg block_arg var_ref var_lhs %type command_rhs arg_rhs %type command_asgn mrhs mrhs_arg superclass block_call block_command %type f_block_optarg f_block_opt %type f_arglist f_args f_arg f_arg_item f_optarg f_marg f_marg_list f_margs %type assoc_list assocs assoc undef_list backref string_dvar for_var %type block_param opt_block_param block_param_def f_opt %type f_kwarg f_kw f_block_kwarg f_block_kw %type bv_decls opt_bv_decl bvar %type lambda f_larglist lambda_body brace_body do_body %type brace_block cmd_brace_block do_block lhs none fitem %type mlhs mlhs_head mlhs_basic mlhs_item mlhs_node mlhs_post mlhs_inner %type fsym keyword_variable user_variable sym symbol operation operation2 operation3 %type cname fname op f_rest_arg f_block_arg opt_f_block_arg f_norm_arg f_bad_arg %type f_kwrest f_label f_arg_asgn call_op call_op2 reswords relop dot_or_colon %token END_OF_INPUT 0 "end-of-input" %token '.' /* escaped chars, should be ignored otherwise */ %token '\\' "backslash" %token tSP "escaped space" %token '\t' "escaped horizontal tab" %token '\f' "escaped form feed" %token '\r' "escaped carriage return" %token '\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 tAREF RUBY_TOKEN(AREF) "[]" %token tASET RUBY_TOKEN(ASET) "[]=" %token tLSHFT RUBY_TOKEN(LSHFT) "<<" %token tRSHFT RUBY_TOKEN(RSHFT) ">>" %token tANDDOT RUBY_TOKEN(ANDDOT) "&." %token tCOLON2 RUBY_TOKEN(COLON2) "::" %token tCOLON3 ":: at EXPR_BEG" %token tOP_ASGN /* +=, -= 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 tSTRING_BEG tXSTRING_BEG tREGEXP_BEG tWORDS_BEG tQWORDS_BEG tSYMBOLS_BEG tQSYMBOLS_BEG %token tSTRING_DBEG tSTRING_DEND tSTRING_DVAR tSTRING_END tLAMBEG tLABEL_END /* * precedence table */ %nonassoc tLOWEST %nonassoc tLBRACE_ARG %nonassoc modifier_if modifier_unless modifier_while modifier_until %left keyword_or keyword_and %right keyword_not %nonassoc keyword_defined %right '=' tOP_ASGN %left modifier_rescue %right '?' ':' %nonassoc tDOT2 tDOT3 %left tOROP %left tANDOP %nonassoc tCMP tEQ tEQQ tNEQ tMATCH tNMATCH %left '>' tGEQ '<' tLEQ %left '|' '^' %left '&' %left tLSHFT tRSHFT %left '+' '-' %left '*' '/' '%' %right tUMINUS_NUM tUMINUS %right tPOW %right '!' '~' tUPLUS %token tLAST_TOKEN %% program : { SET_LEX_STATE(EXPR_BEG); local_push(p, ifndef_ripper(1)+0); } top_compstmt { /*%%%*/ if ($2 && !compile_for_eval) { NODE *node = $2; /* last expression should not be void */ if (nd_type(node) == NODE_BLOCK) { while (node->nd_next) { node = node->nd_next; } node = node->nd_head; } node = remove_begin(node); void_expr(p, node); } 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($2, &@$)); $$ = NEW_BEGIN(0, &@$); /*% %*/ /*% ripper: BEGIN!($2) %*/ } ; bodystmt : compstmt opt_rescue k_else {if (!$2) {yyerror1(&@3, "else without rescue is useless");}} compstmt opt_ensure { /*%%%*/ $$ = 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(p, $1); } ; stmts : none { /*%%%*/ $$ = NEW_BEGIN(0, &@$); /*% %*/ /*% ripper: stmts_add!(stmts_new!, void_stmt!) %*/ } | stmt_or_begin { /*%%%*/ $$ = newline_node($1); /*% %*/ /*% ripper: stmts_add!(stmts_new!, $1) %*/ } | stmts terms stmt_or_begin { /*%%%*/ $$ = block_append(p, $1, newline_node($3)); /*% %*/ /*% ripper: stmts_add!($1, $3) %*/ } | error stmt { $$ = remove_begin($2); } ; stmt_or_begin : stmt { $$ = $1; } | keyword_BEGIN { yyerror1(&@1, "BEGIN is permitted only at toplevel"); } begin_block { $$ = $3; } stmt : keyword_alias fitem {SET_LEX_STATE(EXPR_FNAME|EXPR_FITEM);} fitem { /*%%%*/ $$ = NEW_ALIAS($2, $4, &@$); /*% %*/ /*% ripper: alias!($2, $4) %*/ } | keyword_alias tGVAR tGVAR { /*%%%*/ $$ = 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; $$ = NEW_VALIAS($2, rb_intern2(buf, 2), &@$); /*% %*/ /*% ripper: var_alias!($2, $3) %*/ } | keyword_alias tGVAR tNTH_REF { /*%%%*/ yyerror1(&@3, "can't make alias for the number variables"); $$ = NEW_BEGIN(0, &@$); /*% %*/ /*% ripper[error]: alias_error!(var_alias!($2, $3)) %*/ } | keyword_undef undef_list { /*%%%*/ $$ = $2; /*% %*/ /*% ripper: undef!($2) %*/ } | stmt modifier_if expr_value { /*%%%*/ $$ = new_if(p, $3, remove_begin($1), 0, &@$); fixpos($$, $3); /*% %*/ /*% ripper: if_mod!($3, $1) %*/ } | stmt modifier_unless expr_value { /*%%%*/ $$ = new_unless(p, $3, remove_begin($1), 0, &@$); fixpos($$, $3); /*% %*/ /*% ripper: unless_mod!($3, $1) %*/ } | stmt modifier_while expr_value { /*%%%*/ if ($1 && nd_type($1) == NODE_BEGIN) { $$ = NEW_WHILE(cond(p, $3, &@3), $1->nd_body, 0, &@$); } else { $$ = NEW_WHILE(cond(p, $3, &@3), $1, 1, &@$); } /*% %*/ /*% ripper: while_mod!($3, $1) %*/ } | stmt modifier_until expr_value { /*%%%*/ if ($1 && nd_type($1) == NODE_BEGIN) { $$ = NEW_UNTIL(cond(p, $3, &@3), $1->nd_body, 0, &@$); } else { $$ = NEW_UNTIL(cond(p, $3, &@3), $1, 1, &@$); } /*% %*/ /*% ripper: until_mod!($3, $1) %*/ } | stmt modifier_rescue stmt { /*%%%*/ NODE *resq; YYLTYPE loc = code_loc_gen(&@2, &@3); resq = NEW_RESBODY(0, remove_begin($3), 0, &loc); $$ = NEW_RESCUE(remove_begin($1), resq, 0, &@$); /*% %*/ /*% ripper: rescue_mod!($1, $3) %*/ } | keyword_END '{' compstmt '}' { if (p->in_def) { rb_warn0("END in method; use at_exit"); } /*%%%*/ { NODE *scope = NEW_NODE( NODE_SCOPE, 0 /* tbl */, $3 /* body */, 0 /* args */, &@$); $$ = NEW_POSTEXE(scope, &@$); } /*% %*/ /*% ripper: END!($3) %*/ } | command_asgn | mlhs '=' command_call { /*%%%*/ value_expr($3); $$ = node_assign(p, $1, $3, &@$); /*% %*/ /*% ripper: massign!($1, $3) %*/ } | lhs '=' mrhs { /*%%%*/ value_expr($3); $$ = node_assign(p, $1, $3, &@$); /*% %*/ /*% ripper: assign!($1, $3) %*/ } | mlhs '=' mrhs_arg { /*%%%*/ $$ = node_assign(p, $1, $3, &@$); /*% %*/ /*% ripper: massign!($1, $3) %*/ } | expr ; command_asgn : lhs '=' command_rhs { /*%%%*/ value_expr($3); $$ = node_assign(p, $1, $3, &@$); /*% %*/ /*% ripper: assign!($1, $3) %*/ } | var_lhs tOP_ASGN command_rhs { /*%%%*/ value_expr($3); $$ = new_op_assign(p, $1, $2, $3, &@$); /*% %*/ /*% ripper: opassign!($1, $2, $3) %*/ } | primary_value '[' opt_call_args rbracket tOP_ASGN command_rhs { /*%%%*/ value_expr($6); $$ = new_ary_op_assign(p, $1, $3, $5, $6, &@3, &@$); /*% %*/ /*% ripper: opassign!(aref_field!($1, escape_Qundef($3)), $5, $6) %*/ } | primary_value call_op tIDENTIFIER tOP_ASGN command_rhs { /*%%%*/ value_expr($5); $$ = new_attr_op_assign(p, $1, $2, $3, $4, $5, &@$); /*% %*/ /*% ripper: opassign!(field!($1, $2, $3), $4, $5) %*/ } | primary_value call_op tCONSTANT tOP_ASGN command_rhs { /*%%%*/ value_expr($5); $$ = new_attr_op_assign(p, $1, $2, $3, $4, $5, &@$); /*% %*/ /*% ripper: opassign!(field!($1, $2, $3), $4, $5) %*/ } | primary_value tCOLON2 tCONSTANT tOP_ASGN command_rhs { /*%%%*/ YYLTYPE loc = code_loc_gen(&@1, &@3); $$ = new_const_op_assign(p, NEW_COLON2($1, $3, &loc), $4, $5, &@$); /*% %*/ /*% ripper: opassign!(const_path_field!($1, $3), $4, $5) %*/ } | primary_value tCOLON2 tIDENTIFIER tOP_ASGN command_rhs { /*%%%*/ value_expr($5); $$ = new_attr_op_assign(p, $1, ID2VAL(idCOLON2), $3, $4, $5, &@$); /*% %*/ /*% ripper: opassign!(field!($1, ID2VAL(idCOLON2), $3), $4, $5) %*/ } | backref tOP_ASGN command_rhs { /*%%%*/ rb_backref_error(p, $1); $$ = NEW_BEGIN(0, &@$); /*% %*/ /*% ripper[error]: assign_error!(assign!(var_field(p, $1), $3)) %*/ } ; command_rhs : command_call %prec tOP_ASGN { value_expr($1); $$ = $1; } | command_call modifier_rescue stmt { /*%%%*/ YYLTYPE loc = code_loc_gen(&@2, &@3); value_expr($1); $$ = NEW_RESCUE($1, NEW_RESBODY(0, remove_begin($3), 0, &loc), 0, &@$); /*% %*/ /*% ripper: rescue_mod!($1, $3) %*/ } | command_asgn ; expr : command_call | expr keyword_and expr { $$ = logop(p, idAND, $1, $3, &@2, &@$); } | expr keyword_or expr { $$ = logop(p, idOR, $1, $3, &@2, &@$); } | keyword_not opt_nl expr { $$ = call_uni_op(p, method_cond(p, $3, &@3), METHOD_NOT, &@1, &@$); } | '!' command_call { $$ = call_uni_op(p, method_cond(p, $2, &@2), '!', &@1, &@$); } | arg ; expr_value : expr { value_expr($1); $$ = $1; } ; expr_value_do : {COND_PUSH(1);} expr_value do {COND_POP();} { $$ = $2; } command_call : command | block_command ; block_command : block_call | block_call call_op2 operation2 command_args { /*%%%*/ $$ = new_qcall(p, $2, $1, $3, $4, &@3, &@$); /*% %*/ /*% ripper: method_add_arg!(call!($1, $2, $3), $4) %*/ } ; cmd_brace_block : tLBRACE_ARG brace_body '}' { $$ = $2; /*%%%*/ $$->nd_body->nd_loc = code_loc_gen(&@1, &@3); nd_set_line($$, @1.end_pos.lineno); /*% %*/ } ; fcall : operation { /*%%%*/ $$ = NEW_FCALL($1, 0, &@$); nd_set_line($$, p->tokline); /*% %*/ /*% ripper: $1 %*/ } ; command : fcall command_args %prec tLOWEST { /*%%%*/ $1->nd_args = $2; nd_set_last_loc($1, @2.end_pos); $$ = $1; /*% %*/ /*% ripper: command!($1, $2) %*/ } | fcall command_args cmd_brace_block { /*%%%*/ block_dup_check(p, $2, $3); $1->nd_args = $2; $$ = 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_command_qcall(p, $2, $1, $3, $4, Qnull, &@3, &@$); /*% %*/ /*% ripper: command_call!($1, $2, $3, $4) %*/ } | primary_value call_op operation2 command_args cmd_brace_block { /*%%%*/ $$ = new_command_qcall(p, $2, $1, $3, $4, $5, &@3, &@$); /*% %*/ /*% ripper: method_add_block!(command_call!($1, $2, $3, $4), $5) %*/ } | primary_value tCOLON2 operation2 command_args %prec tLOWEST { /*%%%*/ $$ = new_command_qcall(p, ID2VAL(idCOLON2), $1, $3, $4, Qnull, &@3, &@$); /*% %*/ /*% ripper: command_call!($1, ID2VAL(idCOLON2), $3, $4) %*/ } | primary_value tCOLON2 operation2 command_args cmd_brace_block { /*%%%*/ $$ = new_command_qcall(p, ID2VAL(idCOLON2), $1, $3, $4, $5, &@3, &@$); /*% %*/ /*% ripper: method_add_block!(command_call!($1, ID2VAL(idCOLON2), $3, $4), $5) %*/ } | keyword_super command_args { /*%%%*/ $$ = 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 { /*%%%*/ $$ = NEW_BREAK(ret_args(p, $2), &@$); /*% %*/ /*% ripper: break!($2) %*/ } | keyword_next call_args { /*%%%*/ $$ = NEW_NEXT(ret_args(p, $2), &@$); /*% %*/ /*% ripper: next!($2) %*/ } ; mlhs : mlhs_basic | tLPAREN mlhs_inner rparen { /*%%%*/ $$ = $2; /*% %*/ /*% ripper: mlhs_paren!($2) %*/ } ; mlhs_inner : mlhs_basic | tLPAREN mlhs_inner rparen { /*%%%*/ $$ = NEW_MASGN(NEW_LIST($2, &@$), 0, &@$); /*% %*/ /*% ripper: mlhs_paren!($2) %*/ } ; mlhs_basic : mlhs_head { /*%%%*/ $$ = NEW_MASGN($1, 0, &@$); /*% %*/ /*% ripper: $1 %*/ } | mlhs_head mlhs_item { /*%%%*/ $$ = NEW_MASGN(list_append(p, $1,$2), 0, &@$); /*% %*/ /*% ripper: mlhs_add!($1, $2) %*/ } | mlhs_head tSTAR mlhs_node { /*%%%*/ $$ = NEW_MASGN($1, $3, &@$); /*% %*/ /*% ripper: mlhs_add_star!($1, $3) %*/ } | mlhs_head tSTAR mlhs_node ',' mlhs_post { /*%%%*/ $$ = NEW_MASGN($1, NEW_POSTARG($3,$5,&@$), &@$); /*% %*/ /*% ripper: mlhs_add_post!(mlhs_add_star!($1, $3), $5) %*/ } | mlhs_head tSTAR { /*%%%*/ $$ = NEW_MASGN($1, NODE_SPECIAL_NO_NAME_REST, &@$); /*% %*/ /*% ripper: mlhs_add_star!($1, Qnil) %*/ } | mlhs_head tSTAR ',' mlhs_post { /*%%%*/ $$ = NEW_MASGN($1, NEW_POSTARG(NODE_SPECIAL_NO_NAME_REST, $4, &@$), &@$); /*% %*/ /*% ripper: mlhs_add_post!(mlhs_add_star!($1, Qnil), $4) %*/ } | tSTAR mlhs_node { /*%%%*/ $$ = NEW_MASGN(0, $2, &@$); /*% %*/ /*% ripper: mlhs_add_star!(mlhs_new!, $2) %*/ } | tSTAR mlhs_node ',' mlhs_post { /*%%%*/ $$ = NEW_MASGN(0, NEW_POSTARG($2,$4,&@$), &@$); /*% %*/ /*% ripper: mlhs_add_post!(mlhs_add_star!(mlhs_new!, $2), $4) %*/ } | tSTAR { /*%%%*/ $$ = NEW_MASGN(0, NODE_SPECIAL_NO_NAME_REST, &@$); /*% %*/ /*% ripper: mlhs_add_star!(mlhs_new!, Qnil) %*/ } | tSTAR ',' mlhs_post { /*%%%*/ $$ = 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_inner rparen { /*%%%*/ $$ = $2; /*% %*/ /*% ripper: mlhs_paren!($2) %*/ } ; mlhs_head : mlhs_item ',' { /*%%%*/ $$ = NEW_LIST($1, &@1); /*% %*/ /*% ripper: mlhs_add!(mlhs_new!, $1) %*/ } | mlhs_head mlhs_item ',' { /*%%%*/ $$ = list_append(p, $1, $2); /*% %*/ /*% ripper: mlhs_add!($1, $2) %*/ } ; mlhs_post : mlhs_item { /*%%%*/ $$ = NEW_LIST($1, &@$); /*% %*/ /*% ripper: mlhs_add!(mlhs_new!, $1) %*/ } | mlhs_post ',' mlhs_item { /*%%%*/ $$ = list_append(p, $1, $3); /*% %*/ /*% ripper: mlhs_add!($1, $3) %*/ } ; mlhs_node : user_variable { /*%%%*/ $$ = 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 { /*%%%*/ $$ = attrset(p, $1, $2, $3, &@$); /*% %*/ /*% ripper: field!($1, $2, $3) %*/ } | primary_value tCOLON2 tIDENTIFIER { /*%%%*/ $$ = attrset(p, $1, idCOLON2, $3, &@$); /*% %*/ /*% ripper: const_path_field!($1, $3) %*/ } | primary_value call_op tCONSTANT { /*%%%*/ $$ = attrset(p, $1, $2, $3, &@$); /*% %*/ /*% ripper: field!($1, $2, $3) %*/ } | primary_value tCOLON2 tCONSTANT { /*%%%*/ $$ = const_decl(p, NEW_COLON2($1, $3, &@$), &@$); /*% %*/ /*% ripper: const_decl(p, const_path_field!($1, $3)) %*/ } | tCOLON3 tCONSTANT { /*%%%*/ $$ = const_decl(p, NEW_COLON3($2, &@$), &@$); /*% %*/ /*% ripper: const_decl(p, top_const_field!($2)) %*/ } | backref { /*%%%*/ rb_backref_error(p, $1); $$ = NEW_BEGIN(0, &@$); /*% %*/ /*% ripper[error]: assign_error!(var_field(p, $1)) %*/ } ; lhs : 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(p, $1, $2, $3, &@$); /*% %*/ /*% ripper: field!($1, $2, $3) %*/ } | primary_value tCOLON2 tIDENTIFIER { /*%%%*/ $$ = attrset(p, $1, idCOLON2, $3, &@$); /*% %*/ /*% ripper: field!($1, ID2VAL(idCOLON2), $3) %*/ } | primary_value call_op tCONSTANT { /*%%%*/ $$ = attrset(p, $1, $2, $3, &@$); /*% %*/ /*% ripper: field!($1, $2, $3) %*/ } | primary_value tCOLON2 tCONSTANT { /*%%%*/ $$ = const_decl(p, NEW_COLON2($1, $3, &@$), &@$); /*% %*/ /*% ripper: const_decl(p, const_path_field!($1, $3)) %*/ } | tCOLON3 tCONSTANT { /*%%%*/ $$ = const_decl(p, NEW_COLON3($2, &@$), &@$); /*% %*/ /*% ripper: const_decl(p, top_const_field!($2)) %*/ } | backref { /*%%%*/ rb_backref_error(p, $1); $$ = NEW_BEGIN(0, &@$); /*% %*/ /*% ripper[error]: assign_error!(var_field(p, $1)) %*/ } ; cname : tIDENTIFIER { /*%%%*/ yyerror1(&@1, "class/module name must be CONSTANT"); /*% %*/ /*% ripper[error]: class_name_error!($1) %*/ } | tCONSTANT ; cpath : tCOLON3 cname { /*%%%*/ $$ = NEW_COLON3($2, &@$); /*% %*/ /*% ripper: top_const_ref!($2) %*/ } | cname { /*%%%*/ $$ = NEW_COLON2(0, $$, &@$); /*% %*/ /*% ripper: const_ref!($1) %*/ } | primary_value tCOLON2 cname { /*%%%*/ $$ = NEW_COLON2($1, $3, &@$); /*% %*/ /*% ripper: const_path_ref!($1, $3) %*/ } ; fname : tIDENTIFIER | tCONSTANT | tFID | op { SET_LEX_STATE(EXPR_ENDFN); $$ = $1; } | reswords { SET_LEX_STATE(EXPR_ENDFN); $$ = $1; } ; fsym : fname | symbol ; fitem : fsym { /*%%%*/ $$ = NEW_LIT(ID2SYM($1), &@$); /*% %*/ /*% ripper: symbol_literal!($1) %*/ } | dsym ; undef_list : fitem { /*%%%*/ $$ = NEW_UNDEF($1, &@$); /*% %*/ /*% ripper: rb_ary_new3(1, get_value($1)) %*/ } | undef_list ',' {SET_LEX_STATE(EXPR_FNAME|EXPR_FITEM);} fitem { /*%%%*/ NODE *undef = NEW_UNDEF($4, &@4); $$ = block_append(p, $1, undef); /*% %*/ /*% ripper: rb_ary_push($1, get_value($4)) %*/ } ; 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 : 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_rhs { /*%%%*/ $$ = node_assign(p, $1, $3, &@$); /*% %*/ /*% ripper: assign!($1, $3) %*/ } | var_lhs tOP_ASGN arg_rhs { /*%%%*/ $$ = new_op_assign(p, $1, $2, $3, &@$); /*% %*/ /*% ripper: opassign!($1, $2, $3) %*/ } | primary_value '[' opt_call_args rbracket tOP_ASGN arg_rhs { /*%%%*/ value_expr($6); $$ = new_ary_op_assign(p, $1, $3, $5, $6, &@3, &@$); /*% %*/ /*% ripper: opassign!(aref_field!($1, escape_Qundef($3)), $5, $6) %*/ } | primary_value call_op tIDENTIFIER tOP_ASGN arg_rhs { /*%%%*/ value_expr($5); $$ = new_attr_op_assign(p, $1, $2, $3, $4, $5, &@$); /*% %*/ /*% ripper: opassign!(field!($1, $2, $3), $4, $5) %*/ } | primary_value call_op tCONSTANT tOP_ASGN arg_rhs { /*%%%*/ value_expr($5); $$ = new_attr_op_assign(p, $1, $2, $3, $4, $5, &@$); /*% %*/ /*% ripper: opassign!(field!($1, $2, $3), $4, $5) %*/ } | primary_value tCOLON2 tIDENTIFIER tOP_ASGN arg_rhs { /*%%%*/ value_expr($5); $$ = new_attr_op_assign(p, $1, ID2VAL(idCOLON2), $3, $4, $5, &@$); /*% %*/ /*% ripper: opassign!(field!($1, ID2VAL(idCOLON2), $3), $4, $5) %*/ } | primary_value tCOLON2 tCONSTANT tOP_ASGN arg_rhs { /*%%%*/ YYLTYPE loc = code_loc_gen(&@1, &@3); $$ = new_const_op_assign(p, NEW_COLON2($1, $3, &loc), $4, $5, &@$); /*% %*/ /*% ripper: opassign!(const_path_field!($1, $3), $4, $5) %*/ } | tCOLON3 tCONSTANT tOP_ASGN arg_rhs { /*%%%*/ $$ = new_const_op_assign(p, NEW_COLON3($2, &@$), $3, $4, &@$); /*% %*/ /*% ripper: opassign!(top_const_field!($2), $3, $4) %*/ } | backref tOP_ASGN arg_rhs { /*%%%*/ rb_backref_error(p, $1); $$ = NEW_BEGIN(0, &@$); /*% %*/ /*% ripper[error]: assign_error!(opassign!(var_field(p, $1), $2, $3)) %*/ } | arg tDOT2 arg { /*%%%*/ value_expr($1); value_expr($3); $$ = NEW_DOT2($1, $3, &@$); /*% %*/ /*% ripper: dot2!($1, $3) %*/ } | arg tDOT3 arg { /*%%%*/ value_expr($1); value_expr($3); $$ = NEW_DOT3($1, $3, &@$); /*% %*/ /*% ripper: dot3!($1, $3) %*/ } | arg tDOT2 { /*%%%*/ YYLTYPE loc; loc.beg_pos = @2.end_pos; loc.end_pos = @2.end_pos; value_expr($1); $$ = NEW_DOT2($1, new_nil(&loc), &@$); /*% %*/ /*% ripper: dot2!($1, Qnil) %*/ } | arg tDOT3 { /*%%%*/ YYLTYPE loc; loc.beg_pos = @2.end_pos; loc.end_pos = @2.end_pos; value_expr($1); $$ = NEW_DOT3($1, new_nil(&loc), &@$); /*% %*/ /*% ripper: dot3!($1, Qnil) %*/ } | arg '+' arg { $$ = call_bin_op(p, $1, '+', $3, &@2, &@$); } | arg '-' arg { $$ = call_bin_op(p, $1, '-', $3, &@2, &@$); } | arg '*' arg { $$ = call_bin_op(p, $1, '*', $3, &@2, &@$); } | arg '/' arg { $$ = call_bin_op(p, $1, '/', $3, &@2, &@$); } | arg '%' arg { $$ = call_bin_op(p, $1, '%', $3, &@2, &@$); } | arg tPOW arg { $$ = call_bin_op(p, $1, idPow, $3, &@2, &@$); } | tUMINUS_NUM simple_numeric tPOW arg { $$ = call_uni_op(p, call_bin_op(p, $2, idPow, $4, &@2, &@$), idUMinus, &@1, &@$); } | tUPLUS arg { $$ = call_uni_op(p, $2, idUPlus, &@1, &@$); } | tUMINUS arg { $$ = call_uni_op(p, $2, idUMinus, &@1, &@$); } | arg '|' arg { $$ = call_bin_op(p, $1, '|', $3, &@2, &@$); } | arg '^' arg { $$ = call_bin_op(p, $1, '^', $3, &@2, &@$); } | arg '&' arg { $$ = call_bin_op(p, $1, '&', $3, &@2, &@$); } | arg tCMP arg { $$ = call_bin_op(p, $1, idCmp, $3, &@2, &@$); } | rel_expr %prec tCMP | arg tEQ arg { $$ = call_bin_op(p, $1, idEq, $3, &@2, &@$); } | arg tEQQ arg { $$ = call_bin_op(p, $1, idEqq, $3, &@2, &@$); } | arg tNEQ arg { $$ = call_bin_op(p, $1, idNeq, $3, &@2, &@$); } | arg tMATCH arg { $$ = match_op(p, $1, $3, &@2, &@$); } | arg tNMATCH arg { $$ = call_bin_op(p, $1, idNeqTilde, $3, &@2, &@$); } | '!' arg { $$ = call_uni_op(p, method_cond(p, $2, &@2), '!', &@1, &@$); } | '~' arg { $$ = call_uni_op(p, $2, '~', &@1, &@$); } | arg tLSHFT arg { $$ = call_bin_op(p, $1, idLTLT, $3, &@2, &@$); } | arg tRSHFT arg { $$ = call_bin_op(p, $1, idGTGT, $3, &@2, &@$); } | arg tANDOP arg { $$ = logop(p, idANDOP, $1, $3, &@2, &@$); } | arg tOROP arg { $$ = logop(p, idOROP, $1, $3, &@2, &@$); } | keyword_defined opt_nl {p->in_defined = 1;} arg { p->in_defined = 0; $$ = new_defined(p, $4, &@$); } | arg '?' arg opt_nl ':' arg { /*%%%*/ value_expr($1); $$ = new_if(p, $1, $3, $6, &@$); fixpos($$, $1); /*% %*/ /*% ripper: ifop!($1, $3, $6) %*/ } | primary { $$ = $1; } ; relop : '>' {$$ = '>';} | '<' {$$ = '<';} | tGEQ {$$ = 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)); $$ = call_bin_op(p, $1, $2, $3, &@2, &@$); } ; arg_value : arg { value_expr($1); $$ = $1; } ; aref_args : none | args trailer { $$ = $1; } | args ',' assocs trailer { /*%%%*/ $$ = $3 ? arg_append(p, $1, new_hash(p, $3, &@3), &@$) : $1; /*% %*/ /*% ripper: args_add!($1, bare_assoc_hash!($3)) %*/ } | assocs trailer { /*%%%*/ $$ = $1 ? NEW_LIST(new_hash(p, $1, &@1), &@$) : 0; /*% %*/ /*% ripper: args_add!(args_new!, bare_assoc_hash!($1)) %*/ } ; arg_rhs : arg %prec tOP_ASGN { value_expr($1); $$ = $1; } | arg modifier_rescue arg { /*%%%*/ YYLTYPE loc = code_loc_gen(&@2, &@3); value_expr($1); $$ = NEW_RESCUE($1, NEW_RESBODY(0, remove_begin($3), 0, &loc), 0, &@$); /*% %*/ /*% ripper: rescue_mod!($1, $3) %*/ } ; paren_args : '(' opt_call_args rparen { /*%%%*/ $$ = $2; /*% %*/ /*% ripper: arg_paren!(escape_Qundef($2)) %*/ } ; opt_paren_args : none | paren_args ; opt_call_args : none | call_args | args ',' { $$ = $1; } | args ',' assocs ',' { /*%%%*/ $$ = $3 ? arg_append(p, $1, new_hash(p, $3, &@3), &@$) : $1; /*% %*/ /*% ripper: args_add!($1, bare_assoc_hash!($3)) %*/ } | assocs ',' { /*%%%*/ $$ = $1 ? NEW_LIST(new_hash(p, $1, &@1), &@1) : 0; /*% %*/ /*% ripper: args_add!(args_new!, bare_assoc_hash!($1)) %*/ } ; call_args : command { /*%%%*/ value_expr($1); $$ = NEW_LIST($1, &@$); /*% %*/ /*% ripper: args_add!(args_new!, $1) %*/ } | args opt_block_arg { /*%%%*/ $$ = arg_blk_pass($1, $2); /*% %*/ /*% ripper: args_add_block!($1, $2) %*/ } | assocs opt_block_arg { /*%%%*/ $$ = $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) %*/ } | args ',' assocs opt_block_arg { /*%%%*/ $$ = $3 ? arg_append(p, $1, new_hash(p, $3, &@3), &@$) : $1; $$ = arg_blk_pass($$, $4); /*% %*/ /*% ripper: args_add_block!(args_add!($1, bare_assoc_hash!($3)), $4) %*/ } | block_arg /*% ripper[brace]: args_add_block!(args_new!, $1) %*/ ; command_args : { /* If call_args starts with a open paren '(' or '[', * look-ahead reading of the letters calls CMDARG_PUSH(0), * but the push must be done after CMDARG_PUSH(1). * So this code makes them consistent by first cancelling * the premature CMDARG_PUSH(0), doing CMDARG_PUSH(1), * and finally redoing CMDARG_PUSH(0). */ int lookahead = 0; switch (yychar) { case '(': case tLPAREN: case tLPAREN_ARG: case '[': case tLBRACK: lookahead = 1; } if (lookahead) CMDARG_POP(); CMDARG_PUSH(1); if (lookahead) CMDARG_PUSH(0); } call_args { /* call_args can be followed by tLBRACE_ARG (that does CMDARG_PUSH(0) in the lexer) * but the push must be done after CMDARG_POP() in the parser. * So this code does CMDARG_POP() to pop 0 pushed by tLBRACE_ARG, * CMDARG_POP() to pop 1 pushed by command_args, * and CMDARG_PUSH(0) to restore back the flag set by tLBRACE_ARG. */ int lookahead = 0; switch (yychar) { case tLBRACE_ARG: lookahead = 1; } if (lookahead) CMDARG_POP(); CMDARG_POP(); if (lookahead) CMDARG_PUSH(0); $$ = $2; } ; block_arg : tAMPER arg_value { /*%%%*/ $$ = NEW_BLOCK_PASS($2, &@$); /*% %*/ /*% ripper: $2 %*/ } ; opt_block_arg : ',' block_arg { $$ = $2; } | none { $$ = 0; } ; args : arg_value { /*%%%*/ $$ = NEW_LIST($1, &@$); /*% %*/ /*% ripper: args_add!(args_new!, $1) %*/ } | tSTAR arg_value { /*%%%*/ $$ = NEW_SPLAT($2, &@$); /*% %*/ /*% ripper: args_add_star!(args_new!, $2) %*/ } | args ',' arg_value { /*%%%*/ $$ = last_arg_append(p, $1, $3, &@$); /*% %*/ /*% ripper: args_add!($1, $3) %*/ } | args ',' tSTAR arg_value { /*%%%*/ $$ = rest_arg_append(p, $1, $4, &@$); /*% %*/ /*% ripper: args_add_star!($1, $4) %*/ } ; mrhs_arg : mrhs | arg_value ; mrhs : args ',' arg_value { /*%%%*/ $$ = last_arg_append(p, $1, $3, &@$); /*% %*/ /*% ripper: mrhs_add!(mrhs_new_from_args!($1), $3) %*/ } | args ',' tSTAR arg_value { /*%%%*/ $$ = rest_arg_append(p, $1, $4, &@$); /*% %*/ /*% ripper: mrhs_add_star!(mrhs_new_from_args!($1), $4) %*/ } | tSTAR arg_value { /*%%%*/ $$ = NEW_SPLAT($2, &@$); /*% %*/ /*% ripper: mrhs_add_star!(mrhs_new!, $2) %*/ } ; primary : literal | strings | xstring | regexp | words | qwords | symbols | qsymbols | var_ref | backref | tFID { /*%%%*/ $$ = NEW_FCALL($1, 0, &@$); /*% %*/ /*% ripper: method_add_arg!(fcall!($1), args_new!) %*/ } | k_begin { CMDARG_PUSH(0); } bodystmt k_end { CMDARG_POP(); /*%%%*/ set_line_body($3, @1.end_pos.lineno); $$ = 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 stmt {SET_LEX_STATE(EXPR_ENDARG);} rparen { /*%%%*/ $$ = $2; /*% %*/ /*% ripper: paren!($2) %*/ } | tLPAREN compstmt ')' { /*%%%*/ $$ = $2; /*% %*/ /*% ripper: paren!($2) %*/ } | primary_value tCOLON2 tCONSTANT { /*%%%*/ $$ = NEW_COLON2($1, $3, &@$); /*% %*/ /*% ripper: const_path_ref!($1, $3) %*/ } | tCOLON3 tCONSTANT { /*%%%*/ $$ = NEW_COLON3($2, &@$); /*% %*/ /*% ripper: top_const_ref!($2) %*/ } | tLBRACK aref_args ']' { /*%%%*/ $$ = make_array($2, &@$); /*% %*/ /*% ripper: array!(escape_Qundef($2)) %*/ } | tLBRACE assoc_list '}' { /*%%%*/ $$ = new_hash(p, $2, &@$); $$->nd_alen = TRUE; /*% %*/ /*% ripper: hash!(escape_Qundef($2)) %*/ } | k_return { /*%%%*/ $$ = NEW_RETURN(0, &@$); /*% %*/ /*% ripper: return0! %*/ } | keyword_yield '(' call_args rparen { /*%%%*/ $$ = new_yield(p, $3, &@$); /*% %*/ /*% ripper: yield!(paren!($3)) %*/ } | keyword_yield '(' rparen { /*%%%*/ $$ = NEW_YIELD(0, &@$); /*% %*/ /*% ripper: yield!(paren!(args_new!)) %*/ } | keyword_yield { /*%%%*/ $$ = NEW_YIELD(0, &@$); /*% %*/ /*% ripper: yield0! %*/ } | keyword_defined opt_nl '(' {p->in_defined = 1;} expr rparen { p->in_defined = 0; $$ = new_defined(p, $5, &@$); } | keyword_not '(' expr rparen { $$ = call_uni_op(p, method_cond(p, $3, &@3), METHOD_NOT, &@1, &@$); } | keyword_not '(' rparen { $$ = call_uni_op(p, method_cond(p, new_nil(&@2), &@2), METHOD_NOT, &@1, &@$); } | fcall brace_block { /*%%%*/ $$ = method_add_block(p, $1, $2, &@$); /*% %*/ /*% ripper: method_add_block!(method_add_arg!(fcall!($1), args_new!), $2) %*/ } | method_call | method_call brace_block { /*%%%*/ block_dup_check(p, $1->nd_args, $2); $$ = method_add_block(p, $1, $2, &@$); /*% %*/ /*% ripper: method_add_block!($1, $2) %*/ } | tLAMBDA { token_info_push(p, "->", &@1); } lambda { $$ = $3; /*%%%*/ nd_set_first_loc($$, @1.beg_pos); /*% %*/ } | k_if expr_value then compstmt if_tail k_end { /*%%%*/ $$ = new_if(p, $2, $4, $5, &@$); fixpos($$, $2); /*% %*/ /*% ripper: if!($2, $4, escape_Qundef($5)) %*/ } | k_unless expr_value then compstmt opt_else k_end { /*%%%*/ $$ = new_unless(p, $2, $4, $5, &@$); fixpos($$, $2); /*% %*/ /*% ripper: unless!($2, $4, escape_Qundef($5)) %*/ } | k_while expr_value_do compstmt k_end { /*%%%*/ $$ = NEW_WHILE(cond(p, $2, &@2), $3, 1, &@$); fixpos($$, $2); /*% %*/ /*% ripper: while!($2, $3) %*/ } | k_until expr_value_do compstmt k_end { /*%%%*/ $$ = NEW_UNTIL(cond(p, $2, &@2), $3, 1, &@$); fixpos($$, $2); /*% %*/ /*% ripper: until!($2, $3) %*/ } | k_case expr_value opt_terms case_body k_end { /*%%%*/ $$ = NEW_CASE($2, $4, &@$); fixpos($$, $2); /*% %*/ /*% ripper: case!($2, $4) %*/ } | k_case opt_terms case_body k_end { /*%%%*/ $$ = NEW_CASE2($3, &@$); /*% %*/ /*% ripper: case!(Qnil, $3) %*/ } | k_for for_var keyword_in expr_value_do compstmt k_end { /*%%%*/ /* * for a, b, c in e * #=> * e.each{|*x| a, b, c = x} * * for a in e * #=> * e.each{|x| a, = x} */ ID id = internal_id(p); NODE *m = NEW_ARGS_AUX(0, 0, &NULL_LOC); NODE *args, *scope, *internal_var = NEW_DVAR(id, &@2); rb_imemo_tmpbuf_t *tmpbuf = new_tmpbuf(); ID *tbl = ALLOC_N(ID, 2); tbl[0] = 1 /* length of local var table */; tbl[1] = id /* internal id */; tmpbuf->ptr = (VALUE *)tbl; switch (nd_type($2)) { case NODE_LASGN: case NODE_DASGN: case NODE_DASGN_CURR: /* e.each {|internal_var| a = internal_var; ... } */ $2->nd_value = internal_var; id = 0; m->nd_plen = 1; m->nd_next = $2; break; case NODE_MASGN: /* e.each {|*internal_var| a, b, c = (internal_var.length == 1 && Array === (tmp = internal_var[0]) ? tmp : internal_var); ... } */ m->nd_next = node_assign(p, $2, NEW_FOR_MASGN(internal_var, &@2), &@2); break; default: /* e.each {|*internal_var| @a, B, c[1], d.attr = internal_val; ... } */ m->nd_next = node_assign(p, NEW_MASGN(NEW_LIST($2, &@2), 0, &@2), internal_var, &@2); } /* {|*internal_id| = internal_id; ... } */ args = new_args(p, m, 0, id, 0, new_args_tail(p, 0, 0, 0, &@2), &@2); scope = NEW_NODE(NODE_SCOPE, tbl, $5, args, &@$); $$ = NEW_FOR($4, scope, &@$); fixpos($$, $2); /*% %*/ /*% ripper: for!($2, $4, $5) %*/ } | k_class cpath superclass { if (p->in_def) { YYLTYPE loc = code_loc_gen(&@1, &@2); yyerror1(&loc, "class definition in method body"); } $1 = p->in_class; p->in_class = 1; local_push(p, 0); } bodystmt k_end { /*%%%*/ $$ = 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->in_class = $1 & 1; } | k_class tLSHFT expr { $$ = (p->in_class << 1) | p->in_def; p->in_def = 0; p->in_class = 0; local_push(p, 0); } term bodystmt k_end { /*%%%*/ $$ = 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->in_def = $4 & 1; p->in_class = ($4 >> 1) & 1; } | k_module cpath { if (p->in_def) { YYLTYPE loc = code_loc_gen(&@1, &@2); yyerror1(&loc, "module definition in method body"); } $1 = p->in_class; p->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($$, @2.end_pos.lineno); /*% %*/ /*% ripper: module!($2, $4) %*/ local_pop(p); p->in_class = $1 & 1; } | k_def fname { local_push(p, 0); $$ = p->cur_arg; p->cur_arg = 0; } { $$ = p->in_def; p->in_def = 1; } f_arglist bodystmt k_end { /*%%%*/ NODE *body = remove_begin($6); reduce_nodes(p, &body); $$ = NEW_DEFN($2, $5, body, &@$); nd_set_line($$->nd_defn, @7.end_pos.lineno); set_line_body(body, @1.beg_pos.lineno); /*% %*/ /*% ripper: def!($2, $5, $6) %*/ local_pop(p); p->in_def = $4 & 1; p->cur_arg = $3; } | k_def singleton dot_or_colon {SET_LEX_STATE(EXPR_FNAME);} fname { $4 = p->in_def; p->in_def = 1; SET_LEX_STATE(EXPR_ENDFN|EXPR_LABEL); /* force for args */ local_push(p, 0); $$ = p->cur_arg; p->cur_arg = 0; } f_arglist bodystmt k_end { /*%%%*/ NODE *body = remove_begin($8); reduce_nodes(p, &body); $$ = NEW_DEFS($2, $5, $7, body, &@$); nd_set_line($$->nd_defn, @9.end_pos.lineno); set_line_body(body, @1.beg_pos.lineno); /*% %*/ /*% ripper: defs!($2, $3, $5, $7, $8) %*/ local_pop(p); p->in_def = $4 & 1; p->cur_arg = $6; } | keyword_break { /*%%%*/ $$ = NEW_BREAK(0, &@$); /*% %*/ /*% ripper: break!(args_new!) %*/ } | keyword_next { /*%%%*/ $$ = NEW_NEXT(0, &@$); /*% %*/ /*% ripper: next!(args_new!) %*/ } | keyword_redo { /*%%%*/ $$ = NEW_REDO(&@$); /*% %*/ /*% ripper: redo! %*/ } | keyword_retry { /*%%%*/ $$ = NEW_RETRY(&@$); /*% %*/ /*% ripper: retry! %*/ } ; primary_value : primary { value_expr($1); $$ = $1; } ; k_begin : keyword_begin { token_info_push(p, "begin", &@$); } ; k_if : keyword_if { token_info_push(p, "if", &@$); } ; k_unless : keyword_unless { token_info_push(p, "unless", &@$); } ; k_while : keyword_while { token_info_push(p, "while", &@$); } ; k_until : keyword_until { token_info_push(p, "until", &@$); } ; k_case : keyword_case { token_info_push(p, "case", &@$); } ; k_for : keyword_for { token_info_push(p, "for", &@$); } ; k_class : keyword_class { token_info_push(p, "class", &@$); } ; k_module : keyword_module { token_info_push(p, "module", &@$); } ; k_def : keyword_def { token_info_push(p, "def", &@$); } ; k_do : keyword_do { token_info_push(p, "do", &@$); } ; k_do_block : keyword_do_block { token_info_push(p, "do", &@$); } ; k_rescue : keyword_rescue { token_info_warn(p, "rescue", p->token_info, 1, &@$); } ; k_ensure : keyword_ensure { token_info_warn(p, "ensure", p->token_info, 1, &@$); } ; k_when : keyword_when { token_info_warn(p, "when", p->token_info, 0, &@$); } ; k_else : keyword_else { token_info *ptinfo_beg = p->token_info; int same = ptinfo_beg && strcmp(ptinfo_beg->token, "case") != 0; token_info_warn(p, "else", p->token_info, same, &@$); } ; k_elsif : keyword_elsif { token_info_warn(p, "elsif", p->token_info, 1, &@$); } ; k_end : keyword_end { token_info_pop(p, "end", &@$); } ; k_return : keyword_return { if (p->in_class && !p->in_def && !dyna_in_block(p)) yyerror1(&@1, "Invalid return in class/module body"); } ; then : term | keyword_then | term keyword_then ; do : term | keyword_do_cond ; if_tail : opt_else | k_elsif expr_value then compstmt if_tail { /*%%%*/ $$ = new_if(p, $2, $4, $5, &@$); fixpos($$, $2); /*% %*/ /*% ripper: elsif!($2, $4, escape_Qundef($5)) %*/ } ; opt_else : none | k_else compstmt { /*%%%*/ $$ = $2; /*% %*/ /*% ripper: else!($2) %*/ } ; for_var : lhs | mlhs ; f_marg : f_norm_arg { /*%%%*/ $$ = assignable(p, $1, 0, &@$); /*% %*/ /*% ripper: assignable(p, $1) %*/ } | tLPAREN f_margs rparen { /*%%%*/ $$ = $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 { /*%%%*/ $$ = NEW_MASGN($1, 0, &@$); /*% %*/ /*% ripper: $1 %*/ } | f_marg_list ',' tSTAR f_norm_arg { /*%%%*/ $$ = NEW_MASGN($1, assignable(p, $4, 0, &@$), &@$); /*% %*/ /*% ripper: mlhs_add_star!($1, assignable(p, $4)) %*/ } | f_marg_list ',' tSTAR f_norm_arg ',' f_marg_list { /*%%%*/ $$ = NEW_MASGN($1, NEW_POSTARG(assignable(p, $4, 0, &@$), $6, &@$), &@$); /*% %*/ /*% ripper: mlhs_add_post!(mlhs_add_star!($1, assignable(p, $4)), $6) %*/ } | f_marg_list ',' tSTAR { /*%%%*/ $$ = NEW_MASGN($1, NODE_SPECIAL_NO_NAME_REST, &@$); /*% %*/ /*% ripper: mlhs_add_star!($1, Qnil) %*/ } | f_marg_list ',' tSTAR ',' f_marg_list { /*%%%*/ $$ = NEW_MASGN($1, NEW_POSTARG(NODE_SPECIAL_NO_NAME_REST, $5, &@$), &@$); /*% %*/ /*% ripper: mlhs_add_post!(mlhs_add_star!($1, Qnil), $5) %*/ } | tSTAR f_norm_arg { /*%%%*/ $$ = NEW_MASGN(0, assignable(p, $2, 0, &@$), &@$); /*% %*/ /*% ripper: mlhs_add_star!(mlhs_new!, assignable(p, $2)) %*/ } | tSTAR f_norm_arg ',' f_marg_list { /*%%%*/ $$ = NEW_MASGN(0, NEW_POSTARG(assignable(p, $2, 0, &@$), $4, &@$), &@$); /*% %*/ /*% ripper: mlhs_add_post!(mlhs_add_star!(mlhs_new!, assignable(p, $2)), $4) %*/ } | tSTAR { /*%%%*/ $$ = NEW_MASGN(0, NODE_SPECIAL_NO_NAME_REST, &@$); /*% %*/ /*% ripper: mlhs_add_star!(mlhs_new!, Qnil) %*/ } | tSTAR ',' f_marg_list { /*%%%*/ $$ = NEW_MASGN(0, NEW_POSTARG(NODE_SPECIAL_NO_NAME_REST, $3, &@$), &@$); /*% %*/ /*% ripper: mlhs_add_post!(mlhs_add_star!(mlhs_new!, Qnil), $3) %*/ } ; block_args_tail : f_block_kwarg ',' f_kwrest opt_f_block_arg { $$ = new_args_tail(p, $1, $3, $4, &@3); } | f_block_kwarg opt_f_block_arg { $$ = new_args_tail(p, $1, Qnone, $2, &@1); } | f_kwrest opt_f_block_arg { $$ = new_args_tail(p, Qnone, $1, $2, &@1); } | f_block_arg { $$ = new_args_tail(p, Qnone, Qnone, $1, &@1); } ; opt_block_args_tail : ',' block_args_tail { $$ = $2; } | /* none */ { $$ = new_args_tail(p, Qnone, Qnone, Qnone, &@0); } ; block_param : f_arg ',' f_block_optarg ',' f_rest_arg opt_block_args_tail { $$ = 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 ',' { /*%%%*/ /* magic number for rest_id in iseq_set_arguments() */ const ID excessed_comma = 1; $$ = new_args(p, $1, Qnone, excessed_comma, Qnone, new_args_tail(p, Qnone, Qnone, Qnone, &@1), &@$); /*% %*/ /*% ripper: new_args(p, $1, Qnone, excessed_comma!, Qnone, new_args_tail(p, Qnone, Qnone, Qnone, NULL), NULL) %*/ } | f_arg ',' f_rest_arg ',' f_arg opt_block_args_tail { $$ = 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; /*%%%*/ $$ = 0; /*% %*/ /*% ripper: block_var!(params_new(Qnil,Qnil,Qnil,Qnil,Qnil,Qnil,Qnil), escape_Qundef($2)) %*/ } | tOROP { /*%%%*/ $$ = 0; /*% %*/ /*% ripper: block_var!(params_new(Qnil,Qnil,Qnil,Qnil,Qnil,Qnil,Qnil), Qnil) %*/ } | '|' block_param opt_bv_decl '|' { p->cur_arg = 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 : { $$ = dyna_push(p); } { $$ = p->lex.lpar_beg; p->lex.lpar_beg = p->lex.paren_nest; } f_larglist { CMDARG_PUSH(0); } lambda_body { p->lex.lpar_beg = $2; CMDARG_POP(); /*%%%*/ { YYLTYPE loc = code_loc_gen(&@3, &@5); $$ = NEW_LAMBDA($3, $5, &loc); nd_set_line($$->nd_body, @5.end_pos.lineno); nd_set_line($$, @3.end_pos.lineno); } /*% %*/ /*% ripper: lambda!($3, $5) %*/ dyna_pop(p, $1); } ; f_larglist : '(' f_args opt_bv_decl ')' { /*%%%*/ $$ = $2; /*% %*/ /*% ripper: paren!($2) %*/ } | f_args { $$ = $1; } ; lambda_body : tLAMBEG compstmt '}' { token_info_pop(p, "}", &@3); $$ = $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 (nd_type($1) == NODE_YIELD) { compile_error(p, "block given to yield"); } 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 { /*%%%*/ $$ = new_qcall(p, $2, $1, $3, $4, &@3, &@$); /*% %*/ /*% ripper: opt_event(:method_add_arg!, call!($1, $2, $3), $4) %*/ } | block_call call_op2 operation2 opt_paren_args brace_block { /*%%%*/ $$ = 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 call_op2 operation2 command_args do_block { /*%%%*/ $$ = new_command_qcall(p, $2, $1, $3, $4, $5, &@3, &@$); /*% %*/ /*% ripper: method_add_block!(command_call!($1, $2, $3, $4), $5) %*/ } ; method_call : fcall paren_args { /*%%%*/ $$ = $1; $$->nd_args = $2; nd_set_last_loc($1, @2.end_pos); /*% %*/ /*% ripper: method_add_arg!(fcall!($1), $2) %*/ } | primary_value call_op operation2 opt_paren_args { /*%%%*/ $$ = 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_qcall(p, ID2VAL(idCOLON2), $1, $3, $4, &@3, &@$); nd_set_line($$, @3.end_pos.lineno); /*% %*/ /*% ripper: method_add_arg!(call!($1, ID2VAL(idCOLON2), $3), $4) %*/ } | primary_value tCOLON2 operation3 { /*%%%*/ $$ = new_qcall(p, ID2VAL(idCOLON2), $1, $3, Qnull, &@3, &@$); /*% %*/ /*% ripper: call!($1, ID2VAL(idCOLON2), $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) %*/ } | primary_value tCOLON2 paren_args { /*%%%*/ $$ = new_qcall(p, ID2VAL(idCOLON2), $1, ID2VAL(idCall), $3, &@2, &@$); nd_set_line($$, @2.end_pos.lineno); /*% %*/ /*% ripper: method_add_arg!(call!($1, ID2VAL(idCOLON2), ID2VAL(idCall)), $3) %*/ } | keyword_super paren_args { /*%%%*/ $$ = NEW_SUPER($2, &@$); /*% %*/ /*% ripper: super!($2) %*/ } | keyword_super { /*%%%*/ $$ = NEW_ZSUPER(&@$); /*% %*/ /*% ripper: zsuper! %*/ } | primary_value '[' opt_call_args rbracket { /*%%%*/ if ($1 && nd_type($1) == NODE_SELF) $$ = NEW_FCALL(tAREF, $3, &@$); else $$ = NEW_CALL($1, tAREF, $3, &@$); fixpos($$, $1); /*% %*/ /*% ripper: aref!($1, escape_Qundef($3)) %*/ } ; brace_block : '{' brace_body '}' { $$ = $2; /*%%%*/ $$->nd_body->nd_loc = code_loc_gen(&@1, &@3); nd_set_line($$, @1.end_pos.lineno); /*% %*/ } | k_do do_body k_end { $$ = $2; /*%%%*/ $$->nd_body->nd_loc = code_loc_gen(&@1, &@3); nd_set_line($$, @1.end_pos.lineno); /*% %*/ } ; brace_body : {$$ = dyna_push(p);} opt_block_param compstmt { /*%%%*/ $$ = NEW_ITER($2, $3, &@$); /*% %*/ /*% ripper: brace_block!(escape_Qundef($2), $3) %*/ dyna_pop(p, $1); } ; do_body : {$$ = dyna_push(p);} {CMDARG_PUSH(0);} opt_block_param bodystmt { /*%%%*/ $$ = NEW_ITER($3, $4, &@$); /*% %*/ /*% ripper: do_block!(escape_Qundef($3), $4) %*/ CMDARG_POP(); dyna_pop(p, $1); } ; case_body : k_when args then compstmt cases { /*%%%*/ $$ = NEW_WHEN($2, $4, $5, &@$); fixpos($$, $2); /*% %*/ /*% ripper: when!($2, $4, escape_Qundef($5)) %*/ } ; cases : opt_else | case_body ; opt_rescue : k_rescue exc_list exc_var then compstmt opt_rescue { /*%%%*/ $$ = NEW_RESBODY($2, $3 ? block_append(p, node_assign(p, $3, NEW_ERRINFO(&@3), &@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($1, &@$); /*% %*/ /*% ripper: rb_ary_new3(1, get_value($1)) %*/ } | mrhs { /*%%%*/ if (!($$ = splat_array($1))) $$ = $1; /*% %*/ /*% ripper: $1 %*/ } | none ; exc_var : tASSOC lhs { $$ = $2; } | none ; opt_ensure : k_ensure compstmt { /*%%%*/ $$ = $2; /*% %*/ /*% ripper: ensure!($2) %*/ } | none ; literal : numeric | symbol { /*%%%*/ $$ = NEW_LIT(ID2SYM($1), &@$); /*% %*/ /*% ripper: symbol_literal!($1) %*/ } | dsym ; strings : string { /*%%%*/ NODE *node = $1; if (!node) { node = NEW_STR(add_mark_object(p, STR_NEW0()), &@$); } else { node = evstr2dstr(p, node); } $$ = node; /*% %*/ /*% ripper: $1 %*/ } ; string : tCHAR | string1 | string string1 { /*%%%*/ $$ = literal_concat(p, $1, $2, &@$); /*% %*/ /*% ripper: string_concat!($1, $2) %*/ } ; string1 : tSTRING_BEG string_contents tSTRING_END { /*%%%*/ $$ = heredoc_dedent(p, $2); if ($$) nd_set_loc($$, &@$); /*% %*/ /*% ripper: string_literal!(heredoc_dedent(p, $2)) %*/ } ; xstring : tXSTRING_BEG xstring_contents tSTRING_END { /*%%%*/ $$ = new_xstring(p, heredoc_dedent(p, $2), &@$); /*% %*/ /*% ripper: xstring_literal!(heredoc_dedent(p, $2)) %*/ } ; regexp : tREGEXP_BEG regexp_contents tREGEXP_END { $$ = new_regexp(p, $2, $3, &@$); } ; words : tWORDS_BEG ' ' word_list tSTRING_END { /*%%%*/ $$ = make_array($3, &@$); /*% %*/ /*% ripper: array!($3) %*/ } ; word_list : /* none */ { /*%%%*/ $$ = 0; /*% %*/ /*% ripper: words_new! %*/ } | word_list word ' ' { /*%%%*/ $$ = list_append(p, $1, evstr2dstr(p, $2)); /*% %*/ /*% ripper: words_add!($1, $2) %*/ } ; word : string_content /*% ripper[brace]: word_add!(word_new!, $1) %*/ | word string_content { /*%%%*/ $$ = literal_concat(p, $1, $2, &@$); /*% %*/ /*% ripper: word_add!($1, $2) %*/ } ; symbols : tSYMBOLS_BEG ' ' symbol_list tSTRING_END { /*%%%*/ $$ = make_array($3, &@$); /*% %*/ /*% ripper: array!($3) %*/ } ; symbol_list : /* none */ { /*%%%*/ $$ = 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_array($3, &@$); /*% %*/ /*% ripper: array!($3) %*/ } ; qsymbols : tQSYMBOLS_BEG ' ' qsym_list tSTRING_END { /*%%%*/ $$ = make_array($3, &@$); /*% %*/ /*% ripper: array!($3) %*/ } ; qword_list : /* none */ { /*%%%*/ $$ = 0; /*% %*/ /*% ripper: qwords_new! %*/ } | qword_list tSTRING_CONTENT ' ' { /*%%%*/ $$ = list_append(p, $1, $2); /*% %*/ /*% ripper: qwords_add!($1, $2) %*/ } ; qsym_list : /* none */ { /*%%%*/ $$ = 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! %*/ } | string_contents string_content { /*%%%*/ $$ = literal_concat(p, $1, $2, &@$); /*% %*/ /*% ripper: string_add!($1, $2) %*/ } ; xstring_contents: /* none */ { /*%%%*/ $$ = 0; /*% %*/ /*% ripper: xstring_new! %*/ } | xstring_contents string_content { /*%%%*/ $$ = literal_concat(p, $1, $2, &@$); /*% %*/ /*% ripper: xstring_add!($1, $2) %*/ } ; regexp_contents: /* none */ { /*%%%*/ $$ = 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 | tSTRING_DVAR { /* need to backup p->lex.strterm so that a string literal `%&foo,#$&,bar&` can be parsed */ $$ = p->lex.strterm; p->lex.strterm = 0; SET_LEX_STATE(EXPR_BEG); } string_dvar { p->lex.strterm = $2; /*%%%*/ $$ = NEW_EVSTR($3, &@$); nd_set_line($$, @3.end_pos.lineno); /*% %*/ /*% ripper: string_dvar!($3) %*/ } | tSTRING_DBEG { CMDARG_PUSH(0); COND_PUSH(0); } { /* need to backup p->lex.strterm so that a string literal `%!foo,#{ !0 },bar!` can be parsed */ $$ = p->lex.strterm; p->lex.strterm = 0; } { $$ = p->lex.state; SET_LEX_STATE(EXPR_BEG); } { $$ = p->lex.brace_nest; p->lex.brace_nest = 0; } { $$ = p->heredoc_indent; p->heredoc_indent = 0; } compstmt tSTRING_DEND { COND_POP(); CMDARG_POP(); p->lex.strterm = $3; SET_LEX_STATE($4); p->lex.brace_nest = $5; p->heredoc_indent = $6; p->heredoc_line_indent = -1; /*%%%*/ if ($7) $7->flags &= ~NODE_FL_NEWLINE; $$ = new_evstr(p, $7, &@$); /*% %*/ /*% ripper: string_embexpr!($7) %*/ } ; string_dvar : tGVAR { /*%%%*/ $$ = NEW_GVAR($1, &@$); /*% %*/ /*% ripper: var_ref!($1) %*/ } | tIVAR { /*%%%*/ $$ = NEW_IVAR($1, &@$); /*% %*/ /*% ripper: var_ref!($1) %*/ } | tCVAR { /*%%%*/ $$ = NEW_CVAR($1, &@$); /*% %*/ /*% ripper: var_ref!($1) %*/ } | backref ; symbol : tSYMBEG sym { SET_LEX_STATE(EXPR_END); /*%%%*/ $$ = $2; /*% %*/ /*% ripper: symbol!($2) %*/ } ; sym : fname | tIVAR | tGVAR | tCVAR ; dsym : tSYMBEG xstring_contents tSTRING_END { SET_LEX_STATE(EXPR_END); /*%%%*/ $$ = dsym_node(p, $2, &@$); /*% %*/ /*% ripper: dyna_symbol!($2) %*/ } ; numeric : simple_numeric | tUMINUS_NUM simple_numeric %prec tLOWEST { /*%%%*/ $$ = $2; add_mark_object(p, $$->nd_lit = negate_lit(p, $$->nd_lit)); /*% %*/ /*% ripper: unary!(ID2VAL(idUMinus), $2) %*/ } ; simple_numeric : tINTEGER | tFLOAT | tRATIONAL | tIMAGINARY ; user_variable : tIDENTIFIER | tIVAR | tGVAR | tCONSTANT | tCVAR ; 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 { $$ = dispatch1(vcall, $1); } %*/ } | keyword_variable { /*%%%*/ if (!($$ = gettable(p, $1, &@$))) $$ = NEW_BEGIN(0, &@$); /*% %*/ /*% ripper: var_ref!($1) %*/ } ; var_lhs : user_variable { /*%%%*/ $$ = assignable(p, $1, 0, &@$); /*% %*/ /*% ripper: assignable(p, var_field(p, $1)) %*/ } | keyword_variable { /*%%%*/ $$ = assignable(p, $1, 0, &@$); /*% %*/ /*% ripper: assignable(p, var_field(p, $1)) %*/ } ; backref : tNTH_REF | tBACK_REF ; superclass : '<' { SET_LEX_STATE(EXPR_BEG); p->command_start = TRUE; } expr_value term { $$ = $3; } | /* none */ { /*%%%*/ $$ = 0; /*% %*/ /*% ripper: Qnil %*/ } ; f_arglist : '(' f_args rparen { /*%%%*/ $$ = $2; /*% %*/ /*% ripper: paren!($2) %*/ SET_LEX_STATE(EXPR_BEG); p->command_start = TRUE; } | { $$ = p->in_kwarg; p->in_kwarg = 1; SET_LEX_STATE(p->lex.state|EXPR_LABEL); /* force for args */ } f_args term { p->in_kwarg = !!$1; $$ = $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_kwrest opt_f_block_arg { $$ = new_args_tail(p, Qnone, $1, $2, &@1); } | f_block_arg { $$ = new_args_tail(p, Qnone, Qnone, $1, &@1); } ; opt_args_tail : ',' args_tail { $$ = $2; } | /* none */ { $$ = new_args_tail(p, Qnone, Qnone, Qnone, &@0); } ; f_args : f_arg ',' f_optarg ',' f_rest_arg opt_args_tail { $$ = new_args(p, $1, $3, $5, Qnone, $6, &@$); } | f_arg ',' f_optarg ',' f_rest_arg ',' f_arg opt_args_tail { $$ = new_args(p, $1, $3, $5, $7, $8, &@$); } | f_arg ',' f_optarg opt_args_tail { $$ = new_args(p, $1, $3, Qnone, Qnone, $4, &@$); } | f_arg ',' f_optarg ',' f_arg opt_args_tail { $$ = new_args(p, $1, $3, Qnone, $5, $6, &@$); } | f_arg ',' f_rest_arg opt_args_tail { $$ = new_args(p, $1, Qnone, $3, Qnone, $4, &@$); } | f_arg ',' f_rest_arg ',' f_arg opt_args_tail { $$ = new_args(p, $1, Qnone, $3, $5, $6, &@$); } | f_arg opt_args_tail { $$ = new_args(p, $1, Qnone, Qnone, Qnone, $2, &@$); } | f_optarg ',' f_rest_arg opt_args_tail { $$ = 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 { $$ = new_args(p, Qnone, $1, Qnone, $3, $4, &@$); } | f_rest_arg opt_args_tail { $$ = 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 */ { $$ = new_args_tail(p, Qnone, Qnone, Qnone, &@0); $$ = new_args(p, Qnone, Qnone, Qnone, Qnone, $$, &@0); } ; f_bad_arg : tCONSTANT { /*%%%*/ yyerror1(&@1, "formal argument cannot be a constant"); $$ = 0; /*% %*/ /*% ripper[error]: param_error!($1) %*/ } | tIVAR { /*%%%*/ yyerror1(&@1, "formal argument cannot be an instance variable"); $$ = 0; /*% %*/ /*% ripper[error]: param_error!($1) %*/ } | tGVAR { /*%%%*/ yyerror1(&@1, "formal argument cannot be a global variable"); $$ = 0; /*% %*/ /*% ripper[error]: param_error!($1) %*/ } | tCVAR { /*%%%*/ yyerror1(&@1, "formal argument cannot be a class variable"); $$ = 0; /*% %*/ /*% ripper[error]: param_error!($1) %*/ } ; f_norm_arg : f_bad_arg | tIDENTIFIER { formal_argument(p, get_id($1)); $$ = $1; } ; f_arg_asgn : f_norm_arg { ID id = get_id($1); arg_var(p, id); p->cur_arg = id; $$ = $1; } ; f_arg_item : f_arg_asgn { p->cur_arg = 0; /*%%%*/ $$ = 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 { ID id = get_id($1); arg_var(p, formal_argument(p, id)); p->cur_arg = id; $$ = $1; } ; f_kw : f_label arg_value { p->cur_arg = 0; /*%%%*/ $$ = 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; /*%%%*/ $$ = new_kw_arg(p, assignable(p, $1, NODE_SPECIAL_REQUIRED_KEYWORD, &@$), &@$); /*% %*/ /*% ripper: rb_assoc_new(get_value(assignable(p, $1)), 0) %*/ } ; f_block_kw : f_label primary_value { /*%%%*/ $$ = new_kw_arg(p, assignable(p, $1, $2, &@$), &@$); /*% %*/ /*% ripper: rb_assoc_new(get_value(assignable(p, $1)), get_value($2)) %*/ } | f_label { /*%%%*/ $$ = new_kw_arg(p, assignable(p, $1, NODE_SPECIAL_REQUIRED_KEYWORD, &@$), &@$); /*% %*/ /*% ripper: rb_assoc_new(get_value(assignable(p, $1)), 0) %*/ } ; f_block_kwarg : f_block_kw { /*%%%*/ $$ = $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 { /*%%%*/ $$ = kwd_append($1, $3); /*% %*/ /*% ripper: rb_ary_push($1, get_value($3)) %*/ } ; kwrest_mark : tPOW | tDSTAR ; f_kwrest : kwrest_mark tIDENTIFIER { arg_var(p, shadowing_lvar(p, get_id($2))); /*%%%*/ $$ = $2; /*% %*/ /*% ripper: kwrest_param!($2) %*/ } | kwrest_mark { /*%%%*/ $$ = internal_id(p); arg_var(p, $$); /*% %*/ /*% ripper: kwrest_param!(Qnil) %*/ } ; f_opt : f_arg_asgn '=' arg_value { p->cur_arg = 0; /*%%%*/ $$ = NEW_OPT_ARG(0, assignable(p, $1, $3, &@$), &@$); /*% %*/ /*% ripper: rb_assoc_new(get_value(assignable(p, $1)), get_value($3)) %*/ } ; f_block_opt : f_arg_asgn '=' primary_value { p->cur_arg = 0; /*%%%*/ $$ = 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 { /*%%%*/ $$ = $1; /*% %*/ /*% ripper: rb_ary_new3(1, get_value($1)) %*/ } | f_optarg ',' f_opt { /*%%%*/ $$ = opt_arg_append($1, $3); /*% %*/ /*% ripper: rb_ary_push($1, get_value($3)) %*/ } ; restarg_mark : '*' | tSTAR ; f_rest_arg : restarg_mark tIDENTIFIER { arg_var(p, shadowing_lvar(p, get_id($2))); /*%%%*/ $$ = $2; /*% %*/ /*% ripper: rest_param!($2) %*/ } | restarg_mark { /*%%%*/ $$ = internal_id(p); arg_var(p, $$); /*% %*/ /*% ripper: rest_param!(Qnil) %*/ } ; blkarg_mark : '&' | tAMPER ; f_block_arg : blkarg_mark tIDENTIFIER { arg_var(p, shadowing_lvar(p, get_id($2))); /*%%%*/ $$ = $2; /*% %*/ /*% ripper: blockarg!($2) %*/ } ; opt_f_block_arg : ',' f_block_arg { $$ = $2; } | none { /*%%%*/ $$ = 0; /*% %*/ /*% ripper: Qundef %*/ } ; singleton : var_ref { value_expr($1); $$ = $1; } | '(' {SET_LEX_STATE(EXPR_BEG);} expr rparen { /*%%%*/ switch (nd_type($3)) { case NODE_STR: case NODE_DSTR: case NODE_XSTR: case NODE_DXSTR: case NODE_DREGX: case NODE_LIT: case NODE_ARRAY: case NODE_ZARRAY: yyerror1(&@3, "can't define singleton method for literals"); break; default: value_expr($3); break; } $$ = $3; /*% %*/ /*% ripper: paren!($3) %*/ } ; assoc_list : none | assocs trailer { /*%%%*/ $$ = $1; /*% %*/ /*% ripper: assoclist_from_args!($1) %*/ } ; assocs : assoc /*% ripper[brace]: rb_ary_new3(1, get_value($1)) %*/ | assocs ',' assoc { /*%%%*/ NODE *assocs = $1; NODE *tail = $3; if (!assocs) { assocs = tail; } else if (tail) { if (assocs->nd_head && !tail->nd_head && nd_type(tail->nd_next) == NODE_ARRAY && nd_type(tail->nd_next->nd_head) == NODE_HASH) { /* DSTAR */ tail = tail->nd_next->nd_head->nd_head; } assocs = list_concat(assocs, tail); } $$ = assocs; /*% %*/ /*% ripper: rb_ary_push($1, get_value($3)) %*/ } ; assoc : arg_value tASSOC arg_value { /*%%%*/ if (nd_type($1) == NODE_STR) { nd_set_type($1, NODE_LIT); add_mark_object(p, $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) %*/ } | tSTRING_BEG string_contents tLABEL_END arg_value { /*%%%*/ YYLTYPE loc = code_loc_gen(&@1, &@3); $$ = list_append(p, NEW_LIST(dsym_node(p, $2, &loc), &loc), $4); /*% %*/ /*% ripper: assoc_new!(dyna_symbol!($2), $4) %*/ } | tDSTAR arg_value { /*%%%*/ if (nd_type($2) == NODE_HASH && !($2->nd_head && $2->nd_head->nd_alen)) $$ = 0; else $$ = list_append(p, NEW_LIST(0, &@$), $2); /*% %*/ /*% ripper: assoc_splat!($2) %*/ } ; operation : tIDENTIFIER | tCONSTANT | tFID ; operation2 : tIDENTIFIER | tCONSTANT | tFID | op ; operation3 : tIDENTIFIER | tFID | op ; dot_or_colon : '.' | tCOLON2 ; call_op : '.' | tANDDOT ; call_op2 : call_op | tCOLON2 ; opt_terms : /* none */ | terms ; opt_nl : /* none */ | '\n' ; rparen : opt_nl ')' ; rbracket : opt_nl ']' ; trailer : /* none */ | '\n' | ',' ; term : ';' {yyerrok;token_flush(p);} | '\n' {token_flush(p);} ; terms : term | terms ';' {yyerrok;} ; none : /* none */ { $$ = Qnull; } ; %% # 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**); 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) set_yylval_node(NEW_STR(x, &_cur_loc)) # define set_yylval_literal(x) set_yylval_node(NEW_LIT(x, &_cur_loc)) # 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 = (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) (void)(x) # define set_yylval_node(x) (void)(x) # define yylval_id() yylval.id # define _cur_loc NULL_LOC /* dummy */ #endif #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, int 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, int 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, int t) { int saved_line = p->ruby_sourceline; const char *saved_tokp = p->lex.ptok; 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)); p->delayed = 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)) #endif /* RIPPER */ #include "ruby/regex.h" #include "ruby/util.h" 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 setup_token_info(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->linenum = loc->beg_pos.lineno; ptinfo->column = 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; setup_token_info(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_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; setup_token_info(ptinfo_end, p->lex.pbeg, loc); if (ptinfo_beg->linenum == ptinfo_end->linenum) return; /* ignore one-line block */ if (ptinfo_beg->nonspc || ptinfo_end->nonspc) return; /* ignore keyword in the middle of a line */ if (ptinfo_beg->column == ptinfo_end->column) return; /* the indents are matched */ if (!same && ptinfo_beg->column < ptinfo_end->column) return; rb_warn3L(ptinfo_end->linenum, "mismatched indentations at '%s' with '%s' at %d", WARN_S(token), WARN_S(ptinfo_beg->token), WARN_I(ptinfo_beg->linenum)); } 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 int parser_yyerror(struct parser_params *p, const YYLTYPE *yylloc, const char *msg) { #ifndef RIPPER const int max_line_margin = 30; const char *ptr, *ptr_end, *pt, *pb; const char *pre = "", *post = "", *pend; const char *code = "", *caret = "", *newline = ""; const char *lim; char *buf; long len; int i; YYLTYPE current; if (!yylloc) { RUBY_SET_YYLLOC(current); yylloc = ¤t; } else if ((p->ruby_sourceline != yylloc->beg_pos.lineno && p->ruby_sourceline != yylloc->end_pos.lineno) || (yylloc->beg_pos.lineno == yylloc->end_pos.lineno && yylloc->beg_pos.column == yylloc->end_pos.column)) { compile_error(p, "%s", msg); return 0; } pend = p->lex.pend; if (pend > p->lex.pbeg && pend[-1] == '\n') { if (--pend > p->lex.pbeg && pend[-1] == '\r') --pend; } pt = (p->ruby_sourceline == yylloc->end_pos.lineno) ? p->lex.pbeg + yylloc->end_pos.column : p->lex.pend; ptr = ptr_end = pt < pend ? pt : pend; lim = ptr - p->lex.pbeg > max_line_margin ? ptr - max_line_margin : p->lex.pbeg; while ((lim < ptr) && (*(ptr-1) != '\n')) ptr--; lim = pend - ptr_end > max_line_margin ? ptr_end + max_line_margin : pend; while ((ptr_end < lim) && (*ptr_end != '\n')) ptr_end++; len = ptr_end - ptr; if (len > 4) { if (ptr > p->lex.pbeg) { ptr = rb_enc_prev_char(p->lex.pbeg, ptr, pt, rb_enc_get(p->lex.lastline)); if (ptr > p->lex.pbeg) pre = "..."; } if (ptr_end < pend) { ptr_end = rb_enc_prev_char(pt, ptr_end, pend, rb_enc_get(p->lex.lastline)); if (ptr_end < pend) post = "..."; } } pb = p->lex.pbeg; if (p->ruby_sourceline == 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) { compile_error(p, "%s", msg); } else if (!p->error_buffer && rb_stderr_tty_p()) { #define CSI_BEGIN "\033[" #define CSI_SGR "m" compile_error(p, "%s\n" 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 */, msg, 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++; } if (lim > ptr) { memset(p2, '~', (lim - ptr)); p2 += (lim - ptr); } *p2 = '\0'; newline = "\n"; compile_error(p, "%s%s""%s%.*s%s%s""%s%s", msg, newline, pre, (int)len, code, post, newline, pre, caret); } #else dispatch1(parse_error, STR_NEW2(msg)); ripper_error(p); #endif /* !RIPPER */ return 0; } static int vtable_size(const struct vtable *tbl) { if (!DVARS_TERMINAL_P(tbl)) { return tbl->pos; } else { return 0; } } 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 void vtable_free_gen(struct parser_params *p, int line, const char *name, struct vtable *tbl) { #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 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) #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) { int i; if (!DVARS_TERMINAL_P(tbl)) { for (i = 0; i < tbl->pos; i++) { if (tbl->tbl[i] == id) { return i+1; } } } 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; } } return 0; } static int e_option_supplied(struct parser_params *p) { return strcmp(p->ruby_sourcefile, "-e") == 0; } static VALUE yycompile0(VALUE arg) { int n; NODE *tree; struct parser_params *p = (struct parser_params *)arg; VALUE cov = Qfalse; if (!compile_for_eval && rb_safe_level() == 0 && !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 = STR_NEW0(); n = p->ruby_sourceline; do { rb_ary_push(p->debug_lines, str); } while (--n); } if (!e_option_supplied(p)) { cov = Qtrue; } } 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); add_mark_object(p, opt); tree->nd_body = prelude; p->ast->body.compile_option = opt; } p->ast->body.root = tree; p->ast->body.line_count = 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_str_new_frozen(fname); p->ruby_sourcefile = StringValueCStr(fname); } p->ruby_sourceline = line - 1; p->ast = ast = rb_ast_new(); rb_suppress_tracing(yycompile0, (VALUE)p); p->ast = 0; RB_GC_GUARD(vparser); /* prohibit tail call optimization */ return ast; } #endif /* !RIPPER */ static rb_encoding * must_be_ascii_compatible(VALUE s) { 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); #ifndef RIPPER if (p->debug_lines) { rb_enc_associate(line, p->enc); rb_ary_push(p->debug_lines, line); } #endif p->line_count++; return 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_compile_string(const char *f, VALUE s, int line) { must_be_ascii_compatible(s); return parser_compile_string(rb_parser_new(), rb_filesystem_str_new_cstr(f), s, line); } rb_ast_t* rb_parser_compile_string(VALUE vparser, const char *f, VALUE s, int line) { return rb_parser_compile_string_path(vparser, rb_filesystem_str_new_cstr(f), s, 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); } rb_ast_t* rb_compile_cstr(const char *f, const char *s, int len, int line) { VALUE str = rb_str_new(s, len); return parser_compile_string(rb_parser_new(), rb_filesystem_str_new_cstr(f), str, line); } rb_ast_t* rb_parser_compile_cstr(VALUE vparser, const char *f, const char *s, int len, int line) { VALUE str = rb_str_new(s, len); return parser_compile_string(vparser, rb_filesystem_str_new_cstr(f), str, line); } VALUE rb_io_gets_internal(VALUE io); static VALUE lex_io_gets(struct parser_params *p, VALUE io) { return rb_io_gets_internal(io); } rb_ast_t* rb_compile_file(const char *f, VALUE file, int start) { VALUE vparser = rb_parser_new(); return rb_parser_compile_file(vparser, f, file, start); } rb_ast_t* rb_parser_compile_file(VALUE vparser, const char *f, VALUE file, int start) { return rb_parser_compile_file_path(vparser, rb_filesystem_str_new_cstr(f), file, start); } rb_ast_t* rb_parser_compile_file_path(VALUE vparser, VALUE fname, VALUE file, int start) { struct parser_params *p; 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 VALUE lex_generic_gets(struct parser_params *p, VALUE input) { return (*p->lex.gets_.call)(input, p->line_count); } rb_ast_t* rb_parser_compile_generic(VALUE vparser, VALUE (*lex_gets)(VALUE, int), VALUE fname, VALUE input, int start) { struct parser_params *p; 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 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 = rb_str_buf_new(1024); rb_enc_associate(p->delayed, p->enc); p->delayed_line = p->ruby_sourceline; p->delayed_col = (int)(tok - p->lex.pbeg); } rb_str_buf_cat(p->delayed, tok, end - tok); p->lex.ptok = end; } } #else #define add_delayed_token(p, tok, end) ((void)(tok), (void)(end)) #endif static int nextline(struct parser_params *p) { VALUE v = p->lex.nextline; p->lex.nextline = 0; if (!v) { if (p->eofp) return -1; if (!p->lex.input || NIL_P(v = lex_getline(p))) { p->eofp = 1; lex_goto_eol(p); return -1; } p->cr_seen = FALSE; } 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'; } else 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"); } return c; } static inline int nextc(struct parser_params *p) { int c; if (UNLIKELY((p->lex.pcur == p->lex.pend) || p->eofp || p->lex.nextline)) { if (nextline(p)) return -1; } c = (unsigned char)*p->lex.pcur++; if (UNLIKELY(c == '\r')) { c = parser_cr(p, c); } return c; } static void pushback(struct parser_params *p, int c) { if (c == -1) return; 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(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 static char* 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) { p->lex.ptok = p->lex.pcur; yyerror0("invalid hex escape"); return 0; } p->lex.pcur += *numlen; return c; } #define tokcopy(p, n) memcpy(tokspace(p, n), (p)->lex.pcur - (n), (n)) static int tokadd_codepoint(struct parser_params *p, rb_encoding **encp, int regexp_literal, int wide) { 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) { static const char mixed_utf8[] = "UTF-8 mixed within %s source"; size_t len = sizeof(mixed_utf8) - 2 + strlen(rb_enc_name(*encp)); char *mesg = alloca(len); snprintf(mesg, len, mixed_utf8, rb_enc_name(*encp)); yyerror0(mesg); return wide; } *encp = utf8; tokaddmbc(p, codepoint, *encp); } else { tokadd(p, codepoint); } return TRUE; } /* return value is for ?\u3042 */ static int parser_tokadd_utf8(struct parser_params *p, rb_encoding **encp, int string_literal, int symbol_literal, int regexp_literal) { /* * If string_literal is true, then we allow multiple codepoints * in \u{}, and add the codepoints to the current token. * Otherwise we're parsing a character literal and return a single * codepoint without adding it */ const int open_brace = '{', close_brace = '}'; if (regexp_literal) { tokadd(p, '\\'); tokadd(p, 'u'); } if (peek(p, open_brace)) { /* handle \u{...} form */ 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 (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 (c != close_brace) { unterminated: literal_flush(p, p->lex.pcur); yyerror0("unterminated Unicode escape"); return 0; } if (regexp_literal) tokadd(p, close_brace); nextc(p); } else { /* handle \uxxxx form */ if (!tokadd_codepoint(p, encp, regexp_literal, FALSE)) { return 0; } } return TRUE; } #define ESCAPE_CONTROL 1 #define ESCAPE_META 2 static int read_escape(struct parser_params *p, int flags, rb_encoding **encp) { int c; size_t numlen; switch (c = nextc(p)) { case '\\': /* Backslash */ return c; case 'n': /* newline */ return '\n'; case 't': /* horizontal tab */ return '\t'; case 'r': /* carriage-return */ return '\r'; case 'f': /* form-feed */ return '\f'; case 'v': /* vertical tab */ return '\13'; case 'a': /* alarm(bell) */ return '\007'; case 'e': /* escape */ return 033; case '0': case '1': case '2': case '3': /* octal constant */ case '4': case '5': case '6': case '7': pushback(p, c); c = scan_oct(p->lex.pcur, 3, &numlen); p->lex.pcur += numlen; return c; case 'x': /* hex constant */ c = tok_hex(p, &numlen); if (numlen == 0) return 0; return c; case 'b': /* backspace */ return '\010'; case 's': /* space */ return ' '; case 'M': if (flags & ESCAPE_META) goto eof; if ((c = nextc(p)) != '-') { goto eof; } if ((c = nextc(p)) == '\\') { if (peek(p, 'u')) goto eof; return read_escape(p, flags|ESCAPE_META, encp) | 0x80; } else if (c == -1 || !ISASCII(c)) goto eof; else { return ((c & 0xff) | 0x80); } case 'C': if ((c = nextc(p)) != '-') { goto eof; } case 'c': if (flags & ESCAPE_CONTROL) goto eof; if ((c = nextc(p))== '\\') { if (peek(p, 'u')) goto eof; c = read_escape(p, flags|ESCAPE_CONTROL, encp); } else if (c == '?') return 0177; else if (c == -1 || !ISASCII(c)) goto eof; return c & 0x9f; eof: case -1: yyerror0("Invalid escape character syntax"); pushback(p, c); return '\0'; default: return c; } } 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(struct parser_params *p, rb_encoding **encp) { int c; int flags = 0; size_t numlen; first: 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': { 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 */ { tok_hex(p, &numlen); if (numlen == 0) return -1; tokcopy(p, (int)numlen + 2); } return 0; case 'M': if (flags & ESCAPE_META) goto eof; if ((c = nextc(p)) != '-') { pushback(p, c); goto eof; } tokcopy(p, 3); flags |= ESCAPE_META; goto escaped; case 'C': if (flags & ESCAPE_CONTROL) goto eof; if ((c = nextc(p)) != '-') { pushback(p, c); goto eof; } tokcopy(p, 3); goto escaped; case 'c': if (flags & ESCAPE_CONTROL) goto eof; tokcopy(p, 2); flags |= ESCAPE_CONTROL; escaped: if ((c = nextc(p)) == '\\') { goto first; } else if (c == -1) goto eof; tokadd(p, c); return 0; eof: case -1: yyerror0("Invalid escape character syntax"); return -1; default: tokadd(p, '\\'); tokadd(p, c); } return 0; } static int regx_options(struct parser_params *p) { int kcode = 0; int kopt = 0; int options = 0; int c, opt, kc; 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); } } options |= kopt; pushback(p, c); if (toklen(p)) { tokfix(p); compile_error(p, "unknown regexp option%s - %*s", toklen(p) > 1 ? "s" : "", toklen(p), tok(p)); } return options | RE_OPTION_ENCODING(kcode); } 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; } static inline int simple_re_meta(int c) { switch (c) { case '$': case '*': case '+': case '.': case '?': case '^': case '|': case ')': case ']': case '}': case '>': return TRUE; default: return FALSE; } } static int parser_update_heredoc_indent(struct parser_params *p, int c) { if (p->heredoc_line_indent == -1) { if (c == '\n') p->heredoc_line_indent = 0; } else { if (c == ' ') { p->heredoc_line_indent++; return TRUE; } else if (c == '\t') { int w = (p->heredoc_line_indent / TAB_WIDTH) + 1; p->heredoc_line_indent = w * TAB_WIDTH; return TRUE; } else if (c != '\n') { if (p->heredoc_indent > p->heredoc_line_indent) { p->heredoc_indent = p->heredoc_line_indent; } p->heredoc_line_indent = -1; } } return FALSE; } static int tokadd_string(struct parser_params *p, int func, int term, int paren, long *nest, rb_encoding **encp) { int c; rb_encoding *enc = 0; char *errbuf = 0; static const char mixed_msg[] = "%s mixed within %s source"; #define mixed_error(enc1, enc2) if (!errbuf) { \ size_t len = sizeof(mixed_msg) - 4; \ len += strlen(rb_enc_name(enc1)); \ len += strlen(rb_enc_name(enc2)); \ errbuf = ALLOCA_N(char, len); \ snprintf(errbuf, len, mixed_msg, \ rb_enc_name(enc1), \ rb_enc_name(enc2)); \ yyerror0(errbuf); \ } #define mixed_escape(beg, enc1, enc2) do { \ const char *pos = p->lex.pcur; \ p->lex.pcur = (beg); \ mixed_error((enc1), (enc2)); \ p->lex.pcur = pos; \ } while (0) while ((c = nextc(p)) != -1) { if (p->heredoc_indent > 0) { parser_update_heredoc_indent(p, c); } if (paren && c == paren) { ++*nest; } else if (c == term) { if (!nest || !*nest) { pushback(p, c); break; } --*nest; } else if ((func & STR_FUNC_EXPAND) && c == '#' && p->lex.pcur < p->lex.pend) { int c2 = *p->lex.pcur; if (c2 == '$' || c2 == '@' || c2 == '{') { pushback(p, c); break; } } else if (c == '\\') { 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) { 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(p, c); break; case 'u': if ((func & STR_FUNC_EXPAND) == 0) { tokadd(p, '\\'); break; } if (!parser_tokadd_utf8(p, &enc, term, func & STR_FUNC_SYMBOL, func & STR_FUNC_REGEXP)) { return -1; } 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) { 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(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(p, '\\'); pushback(p, c); continue; } } } 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(p, c); break; } 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) \ new_strterm((VALUE)(func), (VALUE)(paren), (VALUE)(term), 0) #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, 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 = (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; VALUE lit; 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(p);} while (ISSPACE(c)); space = 1; } if (func & STR_FUNC_LIST) { quote->u1.func &= ~STR_FUNC_LIST; space = 1; } if (c == term && !quote->u0.nest) { if (func & STR_FUNC_QWORDS) { 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 ' '; } return parser_string_term(p, func); } if (space) { pushback(p, c); add_delayed_token(p, p->lex.ptok, p->lex.pcur); return ' '; } newtok(p); if ((func & STR_FUNC_EXPAND) && c == '#') { 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) == -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_REGEXP) { unterminated_literal("unterminated regexp meets end of file"); } else { unterminated_literal("unterminated string meets end of file"); } quote->u1.func |= STR_FUNC_TERM; } } tokfix(p); add_mark_object(p, lit = STR_NEW3(tok(p), toklen(p), enc, func)); set_yylval_str(lit); flush_string_content(p, enc); return tSTRING_CONTENT; } static enum yytokentype heredoc_identifier(struct parser_params *p) { /* * term_len is length of `<<"END"` except `END`, * in this case term_len is 4 (<, <, " and "). */ int c = nextc(p), term, func = 0, term_len = 2; enum yytokentype token = tSTRING_BEG; long len; int newline = 0; int indent = 0; if (c == '-') { c = nextc(p); term_len++; func = STR_FUNC_INDENT; } else if (c == '~') { c = nextc(p); term_len++; func = STR_FUNC_INDENT; indent = INT_MAX; } switch (c) { case '\'': term_len++; func |= str_squote; goto quoted; case '"': term_len++; func |= str_dquote; goto quoted; case '`': term_len++; token = tXSTRING_BEG; func |= str_xquote; goto quoted; quoted: term_len++; newtok(p); tokadd(p, term_len); tokadd(p, func); term = c; while ((c = nextc(p)) != -1 && c != term) { if (tokadd_mbchar(p, c) == -1) return 0; if (!newline && c == '\n') newline = 1; else if (newline) newline = 2; } if (c == -1) { yyerror(NULL, p, "unterminated here document identifier"); return -1; } switch (newline) { case 1: rb_warn0("here document identifier ends with a newline"); if (--p->tokidx > 0 && p->tokenbuf[p->tokidx] == '\r') --p->tokidx; break; case 2: compile_error(p, "here document identifier across newlines, never match"); return -1; } break; default: if (!parser_is_identchar(p)) { pushback(p, c); if (func & STR_FUNC_INDENT) { pushback(p, indent > 0 ? '~' : '-'); } return 0; } newtok(p); tokadd(p, term_len); tokadd(p, func |= str_dquote); do { if (tokadd_mbchar(p, c) == -1) return 0; } while ((c = nextc(p)) != -1 && parser_is_identchar(p)); pushback(p, c); break; } tokfix(p); dispatch_scan_event(p, tHEREDOC_BEG); len = p->lex.pcur - p->lex.pbeg; lex_goto_eol(p); p->lex.strterm = new_strterm(STR_NEW(tok(p), toklen(p)), /* term */ p->lex.lastline, /* lastline */ len, /* lastidx */ p->ruby_sourceline); p->lex.strterm->flags |= STRTERM_HEREDOC; token_flush(p); p->heredoc_indent = indent; p->heredoc_line_indent = 0; return token; } static void heredoc_restore(struct parser_params *p, rb_strterm_heredoc_t *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->u3.lastidx; p->heredoc_end = p->ruby_sourceline; p->ruby_sourceline = (int)here->sourceline; token_flush(p); } static int dedent_string(VALUE string, int width) { 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(root) == NODE_ARRAY) 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(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(node) != NODE_ARRAY) 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 (*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; } #define NUM_SUFFIX_R (1<<0) #define NUM_SUFFIX_I (1<<1) #define NUM_SUFFIX_ALL 3 static int 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); if (c == '.') { c = peekc_n(p, 1); if (ISDIGIT(c)) { yyerror0("unexpected fraction part after numeric literal"); p->lex.pcur += 2; while (parser_is_identchar(p)) nextc(p); } } 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); add_mark_object(p, 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; const char *eos, *ptr, *ptr_end; long len; VALUE str = 0; rb_encoding *enc = p->enc; int bol; eos = RSTRING_PTR(here->term); len = RSTRING_LEN(here->term) - 2; /* here->term includes term_len and func */ eos++; /* skip term_len */ indent = (func = *eos++) & STR_FUNC_INDENT; if ((c = nextc(p)) == -1) { error: compile_error(p, "can't find string \"%s\" anywhere before EOF", eos); #ifdef RIPPER if (!has_delayed_token(p)) { dispatch_scan_event(p, tSTRING_CONTENT); } else { if (str) { rb_str_append(p->delayed, str); } 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, p->lex.ptok, len, enc); } dispatch_delayed_token(p, tSTRING_CONTENT); } lex_goto_eol(p); #endif restore: heredoc_restore(p, &p->lex.strterm->u.heredoc); p->lex.strterm = 0; return 0; } bol = was_bol(p); /* `heredoc_line_indent == -1` means * - "after an interpolation in the same line", or * - "in a continuing line" */ if (bol && (p->heredoc_line_indent != -1 || (p->heredoc_line_indent = 0)) && whole_match_p(p, eos, len, indent)) { dispatch_heredoc_end(p); heredoc_restore(p, &p->lex.strterm->u.heredoc); p->lex.strterm = 0; SET_LEX_STATE(EXPR_END); return tSTRING_END; } if (!(func & STR_FUNC_EXPAND)) { do { ptr = RSTRING_PTR(p->lex.lastline); ptr_end = p->lex.pend; if (ptr_end > ptr) { switch (ptr_end[-1]) { case '\n': if (--ptr_end == ptr || ptr_end[-1] != '\r') { ptr_end++; break; } case '\r': --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, ptr, ptr_end - ptr); else 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(p, eos, len, indent)); } else { /* int mb = ENC_CODERANGE_7BIT, *mbp = &mb;*/ newtok(p); if (c == '#') { 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; } if (t) return t; tokadd(p, '#'); c = nextc(p); } do { pushback(p, c); if ((c = tokadd_string(p, func, '\n', 0, NULL, &enc)) == -1) { if (p->eofp) goto error; goto restore; } if (c != '\n') { if (c == '\\') p->heredoc_line_indent = -1; flush: str = STR_NEW3(tok(p), toklen(p), enc, func); flush_str: set_yylval_str(str); add_mark_object(p, str); #ifndef RIPPER if (bol) yylval.node->flags |= NODE_FL_NEWLINE; #endif flush_string_content(p, enc); return tSTRING_CONTENT; } 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); p->lex.strterm = NEW_STRTERM(func | STR_FUNC_TERM, 0, 0); set_yylval_str(str); add_mark_object(p, 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 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 formal_argument(struct parser_params *p, ID lhs) { switch (id_type(lhs)) { case ID_LOCAL: break; #ifndef RIPPER case ID_CONST: yyerror0("formal argument cannot be a constant"); return 0; case ID_INSTANCE: yyerror0("formal argument cannot be an instance variable"); return 0; case ID_GLOBAL: yyerror0("formal argument cannot be a global variable"); return 0; case ID_CLASS: yyerror0("formal argument cannot be a class variable"); return 0; default: yyerror0("formal argument must be local variable"); return 0; #else default: lhs = dispatch1(param_error, lhs); ripper_error(p); return 0; #endif } shadowing_lvar(p, lhs); 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 parser_set_encode(struct parser_params *p, const char *name) { 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 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); } 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; } rb_compile_warning(p->ruby_sourcefile, p->ruby_sourceline, "invalid value for %s: %s", name, 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)), (b ? Qtrue : Qfalse)); } # 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}, {"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; } 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 = nextc(p); 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; 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) { char tmp[30]; trailing_uc: literal_flush(p, p->lex.pcur - 1); snprintf(tmp, sizeof(tmp), "trailing `%c' in number", nondigit); yyerror0(tmp); } 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; VALUE lit; 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 = 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_warn1("invalid character syntax; use ?\\%c", WARN_I(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(); if (!parser_tokadd_utf8(p, &enc, -1, 0, 0)) return 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); add_mark_object(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; if (IS_BEG()) { int term; int paren; c = nextc(p); quotation: if (c == -1 || !ISALNUM(c)) { term = c; c = 'Q'; } else { term = nextc(p); if (rb_enc_isalnum(term, p->enc) || !parser_isascii(p)) { yyerror0("unknown type of %string"); return 0; } } if (c == -1 || term == -1) { 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; 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) { register int c; SET_LEX_STATE(EXPR_END); 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, '$'); do { tokadd(p, c); c = nextc(p); } while (c != -1 && ISDIGIT(c)); pushback(p, c); if (IS_lex_state_for(last_state, EXPR_FNAME)) goto gvar; tokfix(p); set_yylval_node(NEW_NTH_REF(parse_numvar(p), &_cur_loc)); return tNTH_REF; default: if (!parser_is_identchar(p)) { if (c == -1 || ISSPACE(c)) { compile_error(p, "`$' without identifiers is not allowed as a global variable name"); } else { pushback(p, c); compile_error(p, "`$%c' is not allowed as a global variable name", c); } return 0; } case '0': tokadd(p, '$'); } if (tokadd_ident(p, c)) return 0; SET_LEX_STATE(EXPR_END); tokenize_ident(p, last_state); return tGVAR; } static enum yytokentype parse_atmark(struct parser_params *p, const enum lex_state_e last_state) { enum yytokentype result = tIVAR; register int c = nextc(p); newtok(p); tokadd(p, '@'); if (c == '@') { result = tCVAR; tokadd(p, '@'); c = nextc(p); } if (c == -1 || ISSPACE(c)) { 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"); } return 0; } else if (ISDIGIT(c) || !parser_is_identchar(p)) { pushback(p, c); if (result == tIVAR) { compile_error(p, "`@%c' is not allowed as an instance variable name", c); } else { compile_error(p, "`@@%c' is not allowed as a class variable name", c); } return 0; } if (tokadd_ident(p, c)) return 0; SET_LEX_STATE(EXPR_END); 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; SET_LEX_STATE(kw->state); if (IS_lex_state_for(state, EXPR_FNAME)) { set_yylval_name(rb_intern2(tok(p), toklen(p))); return kw->id[0]; } 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]; } } } 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: last_state = p->lex.state; #ifndef RIPPER token_flush(p); #endif switch (c = nextc(p)) { case '\0': /* NUL */ case '\004': /* ^D */ case '\032': /* ^Z */ case -1: /* end of script. */ return 0; /* white spaces */ case ' ': case '\t': case '\f': case '\r': case '\13': /* '\v' */ 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 */ 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': 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->in_kwarg) { goto normal_newline; } goto retry; } while (1) { switch (c = nextc(p)) { case ' ': case '\t': case '\f': case '\r': case '\13': /* '\v' */ space_seen = 1; break; 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(p)) == '*') { if ((c = nextc(p)) == '=') { set_yylval_id(idPow); SET_LEX_STATE(EXPR_BEG); return tOP_ASGN; } 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 == '=') { set_yylval_id('*'); SET_LEX_STATE(EXPR_BEG); return tOP_ASGN; } pushback(p, c); if (IS_SPCARG(c)) { rb_warning0("`*' interpreted as argument prefix"); c = tSTAR; } else if (IS_BEG()) { c = tSTAR; } else { c = warn_balanced('*', "*", "argument prefix"); } } SET_LEX_STATE(IS_AFTER_OPERATOR() ? EXPR_ARG : EXPR_BEG); return c; case '!': 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(p, c); return '!'; case '=': if (was_bol(p)) { /* skip embedded rd document */ if (strncmp(p->lex.pcur, "begin", 5) == 0 && ISSPACE(p->lex.pcur[5])) { int first_p = TRUE; lex_goto_eol(p); dispatch_scan_event(p, tEMBDOC_BEG); for (;;) { lex_goto_eol(p); if (!first_p) { dispatch_scan_event(p, tEMBDOC); } first_p = FALSE; c = nextc(p); if (c == -1) { compile_error(p, "embedded document meets end of file"); return 0; } if (c != '=') continue; if (c == '=' && strncmp(p->lex.pcur, "end", 3) == 0 && (p->lex.pcur + 3 == p->lex.pend || ISSPACE(p->lex.pcur[3]))) { break; } } lex_goto_eol(p); dispatch_scan_event(p, tEMBDOC_END); goto retry; } } SET_LEX_STATE(IS_AFTER_OPERATOR() ? EXPR_ARG : EXPR_BEG); if ((c = nextc(p)) == '=') { if ((c = nextc(p)) == '=') { return tEQQ; } pushback(p, c); return tEQ; } if (c == '~') { return tMATCH; } else if (c == '>') { return tASSOC; } pushback(p, c); return '='; case '<': last_state = p->lex.state; c = nextc(p); if (c == '<' && !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; } 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(p)) == '>') { return tCMP; } pushback(p, c); return tLEQ; } if (c == '<') { if ((c = nextc(p)) == '=') { set_yylval_id(idLTLT); SET_LEX_STATE(EXPR_BEG); return tOP_ASGN; } pushback(p, c); return warn_balanced((enum ruby_method_ids)tLSHFT, "<<", "here document"); } pushback(p, c); return '<'; case '>': SET_LEX_STATE(IS_AFTER_OPERATOR() ? EXPR_ARG : EXPR_BEG); if ((c = nextc(p)) == '=') { return tGEQ; } if (c == '>') { if ((c = nextc(p)) == '=') { set_yylval_id(idGTGT); SET_LEX_STATE(EXPR_BEG); return tOP_ASGN; } pushback(p, c); return tRSHFT; } pushback(p, c); return '>'; case '"': label = (IS_LABEL_POSSIBLE() ? str_label : 0); p->lex.strterm = NEW_STRTERM(str_dquote | label, '"', 0); return tSTRING_BEG; case '`': if (IS_lex_state(EXPR_FNAME)) { SET_LEX_STATE(EXPR_ENDFN); return c; } if (IS_lex_state(EXPR_DOT)) { if (cmd_state) SET_LEX_STATE(EXPR_CMDARG); else SET_LEX_STATE(EXPR_ARG); return c; } p->lex.strterm = NEW_STRTERM(str_xquote, '`', 0); return tXSTRING_BEG; case '\'': label = (IS_LABEL_POSSIBLE() ? str_label : 0); p->lex.strterm = NEW_STRTERM(str_squote | label, '\'', 0); return tSTRING_BEG; case '?': return parse_qmark(p, space_seen); case '&': 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(p, c); return tANDOP; } else if (c == '=') { set_yylval_id('&'); SET_LEX_STATE(EXPR_BEG); return tOP_ASGN; } 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 (IS_BEG()) { c = tAMPER; } else { c = warn_balanced('&', "&", "argument prefix"); } SET_LEX_STATE(IS_AFTER_OPERATOR() ? EXPR_ARG : EXPR_BEG); return c; case '|': 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(p, c); return tOROP; } if (c == '=') { set_yylval_id('|'); SET_LEX_STATE(EXPR_BEG); return tOP_ASGN; } SET_LEX_STATE(IS_AFTER_OPERATOR() ? EXPR_ARG : EXPR_BEG|EXPR_LABEL); pushback(p, c); return '|'; case '+': c = nextc(p); if (IS_AFTER_OPERATOR()) { SET_LEX_STATE(EXPR_ARG); if (c == '@') { return tUPLUS; } pushback(p, c); return '+'; } if (c == '=') { set_yylval_id('+'); SET_LEX_STATE(EXPR_BEG); return tOP_ASGN; } 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; } SET_LEX_STATE(EXPR_BEG); pushback(p, c); return warn_balanced('+', "+", "unary operator"); case '-': c = nextc(p); if (IS_AFTER_OPERATOR()) { SET_LEX_STATE(EXPR_ARG); if (c == '@') { return tUMINUS; } pushback(p, c); return '-'; } if (c == '=') { set_yylval_id('-'); SET_LEX_STATE(EXPR_BEG); return tOP_ASGN; } 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; } SET_LEX_STATE(EXPR_BEG); pushback(p, c); return warn_balanced('-', "-", "unary operator"); case '.': SET_LEX_STATE(EXPR_BEG); if ((c = nextc(p)) == '.') { if ((c = nextc(p)) == '.') { return tDOT3; } pushback(p, c); return tDOT2; } pushback(p, c); if (c != -1 && ISDIGIT(c)) { yyerror0("no . floating literal anymore; put 0 before dot"); } 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 '}': /* 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(p); if (c == ':') { if (IS_BEG() || IS_lex_state(EXPR_CLASS) || IS_SPCARG(-1)) { SET_LEX_STATE(EXPR_BEG); return tCOLON3; } set_yylval_id(idCOLON2); SET_LEX_STATE(EXPR_DOT); return tCOLON2; } if (IS_END() || ISSPACE(c) || c == '#') { pushback(p, c); c = warn_balanced(':', ":", "symbol literal"); SET_LEX_STATE(EXPR_BEG); return c; } switch (c) { case '\'': p->lex.strterm = NEW_STRTERM(str_ssym, c, 0); break; case '"': p->lex.strterm = NEW_STRTERM(str_dsym, c, 0); break; default: pushback(p, c); break; } SET_LEX_STATE(EXPR_FNAME); return tSYMBEG; case '/': if (IS_BEG()) { p->lex.strterm = NEW_STRTERM(str_regexp, '/', 0); return tREGEXP_BEG; } if ((c = nextc(p)) == '=') { set_yylval_id('/'); SET_LEX_STATE(EXPR_BEG); return tOP_ASGN; } pushback(p, c); if (IS_SPCARG(c)) { arg_ambiguous(p, '/'); p->lex.strterm = NEW_STRTERM(str_regexp, '/', 0); return tREGEXP_BEG; } SET_LEX_STATE(IS_AFTER_OPERATOR() ? EXPR_ARG : EXPR_BEG); return warn_balanced('/', "/", "regexp literal"); case '^': if ((c = nextc(p)) == '=') { set_yylval_id('^'); SET_LEX_STATE(EXPR_BEG); return tOP_ASGN; } SET_LEX_STATE(IS_AFTER_OPERATOR() ? EXPR_ARG : EXPR_BEG); pushback(p, c); return '^'; case ';': SET_LEX_STATE(EXPR_BEG); p->command_start = TRUE; return ';'; case ',': SET_LEX_STATE(EXPR_BEG|EXPR_LABEL); return ','; case '~': if (IS_AFTER_OPERATOR()) { if ((c = nextc(p)) != '@') { pushback(p, c); } SET_LEX_STATE(EXPR_ARG); } else { SET_LEX_STATE(EXPR_BEG); } return '~'; case '(': if (IS_BEG()) { c = tLPAREN; } 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); SET_LEX_STATE(EXPR_BEG|EXPR_LABEL); return c; case '[': p->lex.paren_nest++; if (IS_AFTER_OPERATOR()) { if ((c = nextc(p)) == ']') { SET_LEX_STATE(EXPR_ARG); if ((c = nextc(p)) == '=') { return tASET; } pushback(p, c); return tAREF; } pushback(p, c); SET_LEX_STATE(EXPR_ARG|EXPR_LABEL); return '['; } else if (IS_BEG()) { c = tLBRACK; } else if (IS_ARG() && (space_seen || IS_lex_state(EXPR_LABELED))) { c = tLBRACK; } SET_LEX_STATE(EXPR_BEG|EXPR_LABEL); COND_PUSH(0); CMDARG_PUSH(0); return c; case '{': ++p->lex.brace_nest; if (lambda_beginning_p()) { SET_LEX_STATE(EXPR_BEG); COND_PUSH(0); CMDARG_PUSH(0); p->lex.paren_nest++; return tLAMBEG; } p->lex.paren_nest++; 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 (IS_lex_state(EXPR_ENDARG)) c = tLBRACE_ARG; /* block (expr) */ else c = tLBRACE; /* hash */ COND_PUSH(0); CMDARG_PUSH(0); SET_LEX_STATE(c != tLBRACE ? EXPR_BEG : EXPR_BEG|EXPR_LABEL); if (c != tLBRACE) p->command_start = TRUE; return c; case '\\': c = nextc(p); if (c == '\n') { space_seen = 1; dispatch_scan_event(p, tSP); goto retry; /* skip \\n */ } if (c == ' ') return tSP; if (ISSPACE(c)) return c; pushback(p, c); return '\\'; case '%': return parse_percent(p, space_seen, last_state); case '$': return parse_gvar(p, last_state); case '@': return parse_atmark(p, last_state); case '_': 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(p); break; default: if (!parser_is_identchar(p)) { compile_error(p, "Invalid char `\\x%02X' in expression", c); goto retry; } newtok(p); break; } return parse_ident(p, c, cmd_state); } static enum yytokentype yylex(YYSTYPE *lval, YYLTYPE *yylloc, struct parser_params *p) { enum yytokentype t; p->lval = lval; lval->val = Qundef; t = parser_yylex(p); if (has_delayed_token(p)) dispatch_delayed_token(p, t); else if (t != 0) dispatch_scan_event(p, t); if (p->lex.strterm && (p->lex.strterm->flags & STRTERM_HEREDOC)) RUBY_SET_YYLLOC_FROM_STRTERM_HEREDOC(*yylloc); else RUBY_SET_YYLLOC(*yylloc); return t; } #define LVAR_USED ((ID)1 << (sizeof(ID) * CHAR_BIT - 1)) 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); 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 *node) /* for debug */ { return (enum node_type)nd_type(node); } static int nodeline(NODE *node) { return nd_line(node); } static NODE* newline_node(NODE *node) { if (node) { node = remove_begin(node); node->flags |= NODE_FL_NEWLINE; } return node; } static void fixpos(NODE *node, NODE *orig) { if (!node) return; if (!orig) return; nd_set_line(node, nd_line(orig)); } static void parser_warning(struct parser_params *p, NODE *node, const char *mesg) { rb_compile_warning(p->ruby_sourcefile, nd_line(node), "%s", mesg); } static void parser_warn(struct parser_params *p, NODE *node, const char *mesg) { rb_compile_warn(p->ruby_sourcefile, nd_line(node), "%s", mesg); } static NODE* block_append(struct parser_params *p, NODE *head, NODE *tail) { NODE *end, *h = head, *nd; if (tail == 0) return head; if (h == 0) return tail; switch (nd_type(h)) { case NODE_LIT: case NODE_STR: 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, &head->nd_loc); end->nd_end = end; head = end; break; case NODE_BLOCK: end = h->nd_end; 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, &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(struct parser_params *p, NODE *list, NODE *item) { NODE *last; if (list == 0) return NEW_LIST(item, &item->nd_loc); if (list->nd_next) { last = list->nd_next->nd_end; } else { last = list; } list->nd_alen += 1; 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(NODE *head, NODE *tail) { NODE *last; if (head->nd_next) { last = head->nd_next->nd_end; } else { last = head; } head->nd_alen += tail->nd_alen; last->nd_next = tail; if (tail->nd_next) { head->nd_next->nd_end = tail->nd_next->nd_end; } else { 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; } /* concat two string literals */ static NODE * literal_concat(struct parser_params *p, NODE *head, NODE *tail, const YYLTYPE *loc) { enum node_type htype; NODE *headlast; VALUE lit; if (!head) return tail; if (!tail) return head; htype = nd_type(head); if (htype == NODE_EVSTR) { NODE *node = NEW_DSTR(add_mark_object(p, STR_NEW0()), loc); head = list_append(p, node, head); 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 (htype == NODE_DSTR && (headlast = head->nd_next->nd_end->nd_head) && nd_type(headlast) == NODE_STR) { htype = NODE_STR; lit = headlast->nd_lit; } else { lit = head->nd_lit; } if (htype == NODE_STR) { 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(p, head, tail); } break; case NODE_DSTR: if (htype == NODE_STR) { if (!literal_concat0(p, head->nd_lit, tail->nd_lit)) goto error; tail->nd_lit = head->nd_lit; rb_discard_node(p, head); head = tail; } else if (NIL_P(tail->nd_lit)) { append: head->nd_alen += tail->nd_alen - 1; 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 (htype == NODE_DSTR && (headlast = head->nd_next->nd_end->nd_head) && nd_type(headlast) == NODE_STR) { lit = headlast->nd_lit; 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, loc); list_concat(head, tail); } break; case NODE_EVSTR: if (htype == NODE_STR) { nd_set_type(head, NODE_DSTR); head->nd_alen = 1; } list_append(p, head, tail); break; } return head; } static NODE * evstr2dstr(struct parser_params *p, NODE *node) { if (nd_type(node) == NODE_EVSTR) { node = list_append(p, NEW_DSTR(add_mark_object(p, STR_NEW0()), &node->nd_loc), node); } return node; } static NODE * new_evstr(struct parser_params *p, NODE *node, const YYLTYPE *loc) { NODE *head = node; if (node) { switch (nd_type(node)) { case NODE_STR: case NODE_DSTR: case NODE_EVSTR: return node; } } return NEW_EVSTR(head, loc); } 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); 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(node) == NODE_ONCE ? (node)->nd_body : node) static NODE* match_op(struct parser_params *p, NODE *node1, NODE *node2, const YYLTYPE *op_loc, const YYLTYPE *loc) { NODE *n; int line = op_loc->beg_pos.lineno; value_expr(node1); value_expr(node2); if (node1 && (n = nd_once_body(node1)) != 0) { switch (nd_type(n)) { case NODE_DREGX: { NODE *match = NEW_MATCH2(node1, node2, loc); nd_set_line(match, line); return match; } case NODE_LIT: 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 && (n = nd_once_body(node2)) != 0) { NODE *match3; switch (nd_type(n)) { case NODE_LIT: if (!RB_TYPE_P(n->nd_lit, T_REGEXP)) break; /* fallthru */ case NODE_DREGX: match3 = NEW_MATCH3(node2, node1, loc); return match3; } } 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 #define WARN_LOCATION(type) do { \ if (p->warn_location) { \ rb_warning0(type" in eval may not return location in binding;" \ " use Binding#source_location instead"); \ } \ } while (0) static NODE* 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__: WARN_LOCATION("__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(add_mark_object(p, file), loc); } return node; case keyword__LINE__: WARN_LOCATION("__LINE__"); return NEW_LIT(INT2FIX(p->tokline), loc); case keyword__ENCODING__: return NEW_LIT(add_mark_object(p, rb_enc_from_encoding(p->enc)), loc); } switch (id_type(id)) { case ID_LOCAL: if (dyna_in_block(p) && dvar_defined_ref(p, id, &vidp)) { if (id == p->cur_arg) { rb_warn1("circular argument reference - %"PRIsWARN, rb_id2str(id)); } if (vidp) *vidp |= LVAR_USED; node = NEW_DVAR(id, loc); return node; } if (local_id_ref(p, id, &vidp)) { if (id == p->cur_arg) { rb_warn1("circular argument reference - %"PRIsWARN, rb_id2str(id)); } if (vidp) *vidp |= LVAR_USED; node = NEW_LVAR(id, loc); return node; } # if WARN_PAST_SCOPE if (!p->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) { 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; } 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; } 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) { if (nd_type(symbol) == NODE_DSTR) { nd_set_type(symbol, NODE_DSYM); } else { nd_set_type(symbol, NODE_LIT); symbol->nd_lit = add_mark_object(p, rb_str_intern(symbol->nd_lit)); } 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) { return NEW_LIT(add_mark_object(p, reg_compile(p, STR_NEW0(), options)), loc); } switch (nd_type(node)) { case NODE_STR: { VALUE src = node->nd_lit; nd_set_type(node, NODE_LIT); nd_set_loc(node, loc); add_mark_object(p, node->nd_lit = reg_compile(p, src, options)); } break; default: add_mark_object(p, lit = STR_NEW0()); node = NEW_NODE(NODE_DSTR, lit, 1, NEW_LIST(node, loc), loc); 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) { if (nd_type(list->nd_head) == NODE_STR) { VALUE tail = list->nd_head->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); add_mark_object(p, node->nd_lit = reg_compile(p, src, options)); } if (options & RE_OPTION_ONCE) { node = NEW_NODE(NODE_ONCE, 0, node, 0, loc); } break; } 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); add_mark_object(p, lit); return xstr; } 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; } return node; } #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) && dvar_defined(p, id)) return 1; if (local_id(p, id)) return 1; /* method call without arguments */ return 0; } } compile_error(p, "identifier %"PRIsVALUE" is not valid to get", rb_id2str(id)); return 0; } static VALUE new_regexp(struct parser_params *p, VALUE re, VALUE opt, const YYLTYPE *loc) { VALUE src = 0, err; int options = 0; if (ripper_is_node_yylval(re)) { src = RNODE(re)->nd_cval; re = RNODE(re)->nd_rval; } if (ripper_is_node_yylval(opt)) { options = (int)RNODE(opt)->nd_tag; opt = RNODE(opt)->nd_rval; } if (src && NIL_P(parser_reg_compile(p, src, options, &err))) { compile_error(p, "%"PRIsVALUE, err); } return dispatch2(regexp_literal, re, opt); } #endif /* !RIPPER */ #ifndef RIPPER static const char rb_parser_lex_state_names[][13] = { "EXPR_BEG", "EXPR_END", "EXPR_ENDARG", "EXPR_ENDFN", "EXPR_ARG", "EXPR_CMDARG", "EXPR_MID", "EXPR_FNAME", "EXPR_DOT", "EXPR_CLASS", "EXPR_LABEL", "EXPR_LABELED","EXPR_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[] = "EXPR_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]); } } if (!sep) { rb_str_cat(buf, none, sizeof(none)-1); } 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 { 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); } } 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 p error: "); va_start(ap, fmt); rb_str_vcatf(mesg, fmt, ap); va_end(ap); parser_yyerror(p, NULL, RSTRING_PTR(mesg)); RB_GC_GUARD(mesg); mesg = rb_str_new(0, 0); append_lex_state_name(p->lex.state, mesg); compile_error(p, "p->lex.state: %"PRIsVALUE, mesg); rb_str_resize(mesg, 0); append_bitstack_value(p->cond_stack, mesg); compile_error(p, "p->cond_stack: %"PRIsVALUE, mesg); rb_str_resize(mesg, 0); append_bitstack_value(p->cmdarg_stack, mesg); compile_error(p, "p->cmdarg_stack: %"PRIsVALUE, mesg); if (p->debug_output == rb_stdout) p->debug_output = rb_stderr; p->debug = TRUE; } void rb_parser_set_location_from_strterm_heredoc(struct parser_params *p, rb_strterm_heredoc_t *here, YYLTYPE *yylloc) { const char *eos = RSTRING_PTR(here->term); long term_len = RSTRING_LEN(here->term) - 2 + (unsigned char)eos[0]; yylloc->beg_pos.lineno = (int)here->sourceline; yylloc->beg_pos.column = (int)(here->u3.lastidx - term_len); yylloc->end_pos.lineno = (int)here->sourceline; yylloc->end_pos.column = (int)(here->u3.lastidx); } void rb_parser_set_location_of_none(struct parser_params *p, YYLTYPE *yylloc) { yylloc->beg_pos.lineno = p->ruby_sourceline; yylloc->beg_pos.column = (int)(p->lex.ptok - p->lex.pbeg); yylloc->end_pos.lineno = p->ruby_sourceline; yylloc->end_pos.column = (int)(p->lex.ptok - p->lex.pbeg); } void rb_parser_set_location(struct parser_params *p, YYLTYPE *yylloc) { yylloc->beg_pos.lineno = p->ruby_sourceline; yylloc->beg_pos.column = (int)(p->lex.ptok - p->lex.pbeg); yylloc->end_pos.lineno = p->ruby_sourceline; yylloc->end_pos.column = (int)(p->lex.pcur - p->lex.pbeg); } #endif /* !RIPPER */ static void parser_token_value_print(struct parser_params *p, enum yytokentype type, const YYSTYPE *valp) { VALUE v; switch (type) { case tIDENTIFIER: case tFID: case tGVAR: case tIVAR: case tCONSTANT: case tCVAR: case tLABEL: case tOP_ASGN: #ifndef RIPPER v = rb_id2str(valp->id); #else v = valp->val; #endif rb_parser_printf(p, "%"PRIsVALUE, v); break; case tINTEGER: case tFLOAT: case tRATIONAL: case tIMAGINARY: case tSTRING_CONTENT: case tCHAR: #ifndef RIPPER v = valp->node->nd_lit; #else v = valp->val; #endif rb_parser_printf(p, "%+"PRIsVALUE, v); break; case tNTH_REF: #ifndef RIPPER rb_parser_printf(p, "$%ld", valp->node->nd_nth); #else rb_parser_printf(p, "%"PRIsVALUE, valp->val); #endif break; case tBACK_REF: #ifndef RIPPER rb_parser_printf(p, "$%c", (int)valp->node->nd_nth); #else rb_parser_printf(p, "%"PRIsVALUE, valp->val); #endif break; default: break; } } 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; } switch (id_type(id)) { case ID_LOCAL: if (dyna_in_block(p)) { if (dvar_curr(p, id)) return NODE_DASGN_CURR; if (dvar_defined(p, id)) return NODE_DASGN; if (local_id(p, id)) return NODE_LASGN; dyna_var(p, id); return NODE_DASGN_CURR; } 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->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)); } 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_CURR: return NEW_DASGN_CURR(id, val, loc); 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, 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 (is_private_local_id(name)) return 1; if (dyna_in_block(p)) { if (dvar_curr(p, name)) { 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 { if (local_id(p, name)) { yyerror0("duplicated argument name"); } } 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(struct parser_params *p, NODE *recv, NODE *idx, const YYLTYPE *loc) { return NEW_ATTRASGN(recv, tASET, idx, loc); } static void block_dup_check(struct parser_params *p, NODE *node1, NODE *node2) { if (node2 && node1 && nd_type(node1) == NODE_BLOCK_PASS) { compile_error(p, "both block arg and actual block given"); } } static NODE * attrset(struct parser_params *p, NODE *recv, ID atype, ID id, const YYLTYPE *loc) { if (!CALL_Q_P(atype)) id = rb_id_attrset(id); return NEW_ATTRASGN(recv, id, 0, loc); } static void rb_backref_error(struct parser_params *p, NODE *node) { switch (nd_type(node)) { case NODE_NTH_REF: compile_error(p, "Can't set variable $%ld", node->nd_nth); break; case NODE_BACK_REF: compile_error(p, "Can't set variable $%c", (int)node->nd_nth); break; } } 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_ARRAY: 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(node1->nd_body) != NODE_ARRAY) 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(struct parser_params *p, NODE *node1, NODE *node2, const YYLTYPE *loc) { if (!node2) return node1; 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(node2) != NODE_ARRAY) 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(node2) != NODE_ARRAY || nd_type(node1->nd_body) != NODE_ARRAY) 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(rest_arg) == NODE_ARRAY) && (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(node) == NODE_SPLAT) node = node->nd_head; if (nd_type(node) == NODE_ARRAY) 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: case NODE_DASGN_CURR: 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 * node_assign(struct parser_params *p, NODE *lhs, NODE *rhs, const YYLTYPE *loc) { if (!lhs) return 0; switch (nd_type(lhs)) { case NODE_GASGN: case NODE_IASGN: case NODE_LASGN: case NODE_DASGN: case NODE_DASGN_CURR: case NODE_MASGN: case NODE_CDECL: case NODE_CVASGN: lhs->nd_value = rhs; nd_set_loc(lhs, loc); break; case NODE_ATTRASGN: lhs->nd_args = arg_append(p, lhs->nd_args, rhs, loc); nd_set_loc(lhs, loc); break; default: /* should not happen */ break; } return lhs; } static int value_expr_gen(struct parser_params *p, NODE *node) { int cond = 0; if (!node) { rb_warning0("empty expression"); } while (node) { switch (nd_type(node)) { case NODE_RETURN: case NODE_BREAK: case NODE_NEXT: case NODE_REDO: case NODE_RETRY: if (!cond) yyerror1(&node->nd_loc, "void value expression"); /* or "control never reach"? */ return FALSE; case NODE_BLOCK: while (node->nd_next) { node = node->nd_next; } node = node->nd_head; break; case NODE_BEGIN: node = node->nd_body; break; case NODE_IF: case NODE_UNLESS: if (!node->nd_body) { node = node->nd_else; break; } else if (!node->nd_else) { node = node->nd_body; break; } if (!value_expr(node->nd_body)) return FALSE; node = node->nd_else; break; case NODE_AND: case NODE_OR: cond = 1; node = node->nd_2nd; break; case NODE_LASGN: case NODE_DASGN: case NODE_DASGN_CURR: case NODE_MASGN: mark_lvar_used(p, node); return TRUE; default: return TRUE; } } return TRUE; } static void void_expr(struct parser_params *p, NODE *node) { const char *useless = 0; if (!RTEST(ruby_verbose)) return; if (!node || !(node = nd_once_body(node))) return; switch (nd_type(node)) { case NODE_OPCALL: switch (node->nd_mid) { case '+': case '-': case '*': case '/': case '%': case tPOW: case tUPLUS: case tUMINUS: case '|': case '^': case '&': case tCMP: case '>': case tGEQ: case '<': case tLEQ: case tEQ: case tNEQ: useless = rb_id2name(node->nd_mid); break; } break; case NODE_LVAR: case NODE_DVAR: case NODE_GVAR: case NODE_IVAR: case NODE_CVAR: case NODE_NTH_REF: case NODE_BACK_REF: useless = "a variable"; break; case NODE_CONST: useless = "a constant"; break; case NODE_LIT: case NODE_STR: case NODE_DSTR: case NODE_DREGX: useless = "a literal"; break; case NODE_COLON2: case NODE_COLON3: useless = "::"; break; case NODE_DOT2: useless = ".."; break; case NODE_DOT3: useless = "..."; break; case NODE_SELF: useless = "self"; break; case NODE_NIL: useless = "nil"; break; case NODE_TRUE: useless = "true"; break; case NODE_FALSE: useless = "false"; break; case NODE_DEFINED: useless = "defined?"; break; } if (useless) { rb_warn1L(nd_line(node), "possibly useless use of %s in void context", WARN_S(useless)); } } static NODE * void_stmts(struct parser_params *p, NODE *node) { NODE *const n = node; if (!RTEST(ruby_verbose)) return n; if (!node) return n; if (nd_type(node) != NODE_BLOCK) return n; while (node->nd_next) { void_expr(p, node->nd_head); node = node->nd_next; } return n; } static NODE * remove_begin(NODE *node) { NODE **n = &node, *n1 = node; while (n1 && nd_type(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(n1) == NODE_BEGIN) { *n = n1 = n1->nd_body; } return node; } static void reduce_nodes(struct parser_params *p, NODE **body) { 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: *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 = *body; if (newline && node) node->flags |= NODE_FL_NEWLINE; } #undef subnodes } static int is_static_content(NODE *node) { if (!node) return 1; switch (nd_type(node)) { case NODE_HASH: if (!(node = node->nd_head)) break; case NODE_ARRAY: 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_ZARRAY: 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_DASGN_CURR: case NODE_GASGN: case NODE_IASGN: break; default: return 0; } if (!node->nd_value) return 1; if (is_static_content(node->nd_value)) { /* reports always */ parser_warn(p, node->nd_value, "found `= literal' in conditional, should be =="); } return 1; } static void warn_unless_e_option(struct parser_params *p, NODE *node, const char *str) { if (!e_option_supplied(p)) parser_warn(p, node, str); } static void warning_unless_e_option(struct parser_params *p, NODE *node, const char *str) { if (!e_option_supplied(p)) parser_warning(p, node, str); } static NODE *cond0(struct parser_params*,NODE*,int,const YYLTYPE*); static NODE* range_op(struct parser_params *p, NODE *node, const YYLTYPE *loc) { enum node_type type; if (node == 0) return 0; type = nd_type(node); value_expr(node); if (type == NODE_LIT && FIXNUM_P(node->nd_lit)) { warn_unless_e_option(p, node, "integer literal in conditional range"); return NEW_CALL(node, tEQ, NEW_LIST(NEW_GVAR(rb_intern("$."), loc), loc), loc); } return cond0(p, node, FALSE, loc); } static int literal_node(NODE *node) { if (!node) return 1; /* same as NODE_NIL */ if (!(node = nd_once_body(node))) return 1; switch (nd_type(node)) { case NODE_LIT: case NODE_STR: case NODE_DSTR: case NODE_EVSTR: case NODE_DREGX: case NODE_DSYM: return 2; case NODE_TRUE: case NODE_FALSE: case NODE_NIL: return 1; } return 0; } static NODE* cond0(struct parser_params *p, NODE *node, int method_op, const YYLTYPE *loc) { if (node == 0) return 0; 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: if (!method_op) rb_warn0("string literal in condition"); break; case NODE_DREGX: { if (!method_op) warning_unless_e_option(p, node, "regex literal in condition"); return NEW_MATCH2(node, NEW_GVAR(idLASTLINE, loc), loc); } case NODE_AND: case NODE_OR: node->nd_1st = cond0(p, node->nd_1st, FALSE, loc); node->nd_2nd = cond0(p, node->nd_2nd, FALSE, loc); break; case NODE_DOT2: case NODE_DOT3: node->nd_beg = range_op(p, node->nd_beg, loc); node->nd_end = range_op(p, node->nd_end, loc); if (nd_type(node) == NODE_DOT2 || nd_type(node) == NODE_DOT3) { nd_set_type(node, nd_type(node) == NODE_DOT2 ? NODE_FLIP2 : NODE_FLIP3); parser_warn(p, node, "flip-flop is deprecated"); } if (!method_op && !e_option_supplied(p)) { 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(p, node, "range literal in condition"); } } break; case NODE_DSYM: if (!method_op) parser_warning(p, node, "literal in condition"); break; case NODE_LIT: if (RB_TYPE_P(node->nd_lit, T_REGEXP)) { if (!method_op) warn_unless_e_option(p, node, "regex literal in condition"); nd_set_type(node, NODE_MATCH); } else if (node->nd_lit == Qtrue || node->nd_lit == Qfalse) { /* booleans are OK, e.g., while true */ } else { if (!method_op) parser_warning(p, node, "literal in condition"); } default: break; } return node; } static NODE* cond(struct parser_params *p, NODE *node, const YYLTYPE *loc) { if (node == 0) return 0; return cond0(p, node, FALSE, loc); } static NODE* method_cond(struct parser_params *p, NODE *node, const YYLTYPE *loc) { if (node == 0) return 0; return cond0(p, node, TRUE, loc); } 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, FALSE, 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, FALSE, 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 && (enum node_type)nd_type(left) == type) { NODE *node = left, *second; while ((second = node->nd_2nd) != 0 && (enum node_type)nd_type(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; } op = NEW_NODE(type, left, right, 0, loc); nd_set_line(op, op_loc->beg_pos.lineno); return op; } static void no_blockarg(struct parser_params *p, NODE *node) { if (node && nd_type(node) == NODE_BLOCK_PASS) { compile_error(p, "block argument should not be given"); } } static NODE * ret_args(struct parser_params *p, NODE *node) { if (node) { no_blockarg(p, node); if (nd_type(node) == NODE_ARRAY) { if (node->nd_next == 0) { node = node->nd_head; } else { nd_set_type(node, NODE_VALUES); } } } return node; } static NODE * new_yield(struct parser_params *p, NODE *node, const YYLTYPE *loc) { if (node) no_blockarg(p, node); return NEW_YIELD(node, loc); } static VALUE negate_lit(struct parser_params *p, VALUE lit) { 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: BIGNUM_NEGATE(lit); lit = rb_big_norm(lit); break; case T_RATIONAL: RRATIONAL_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(lit)->float_value = -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 lit; } static NODE * 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 NODE* new_args(struct parser_params *p, NODE *pre_args, NODE *opt_args, ID rest_arg, NODE *post_args, NODE *tail, const YYLTYPE *loc) { int saved_line = p->ruby_sourceline; struct rb_args_info *args = tail->nd_ainfo; args->pre_args_num = pre_args ? rb_long2int(pre_args->nd_plen) : 0; args->pre_init = pre_args ? pre_args->nd_next : 0; 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; p->ruby_sourceline = saved_line; nd_set_loc(tail, loc); return tail; } static NODE* new_args_tail(struct parser_params *p, NODE *kw_args, ID kw_rest_arg, ID block, const YYLTYPE *loc) { int saved_line = p->ruby_sourceline; struct rb_args_info *args; NODE *node; rb_imemo_tmpbuf_t *tmpbuf = new_tmpbuf(); args = ZALLOC(struct rb_args_info); tmpbuf->ptr = (VALUE *)args; node = NEW_NODE(NODE_ARGS, 0, 0, args, &NULL_LOC); 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 * #=> * 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; 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, loc); args->kw_rest_arg->nd_cflag = kw_bits; } else if (kw_rest_arg) { args->kw_rest_arg = NEW_DVAR(kw_rest_arg, loc); } p->ruby_sourceline = saved_line; return node; } static NODE* dsym_node(struct parser_params *p, NODE *node, const YYLTYPE *loc) { VALUE lit; if (!node) { return NEW_LIT(ID2SYM(idNULL), loc); } 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; add_mark_object(p, 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; } static NODE * remove_duplicate_keys(struct parser_params *p, NODE *hash) { st_table *literal_keys = st_init_numtable_with_size(hash->nd_alen / 2); NODE *result = 0; while (hash && hash->nd_head && hash->nd_next) { NODE *head = hash->nd_head; NODE *value = hash->nd_next; NODE *next = value->nd_next; VALUE key = (VALUE)head; st_data_t data; if (nd_type(head) == NODE_LIT && st_lookup(literal_keys, (key = head->nd_lit), &data)) { 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)); head = ((NODE *)data)->nd_next; head->nd_head = block_append(p, head->nd_head, value->nd_head); } else { st_insert(literal_keys, (st_data_t)key, (st_data_t)hash); } 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 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 /* !RIPPER */ #ifndef RIPPER static NODE * new_op_assign(struct parser_params *p, NODE *lhs, ID op, NODE *rhs, const YYLTYPE *loc) { NODE *asgn; if (lhs) { ID vid = lhs->nd_vid; YYLTYPE lhs_loc = lhs->nd_loc; if (op == tOROP) { 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; } } } else if (op == tANDOP) { lhs->nd_value = rhs; nd_set_loc(lhs, loc); asgn = NEW_OP_ASGN_AND(gettable(p, vid, &lhs_loc), lhs, loc); } else { asgn = lhs; asgn->nd_value = NEW_CALL(gettable(p, vid, &lhs_loc), op, NEW_LIST(rhs, &rhs->nd_loc), loc); nd_set_loc(asgn, loc); } } else { asgn = NEW_BEGIN(0, loc); } return asgn; } 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) { NODE *asgn; args = make_array(args, args_loc); if (nd_type(args) == NODE_BLOCK_PASS) { args = NEW_ARGSCAT(args, rhs, loc); } else { args = arg_concat(p, args, rhs, loc); } asgn = NEW_OP_ASGN1(ary, op, args, loc); fixpos(asgn, ary); return asgn; } static NODE * new_attr_op_assign(struct parser_params *p, NODE *lhs, ID atype, ID attr, ID op, NODE *rhs, const YYLTYPE *loc) { NODE *asgn; asgn = NEW_OP_ASGN2(lhs, CALL_Q_P(atype), attr, op, rhs, loc); fixpos(asgn, lhs); return asgn; } static NODE * new_const_op_assign(struct parser_params *p, NODE *lhs, ID op, NODE *rhs, const YYLTYPE *loc) { NODE *asgn; if (lhs) { asgn = NEW_OP_CDECL(lhs, op, rhs, loc); } else { asgn = NEW_BEGIN(0, loc); } fixpos(asgn, lhs); return asgn; } static NODE * const_decl(struct parser_params *p, NODE *path, const YYLTYPE *loc) { if (p->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->in_def) { path = dispatch1(assign_error, path); ripper_error(p); } return path; } static VALUE assign_error(struct parser_params *p, VALUE a) { a = dispatch1(assign_error, a); ripper_error(p); return a; } static VALUE var_field(struct parser_params *p, VALUE a) { return ripper_new_yylval(p, get_id(a), dispatch1(var_field, a), 0); } #endif #ifndef RIPPER static NODE * new_bodystmt(struct parser_params *p, NODE *head, NODE *rescue, NODE *rescue_else, NODE *ensure, const YYLTYPE *loc) { NODE *result = head; if (rescue) { NODE *tmp = rescue_else ? rescue_else : rescue; YYLTYPE rescue_loc = code_loc_gen(&head->nd_loc, &tmp->nd_loc); 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); } fixpos(result, head); return result; } #endif static void warn_unused_var(struct parser_params *p, struct local_vars *local) { int i, cnt; ID *v, *u; if (!local->used) return; v = local->vars->tbl; u = local->used->tbl; cnt = local->used->pos; if (cnt != local->vars->pos) { rb_parser_fatal(p, "local->used->pos != local->vars->pos"); } for (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])); } } static void local_push(struct parser_params *p, int toplevel_scope) { struct local_vars *local; int inherits_dvars = toplevel_scope && (compile_for_eval || p->in_main /* is p->in_main really needed? */); int warn_unused_vars = RTEST(ruby_verbose); 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; #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; } 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; } #ifndef RIPPER static ID* local_tbl(struct parser_params *p) { 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; ID *buf; rb_imemo_tmpbuf_t *tmpbuf = new_tmpbuf(); if (cnt <= 0) return 0; buf = ALLOC_N(ID, cnt + 1); tmpbuf->ptr = (void *)buf; MEMCPY(buf+1, p->lvtbl->args->tbl, ID, cnt_args); /* remove IDs duplicated to warn shadowing */ for (i = 0, j = cnt_args+1; i < cnt_vars; ++i) { ID id = p->lvtbl->vars->tbl[i]; if (!vtable_included(p->lvtbl->args, id)) { buf[j++] = id; } } if (--j < cnt) tmpbuf->ptr = (void *)REALLOC_N(buf, ID, (cnt = j) + 1); buf[0] = cnt; return buf; } #endif static void arg_var(struct parser_params *p, ID id) { vtable_add(p->lvtbl->args, id); } static void local_var(struct parser_params *p, ID id) { vtable_add(p->lvtbl->vars, id); if (p->lvtbl->used) { vtable_add(p->lvtbl->used, (ID)p->ruby_sourceline); } } static int 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->base_block); } 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; } } static int local_id(struct parser_params *p, ID id) { return local_id_ref(p, id, NULL); } 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; } 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); } static void dyna_pop_1(struct parser_params *p) { struct vtable *tmp; 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); } 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; } } dyna_pop_1(p); } static int dyna_in_block(struct parser_params *p) { return !DVARS_TERMINAL_P(p->lvtbl->vars) && p->lvtbl->vars->prev != DVARS_TOPSCOPE; } static int dvar_defined_ref(struct parser_params *p, ID id, ID **vidrefp) { struct vtable *vars, *args, *used; int i; args = p->lvtbl->args; vars = p->lvtbl->vars; used = p->lvtbl->used; while (!DVARS_TERMINAL_P(vars)) { if (vtable_included(args, id)) { return 1; } 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) { return rb_dvar_defined(id, p->base_block); } return 0; } 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 { rb_enc_associate(str, rb_ascii8bit_encoding()); } } 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; } return 1; } 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 || (*name != '_' && ISASCII(*name) && !rb_enc_islower(*name, enc)) || (len < MAX_WORD_LENGTH && rb_reserved_word(s, (int)len)) || !rb_enc_symname2_p(s, len, enc)) { return ST_CONTINUE; } var = intern_cstr(s, len, enc); node = node_assign(p, assignable(p, var, 0, arg->loc), NEW_LIT(ID2SYM(var), arg->loc), 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) { reg_named_capture_assign_t arg; 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); if (!arg.succ_block) return 0; return arg.succ_block->nd_next; } 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; 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 #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; } void rb_parser_warn_location(VALUE vparser, int warn) { struct parser_params *p; TypedData_Get_Struct(vparser, struct parser_params, &parser_data_type, p); p->warn_location = warn; } 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; if (p->do_print) { NODE *print = NEW_FCALL(rb_intern("print"), NEW_ARRAY(NEW_GVAR(idLASTLINE, LOC), LOC), LOC); node = block_append(p, node, print); } if (p->do_loop) { if (p->do_split) { NODE *split = NEW_GASGN(rb_intern("$F"), NEW_CALL(NEW_GVAR(idLASTLINE, LOC), rb_intern("split"), 0, LOC), LOC); node = block_append(p, split, node); } 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 node; } void rb_init_parse(void) { /* just to suppress unused-function warnings */ (void)nodetype; (void)nodeline; } #endif /* !RIPPER */ static ID internal_id(struct parser_params *p) { ID id = (ID)vtable_size(p->lvtbl->args) + (ID)vtable_size(p->lvtbl->vars); id += ((tLAST_TOKEN - ID_INTERNAL) >> ID_SCOPE_SHIFT) + 1; return ID_STATIC_SYM | ID_INTERNAL | (id << ID_SCOPE_SHIFT); } static void parser_initialize(struct parser_params *p) { /* 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 = Qnil; p->result = Qnil; p->parsing_thread = Qnil; #else p->error_buffer = Qfalse; #endif p->debug_buffer = Qnil; p->debug_output = rb_stdout; p->enc = rb_utf8_encoding(); } #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); #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); 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 } static void parser_free(void *ptr) { 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); } 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) { return reserved_word(str, len); } VALUE rb_parser_new(void) { struct parser_params *p; VALUE parser = TypedData_Make_Struct(0, struct parser_params, &parser_data_type, p); parser_initialize(p); return parser; } VALUE rb_parser_set_context(VALUE vparser, const struct rb_block *base, int main) { struct parser_params *p; TypedData_Get_Struct(vparser, struct parser_params, &parser_data_type, p); p->error_buffer = main ? Qfalse : Qnil; p->base_block = base; p->in_main = main; return vparser; } #endif #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 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); /* * call-seq: * ripper.error? -> Boolean * * Return true if parsed source has errors. */ static VALUE ripper_error_p(VALUE vparser) { struct parser_params *p; TypedData_Get_Struct(vparser, struct parser_params, &parser_data_type, p); return p->error_p ? Qtrue : Qfalse; } #endif /* * call-seq: * ripper.end_seen? -> Boolean * * Return true if parsed source ended by +\_\_END\_\_+. */ VALUE rb_parser_end_seen_p(VALUE vparser) { struct parser_params *p; TypedData_Get_Struct(vparser, struct parser_params, &parser_data_type, p); return p->ruby__end__seen ? Qtrue : Qfalse; } /* * call-seq: * ripper.encoding -> encoding * * Return encoding of the source. */ VALUE rb_parser_encoding(VALUE vparser) { struct parser_params *p; TypedData_Get_Struct(vparser, struct parser_params, &parser_data_type, p); return rb_enc_from_encoding(p->enc); } /* * call-seq: * ripper.yydebug -> true or false * * Get yydebug. */ VALUE rb_parser_get_yydebug(VALUE self) { struct parser_params *p; TypedData_Get_Struct(self, struct parser_params, &parser_data_type, p); return p->debug ? Qtrue : Qfalse; } /* * call-seq: * ripper.yydebug = flag * * Set yydebug. */ VALUE rb_parser_set_yydebug(VALUE self, VALUE flag) { struct parser_params *p; TypedData_Get_Struct(self, struct parser_params, &parser_data_type, p); p->debug = RTEST(flag); return flag; } #ifndef RIPPER #ifdef YYMALLOC #define HEAPCNT(n, size) ((n) * (size) / sizeof(YYSTYPE)) /* 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); rb_imemo_tmpbuf_t *n = NEWHEAP(); void *ptr = xmalloc(size); return ADD2HEAP(n, cnt, ptr); } void * rb_parser_calloc(struct parser_params *p, size_t nelem, size_t size) { size_t cnt = HEAPCNT(nelem, size); rb_imemo_tmpbuf_t *n = NEWHEAP(); void *ptr = xcalloc(nelem, size); return ADD2HEAP(n, cnt, ptr); } void * rb_parser_realloc(struct parser_params *p, void *ptr, size_t size) { rb_imemo_tmpbuf_t *n; size_t cnt = HEAPCNT(1, size); if (ptr && (n = p->heap) != NULL) { do { if (n->ptr == ptr) { n->ptr = ptr = xrealloc(ptr, size); if (n->cnt) n->cnt = cnt; return ptr; } } while ((n = n->next) != NULL); } n = NEWHEAP(); ptr = xrealloc(ptr, size); return ADD2HEAP(n, cnt, ptr); } void rb_parser_free(struct parser_params *p, void *ptr) { rb_imemo_tmpbuf_t **prev = &p->heap, *n; while ((n = *prev) != NULL) { if (n->ptr == ptr) { *prev = n->next; rb_gc_force_recycle((VALUE)n); break; } 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); } #endif #ifdef RIPPER #ifdef RIPPER_DEBUG extern int rb_is_pointer_to_heap(VALUE); /* :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; if (!rb_is_pointer_to_heap(x)) rb_raise(rb_eArgError, "invalid pointer: %p", 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: return x; case T_NODE: if (nd_type(x) != NODE_RIPPER) { rb_raise(rb_eArgError, "NODE given: %p", x); } return ((NODE *)x)->nd_rval; default: rb_raise(rb_eArgError, "wrong type of ruby object: %p (%s)", x, rb_obj_classname(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(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(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); } #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"); 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, "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, "error?", ripper_error_p, 0); #ifdef RIPPER_DEBUG rb_define_method(rb_mKernel, "assert_Qundef", ripper_assert_Qundef, 2); rb_define_method(rb_mKernel, "rawVALUE", ripper_value, 1); rb_define_method(rb_mKernel, "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 */