diff options
Diffstat (limited to 'ruby.c')
-rw-r--r-- | ruby.c | 3074 |
1 files changed, 1730 insertions, 1344 deletions
@@ -22,10 +22,6 @@ # include <sys/cygwin.h> #endif -#ifdef __hpux -# include <sys/pstat.h> -#endif - #if defined(LOAD_RELATIVE) && defined(HAVE_DLADDR) # include <dlfcn.h> #endif @@ -47,6 +43,8 @@ #include "dln.h" #include "eval_intern.h" #include "internal.h" +#include "internal/cmdlineopt.h" +#include "internal/cont.h" #include "internal/error.h" #include "internal/file.h" #include "internal/inits.h" @@ -55,17 +53,19 @@ #include "internal/loadpath.h" #include "internal/missing.h" #include "internal/object.h" -#include "internal/parse.h" -#include "internal/process.h" +#include "internal/thread.h" +#include "internal/ruby_parser.h" #include "internal/variable.h" -#include "mjit.h" -#include "yjit.h" #include "ruby/encoding.h" #include "ruby/thread.h" #include "ruby/util.h" #include "ruby/version.h" #include "ruby/internal/error.h" +#define singlebit_only_p(x) !((x) & ((x)-1)) +STATIC_ASSERT(Qnil_1bit_from_Qfalse, singlebit_only_p(Qnil^Qfalse)); +STATIC_ASSERT(Qundef_1bit_from_Qnil, singlebit_only_p(Qundef^Qnil)); + #ifndef MAXPATHLEN # define MAXPATHLEN 1024 #endif @@ -73,7 +73,7 @@ # define O_ACCMODE (O_RDONLY | O_WRONLY | O_RDWR) #endif -void Init_ruby_description(void); +void Init_ruby_description(ruby_cmdline_options_t *opt); #ifndef HAVE_STDLIB_H char *getenv(); @@ -99,13 +99,15 @@ void rb_warning_category_update(unsigned int mask, unsigned int bits); SEP \ X(did_you_mean) \ SEP \ + X(syntax_suggest) \ + SEP \ X(rubyopt) \ SEP \ X(frozen_string_literal) \ SEP \ - X(jit) \ + X(rjit) \ SEP \ - X(yjit) + X(yjit) \ /* END OF FEATURES */ #define EACH_DEBUG_FEATURES(X, SEP) \ X(frozen_string_literal) \ @@ -116,11 +118,20 @@ void rb_warning_category_update(unsigned int mask, unsigned int bits); enum feature_flag_bits { EACH_FEATURES(DEFINE_FEATURE, COMMA), feature_debug_flag_first, +#if defined(RJIT_FORCE_ENABLE) || !USE_YJIT + DEFINE_FEATURE(jit) = feature_rjit, +#else + DEFINE_FEATURE(jit) = feature_yjit, +#endif + feature_jit_mask = FEATURE_BIT(rjit) | FEATURE_BIT(yjit), + feature_debug_flag_begin = feature_debug_flag_first - 1, EACH_DEBUG_FEATURES(DEFINE_DEBUG_FEATURE, COMMA), feature_flag_count }; +#define MULTI_BITS_P(bits) ((bits) & ((bits) - 1)) + #define DEBUG_BIT(bit) (1U << feature_debug_##bit) #define DUMP_BIT(bit) (1U << dump_##bit) @@ -140,25 +151,21 @@ enum feature_flag_bits { SEP \ X(parsetree) \ SEP \ - X(parsetree_with_comment) \ - SEP \ X(insns) \ /* END OF DUMPS */ enum dump_flag_bits { dump_version_v, + dump_opt_error_tolerant, + dump_opt_comment, + dump_opt_optimize, EACH_DUMPS(DEFINE_DUMP, COMMA), dump_exit_bits = (DUMP_BIT(yydebug) | DUMP_BIT(syntax) | - DUMP_BIT(parsetree) | DUMP_BIT(parsetree_with_comment) | - DUMP_BIT(insns)) + DUMP_BIT(parsetree) | DUMP_BIT(insns)), + dump_optional_bits = (DUMP_BIT(opt_error_tolerant) | + DUMP_BIT(opt_comment) | + DUMP_BIT(opt_optimize)) }; -typedef struct ruby_cmdline_options ruby_cmdline_options_t; - -typedef struct { - unsigned int mask; - unsigned int set; -} ruby_features_t; - static inline void rb_feature_set_to(ruby_features_t *feat, unsigned int bit_mask, unsigned int bit_set) { @@ -170,37 +177,9 @@ rb_feature_set_to(ruby_features_t *feat, unsigned int bit_mask, unsigned int bit rb_feature_set_to(&(feat), bit_mask, bit_set) #define FEATURE_SET(feat, bits) FEATURE_SET_TO(feat, bits, bits) #define FEATURE_SET_RESTORE(feat, save) FEATURE_SET_TO(feat, (save).mask, (save).set & (save).mask) -#define FEATURE_SET_P(feat, bits) ((feat).set & (bits)) - -struct ruby_cmdline_options { - const char *script; - VALUE script_name; - VALUE e_script; - struct { - struct { - VALUE name; - int index; - } enc; - } src, ext, intern; - VALUE req_list; - ruby_features_t features; - ruby_features_t warn; - unsigned int dump; -#if USE_MJIT - struct mjit_options mjit; -#endif - struct rb_yjit_options yjit; - - int sflag, xflag; - unsigned int warning: 1; - unsigned int verbose: 1; - unsigned int do_loop: 1; - unsigned int do_print: 1; - unsigned int do_line: 1; - unsigned int do_split: 1; - unsigned int do_search: 1; - unsigned int setids: 2; -}; +#define FEATURE_SET_P(feat, bits) ((feat).set & FEATURE_BIT(bits)) +#define FEATURE_USED_P(feat, bits) ((feat).mask & FEATURE_BIT(bits)) +#define FEATURE_SET_BITS(feat) ((feat).set & (feat).mask) static void init_ids(ruby_cmdline_options_t *); @@ -208,21 +187,24 @@ static void init_ids(ruby_cmdline_options_t *); enum { COMPILATION_FEATURES = ( - 0 - | FEATURE_BIT(frozen_string_literal) - | FEATURE_BIT(debug_frozen_string_literal) - ), + 0 + | FEATURE_BIT(frozen_string_literal) + | FEATURE_BIT(debug_frozen_string_literal) + ), DEFAULT_FEATURES = ( - (FEATURE_BIT(debug_flag_first)-1) + (FEATURE_BIT(debug_flag_first)-1) #if DISABLE_RUBYGEMS - & ~FEATURE_BIT(gems) + & ~FEATURE_BIT(gems) #endif - & ~FEATURE_BIT(frozen_string_literal) - & ~FEATURE_BIT(jit) - & ~FEATURE_BIT(yjit) - ) + & ~FEATURE_BIT(frozen_string_literal) + & ~feature_jit_mask + ) }; +#define BACKTRACE_LENGTH_LIMIT_VALID_P(n) ((n) >= -1) +#define OPT_BACKTRACE_LENGTH_LIMIT_VALID_P(opt) \ + BACKTRACE_LENGTH_LIMIT_VALID_P((opt)->backtrace_length_limit) + static ruby_cmdline_options_t * cmdline_options_init(ruby_cmdline_options_t *opt) { @@ -232,21 +214,19 @@ cmdline_options_init(ruby_cmdline_options_t *opt) opt->ext.enc.index = -1; opt->intern.enc.index = -1; opt->features.set = DEFAULT_FEATURES; -#ifdef MJIT_FORCE_ENABLE /* to use with: ./configure cppflags="-DMJIT_FORCE_ENABLE" */ - opt->features.set |= FEATURE_BIT(jit); +#ifdef RJIT_FORCE_ENABLE /* to use with: ./configure cppflags="-DRJIT_FORCE_ENABLE" */ + opt->features.set |= FEATURE_BIT(rjit); #elif defined(YJIT_FORCE_ENABLE) opt->features.set |= FEATURE_BIT(yjit); #endif - - if (getenv("RUBY_YJIT_ENABLE")) { - opt->features.set |= FEATURE_BIT(yjit); - } + opt->dump |= DUMP_BIT(opt_optimize); + opt->backtrace_length_limit = LONG_MIN; return opt; } static rb_ast_t *load_file(VALUE parser, VALUE fname, VALUE f, int script, - ruby_cmdline_options_t *opt); + ruby_cmdline_options_t *opt); static VALUE open_load_file(VALUE fname_v, int *xflag); static void forbid_setid(const char *, const ruby_cmdline_options_t *); #define forbid_setid(s) forbid_setid((s), opt) @@ -260,214 +240,254 @@ static const char esc_standout[] = "\n\033[1;7m"; static const char esc_bold[] = "\033[1m"; static const char esc_reset[] = "\033[0m"; static const char esc_none[] = ""; +#define USAGE_INDENT " " /* macro for concatenation */ static void -show_usage_line(const char *str, unsigned int namelen, unsigned int secondlen, int help, int highlight, unsigned int w) +show_usage_part(const char *str, const unsigned int namelen, + const char *str2, const unsigned int secondlen, + const char *desc, + int help, int highlight, unsigned int w, int columns) { + static const int indent_width = (int)rb_strlen_lit(USAGE_INDENT); const char *sb = highlight ? esc_bold : esc_none; const char *se = highlight ? esc_reset : esc_none; - const int wrap = help && namelen + secondlen - 1 > w; - printf(" %s%.*s%-*.*s%s%-*s%s\n", sb, namelen-1, str, - (wrap ? 0 : w - namelen + 1), - (help ? secondlen-1 : 0), str + namelen, se, - (wrap ? w + 3 : 0), (wrap ? "\n" : ""), - str + namelen + secondlen); + unsigned int desclen = (unsigned int)strcspn(desc, "\n"); + if (!help && desclen > 0 && strchr(".;:", desc[desclen-1])) --desclen; + if (help && (namelen + 1 > w) && /* a padding space */ + (int)(namelen + secondlen + indent_width) >= columns) { + printf(USAGE_INDENT "%s" "%.*s" "%s\n", sb, namelen, str, se); + if (secondlen > 0) { + const int second_end = secondlen; + int n = 0; + if (str2[n] == ',') n++; + if (str2[n] == ' ') n++; + printf(USAGE_INDENT "%s" "%.*s" "%s\n", sb, second_end-n, str2+n, se); + } + printf("%-*s%.*s\n", w + indent_width, USAGE_INDENT, desclen, desc); + } + else { + const int wrap = help && namelen + secondlen >= w; + printf(USAGE_INDENT "%s%.*s%-*.*s%s%-*s%.*s\n", sb, namelen, str, + (wrap ? 0 : w - namelen), + (help ? secondlen : 0), str2, se, + (wrap ? (int)(w + rb_strlen_lit("\n" USAGE_INDENT)) : 0), + (wrap ? "\n" USAGE_INDENT : ""), + desclen, desc); + } + if (help) { + while (desc[desclen]) { + desc += desclen + rb_strlen_lit("\n"); + desclen = (unsigned int)strcspn(desc, "\n"); + printf("%-*s%.*s\n", w + indent_width, USAGE_INDENT, desclen, desc); + } + } +} + +static void +show_usage_line(const struct ruby_opt_message *m, + int help, int highlight, unsigned int w, int columns) +{ + const char *str = m->str; + const unsigned int namelen = m->namelen, secondlen = m->secondlen; + const char *desc = str + namelen + secondlen; + show_usage_part(str, namelen - 1, str + namelen, secondlen - 1, desc, + help, highlight, w, columns); +} + +void +ruby_show_usage_line(const char *name, const char *secondary, const char *description, + int help, int highlight, unsigned int width, int columns) +{ + unsigned int namelen = (unsigned int)strlen(name); + unsigned int secondlen = (secondary ? (unsigned int)strlen(secondary) : 0); + show_usage_part(name, namelen, secondary, secondlen, + description, help, highlight, width, columns); } static void usage(const char *name, int help, int highlight, int columns) { +#define M(shortopt, longopt, desc) RUBY_OPT_MESSAGE(shortopt, longopt, desc) + +#if USE_YJIT +# define PLATFORM_JIT_OPTION "--yjit" +#else +# define PLATFORM_JIT_OPTION "--rjit (experimental)" +#endif + /* This message really ought to be max 23 lines. * Removed -h because the user already knows that option. Others? */ - - struct message { - const char *str; - unsigned short namelen, secondlen; - }; -#define M(shortopt, longopt, desc) { \ - shortopt " " longopt " " desc, \ - (unsigned short)sizeof(shortopt), \ - (unsigned short)sizeof(longopt), \ -} - static const struct message usage_msg[] = { - M("-0[octal]", "", "specify record separator (\\0, if no argument)"), - M("-a", "", "autosplit mode with -n or -p (splits $_ into $F)"), - M("-c", "", "check syntax only"), - M("-Cdirectory", "", "cd to directory before executing your script"), - M("-d", ", --debug", "set debugging flags (set $DEBUG to true)"), - M("-e 'command'", "", "one line of script. Several -e's allowed. Omit [programfile]"), - M("-Eex[:in]", ", --encoding=ex[:in]", "specify the default external and internal character encodings"), - M("-Fpattern", "", "split() pattern for autosplit (-a)"), - M("-i[extension]", "", "edit ARGV files in place (make backup if extension supplied)"), - M("-Idirectory", "", "specify $LOAD_PATH directory (may be used more than once)"), - M("-l", "", "enable line ending processing"), - M("-n", "", "assume 'while gets(); ... end' loop around your script"), - M("-p", "", "assume loop like -n but print line also like sed"), - M("-rlibrary", "", "require the library before executing your script"), - M("-s", "", "enable some switch parsing for switches after script name"), - M("-S", "", "look for the script using PATH environment variable"), - M("-v", "", "print the version number, then turn on verbose mode"), - M("-w", "", "turn warnings on for your script"), - M("-W[level=2|:category]", "", "set warning level; 0=silence, 1=medium, 2=verbose"), - M("-x[directory]", "", "strip off text before #!ruby line and perhaps cd to directory"), - M("--jit", "", "enable JIT with default options (experimental)"), - M("--jit-[option]","", "enable JIT with an option (experimental)"), - M("--yjit", "", "enable in-process JIT compiler (experimental)"), - M("-h", "", "show this message, --help for more info"), - }; - static const struct message help_msg[] = { - M("--copyright", "", "print the copyright"), - M("--dump={insns|parsetree|...}[,...]", "", - "dump debug information. see below for available dump list"), - M("--enable={jit|rubyopt|...}[,...]", ", --disable={jit|rubyopt|...}[,...]", - "enable or disable features. see below for available features"), - M("--external-encoding=encoding", ", --internal-encoding=encoding", - "specify the default external or internal character encoding"), - M("--backtrace-limit=num", "", "limit the maximum length of backtrace"), - M("--verbose", "", "turn on verbose mode and disable script from stdin"), - M("--version", "", "print the version number, then exit"), - M("--help", "", "show this message, -h for short message"), - }; - static const struct message dumps[] = { - M("insns", "", "instruction sequences"), - M("yydebug", "", "yydebug of yacc parser generator"), - M("parsetree", "", "AST"), - M("parsetree_with_comment", "", "AST with comments"), + static const struct ruby_opt_message usage_msg[] = { + M("-0[octal]", "", "Set input record separator ($/):\n" + "-0 for \\0; -00 for paragraph mode; -0777 for slurp mode."), + M("-a", "", "Split each input line ($_) into fields ($F)."), + M("-c", "", "Check syntax (no execution)."), + M("-Cdirpath", "", "Execute program in specified directory."), + M("-d", ", --debug", "Set debugging flag ($DEBUG) to true."), + M("-e 'code'", "", "Execute given Ruby code; multiple -e allowed."), + M("-Eex[:in]", ", --encoding=ex[:in]", "Set default external and internal encodings."), + M("-Fpattern", "", "Set input field separator ($;); used with -a."), + M("-i[extension]", "", "Set ARGF in-place mode;\n" + "create backup files with given extension."), + M("-Idirpath", "", "Add specified directory to load paths ($LOAD_PATH);\n" + "multiple -I allowed."), + M("-l", "", "Set output record separator ($\\) to $/;\n" + "used for line-oriented output."), + M("-n", "", "Run program in gets loop."), + M("-p", "", "Like -n, with printing added."), + M("-rlibrary", "", "Require the given library."), + M("-s", "", "Define global variables using switches following program path."), + M("-S", "", "Search directories found in the PATH environment variable."), + M("-v", "", "Print version; set $VERBOSE to true."), + M("-w", "", "Synonym for -W1."), + M("-W[level=2|:category]", "", "Set warning flag ($-W):\n" + "0 for silent; 1 for moderate; 2 for verbose."), + M("-x[dirpath]", "", "Execute Ruby code starting from a #!ruby line."), + M("--jit", "", "Enable JIT for the platform; same as " PLATFORM_JIT_OPTION "."), +#if USE_YJIT + M("--yjit", "", "Enable in-process JIT compiler."), +#endif +#if USE_RJIT + M("--rjit", "", "Enable pure-Ruby JIT compiler (experimental)."), +#endif + M("-h", "", "Print this help message; use --help for longer message."), }; - static const struct message features[] = { - M("gems", "", "rubygems (only for debugging, default: "DEFAULT_RUBYGEMS_ENABLED")"), - M("error_highlight", "", "error_highlight (default: "DEFAULT_RUBYGEMS_ENABLED")"), - M("did_you_mean", "", "did_you_mean (default: "DEFAULT_RUBYGEMS_ENABLED")"), - M("rubyopt", "", "RUBYOPT environment variable (default: enabled)"), - M("frozen-string-literal", "", "freeze all string literals (default: disabled)"), - M("jit", "", "JIT compiler (default: disabled)"), - M("yjit", "", "in-process JIT compiler (default: disabled)"), + STATIC_ASSERT(usage_msg_size, numberof(usage_msg) < 25); + + static const struct ruby_opt_message help_msg[] = { + M("--backtrace-limit=num", "", "Set backtrace limit."), + M("--copyright", "", "Print Ruby copyright."), + M("--crash-report=template", "", "Set template for crash report file."), + M("--disable=features", "", "Disable features; see list below."), + M("--dump=items", "", "Dump items; see list below."), + M("--enable=features", "", "Enable features; see list below."), + M("--external-encoding=encoding", "", "Set default external encoding."), + M("--help", "", "Print long help message; use -h for short message."), + M("--internal-encoding=encoding", "", "Set default internal encoding."), + M("--parser=parser", "", "Set Ruby parser: parse.y or prism."), + M("--verbose", "", "Set $VERBOSE to true; ignore input from $stdin."), + M("--version", "", "Print Ruby version."), + M("-y", ", --yydebug", "Print parser log; backward compatibility not guaranteed."), }; - static const struct message warn_categories[] = { - M("deprecated", "", "deprecated features"), - M("experimental", "", "experimental features"), + static const struct ruby_opt_message dumps[] = { + M("insns", "", "Instruction sequences."), + M("yydebug", "", "yydebug of yacc parser generator."), + M("parsetree", "", "Abstract syntax tree (AST)."), + M("-optimize", "", "Disable optimization (affects insns)."), + M("+error-tolerant", "", "Error-tolerant parsing (affects yydebug, parsetree)."), + M("+comment", "", "Add comments to AST (affects parsetree)."), }; - static const struct message mjit_options[] = { - M("--jit-warnings", "", "Enable printing JIT warnings"), - M("--jit-debug", "", "Enable JIT debugging (very slow), or add cflags if specified"), - M("--jit-wait", "", "Wait until JIT compilation finishes every time (for testing)"), - M("--jit-save-temps", "", "Save JIT temporary files in $TMP or /tmp (for testing)"), - M("--jit-verbose=num", "", "Print JIT logs of level num or less to stderr (default: 0)"), - M("--jit-max-cache=num", "", "Max number of methods to be JIT-ed in a cache (default: 100)"), - M("--jit-min-calls=num", "", "Number of calls to trigger JIT (for testing, default: 10000)"), + static const struct ruby_opt_message features[] = { + M("gems", "", "Rubygems (only for debugging, default: "DEFAULT_RUBYGEMS_ENABLED")."), + M("error_highlight", "", "error_highlight (default: "DEFAULT_RUBYGEMS_ENABLED")."), + M("did_you_mean", "", "did_you_mean (default: "DEFAULT_RUBYGEMS_ENABLED")."), + M("syntax_suggest", "", "syntax_suggest (default: "DEFAULT_RUBYGEMS_ENABLED")."), + M("rubyopt", "", "RUBYOPT environment variable (default: enabled)."), + M("frozen-string-literal", "", "Freeze all string literals (default: disabled)."), +#if USE_YJIT + M("yjit", "", "In-process JIT compiler (default: disabled)."), +#endif +#if USE_RJIT + M("rjit", "", "Pure-Ruby JIT compiler (experimental, default: disabled)."), +#endif }; - static const struct message yjit_options[] = { -#if YJIT_STATS - M("--yjit-stats", "", "Enable collecting YJIT statistics"), -#endif - M("--yjit-exec-mem-size=num", "", "Size of executable memory block in MiB (default: 256)"), - M("--yjit-call-threshold", "", "Number of calls to trigger JIT (default: 10)"), - M("--yjit-max-versions", "", "Maximum number of versions per basic block (default: 4)"), - M("--yjit-greedy-versioning", "", "Greedy versioning mode (default: disabled)"), + static const struct ruby_opt_message warn_categories[] = { + M("deprecated", "", "Deprecated features."), + M("experimental", "", "Experimental features."), + M("performance", "", "Performance issues."), }; +#if USE_RJIT + extern const struct ruby_opt_message rb_rjit_option_messages[]; +#endif int i; const char *sb = highlight ? esc_standout+1 : esc_none; const char *se = highlight ? esc_reset : esc_none; const int num = numberof(usage_msg) - (help ? 1 : 0); unsigned int w = (columns > 80 ? (columns - 79) / 2 : 0) + 16; -#define SHOW(m) show_usage_line((m).str, (m).namelen, (m).secondlen, help, highlight, w) +#define SHOW(m) show_usage_line(&(m), help, highlight, w, columns) - printf("%sUsage:%s %s [switches] [--] [programfile] [arguments]\n", sb, se, name); + printf("%sUsage:%s %s [options] [--] [filepath] [arguments]\n", sb, se, name); for (i = 0; i < num; ++i) - SHOW(usage_msg[i]); + SHOW(usage_msg[i]); if (!help) return; if (highlight) sb = esc_standout; for (i = 0; i < numberof(help_msg); ++i) - SHOW(help_msg[i]); + SHOW(help_msg[i]); printf("%s""Dump List:%s\n", sb, se); for (i = 0; i < numberof(dumps); ++i) - SHOW(dumps[i]); + SHOW(dumps[i]); printf("%s""Features:%s\n", sb, se); for (i = 0; i < numberof(features); ++i) - SHOW(features[i]); + SHOW(features[i]); printf("%s""Warning categories:%s\n", sb, se); for (i = 0; i < numberof(warn_categories); ++i) - SHOW(warn_categories[i]); - printf("%s""MJIT options (experimental):%s\n", sb, se); - for (i = 0; i < numberof(mjit_options); ++i) - SHOW(mjit_options[i]); - printf("%s""YJIT options (experimental):%s\n", sb, se); - for (i = 0; i < numberof(yjit_options); ++i) - SHOW(yjit_options[i]); + SHOW(warn_categories[i]); +#if USE_YJIT + printf("%s""YJIT options:%s\n", sb, se); + rb_yjit_show_usage(help, highlight, w, columns); +#endif +#if USE_RJIT + printf("%s""RJIT options (experimental):%s\n", sb, se); + for (i = 0; rb_rjit_option_messages[i].str; ++i) + SHOW(rb_rjit_option_messages[i]); +#endif } #define rubylib_path_new rb_str_new static void -push_include(const char *path, VALUE (*filter)(VALUE)) +ruby_push_include(const char *path, VALUE (*filter)(VALUE)) { const char sep = PATH_SEP_CHAR; const char *p, *s; VALUE load_path = GET_VM()->load_path; - - p = path; - while (*p) { - while (*p == sep) - p++; - if (!*p) break; - for (s = p; *s && *s != sep; s = CharNext(s)); - rb_ary_push(load_path, (*filter)(rubylib_path_new(p, s - p))); - p = s; - } -} - #ifdef __CYGWIN__ -static void -push_include_cygwin(const char *path, VALUE (*filter)(VALUE)) -{ - const char *p, *s; char rubylib[FILENAME_MAX]; VALUE buf = 0; +# define is_path_sep(c) ((c) == sep || (c) == ';') +#else +# define is_path_sep(c) ((c) == sep) +#endif + if (path == 0) return; p = path; while (*p) { - unsigned int len; - while (*p == ';') - p++; - if (!*p) break; - for (s = p; *s && *s != ';'; s = CharNext(s)); - len = s - p; - if (*s) { - if (!buf) { - buf = rb_str_new(p, len); - p = RSTRING_PTR(buf); - } - else { - rb_str_resize(buf, len); - p = strncpy(RSTRING_PTR(buf), p, len); - } - } + long len; + while (is_path_sep(*p)) + p++; + if (!*p) break; + for (s = p; *s && !is_path_sep(*s); s = CharNext(s)); + len = s - p; +#undef is_path_sep + +#ifdef __CYGWIN__ + if (*s) { + if (!buf) { + buf = rb_str_new(p, len); + p = RSTRING_PTR(buf); + } + else { + rb_str_resize(buf, len); + p = strncpy(RSTRING_PTR(buf), p, len); + } + } #ifdef HAVE_CYGWIN_CONV_PATH #define CONV_TO_POSIX_PATH(p, lib) \ - cygwin_conv_path(CCP_WIN_A_TO_POSIX|CCP_RELATIVE, (p), (lib), sizeof(lib)) + cygwin_conv_path(CCP_WIN_A_TO_POSIX|CCP_RELATIVE, (p), (lib), sizeof(lib)) #else # error no cygwin_conv_path #endif - if (CONV_TO_POSIX_PATH(p, rubylib) == 0) - p = rubylib; - push_include(p, filter); - if (!*s) break; - p = s + 1; - } -} - -#define push_include push_include_cygwin + if (CONV_TO_POSIX_PATH(p, rubylib) == 0) { + p = rubylib; + len = strlen(p); + } #endif - -void -ruby_push_include(const char *path, VALUE (*filter)(VALUE)) -{ - if (path == 0) - return; - push_include(path, filter); + rb_ary_push(load_path, (*filter)(rubylib_path_new(p, len))); + p = s; + } } static VALUE @@ -475,6 +495,7 @@ identical_path(VALUE path) { return path; } + static VALUE locale_path(VALUE path) { @@ -493,9 +514,9 @@ expand_include_path(VALUE path) { char *p = RSTRING_PTR(path); if (!p) - return path; + return path; if (*p == '.' && p[1] == '/') - return path; + return path; return rb_file_expand_path(path, Qnil); } @@ -513,7 +534,7 @@ BOOL WINAPI DllMain(HINSTANCE dll, DWORD reason, LPVOID reserved) { if (reason == DLL_PROCESS_ATTACH) - libruby = dll; + libruby = dll; return TRUE; } @@ -527,9 +548,9 @@ static inline void translit_char_bin(char *p, int from, int to) { while (*p) { - if ((unsigned char)*p == from) - *p = to; - p++; + if ((unsigned char)*p == from) + *p = to; + p++; } } #endif @@ -552,8 +573,8 @@ static VALUE str_conv_enc(VALUE str, rb_encoding *from, rb_encoding *to) { return rb_str_conv_enc_opts(str, from, to, - ECONV_UNDEF_REPLACE|ECONV_INVALID_REPLACE, - Qnil); + ECONV_UNDEF_REPLACE|ECONV_INVALID_REPLACE, + Qnil); } #else # define str_conv_enc(str, from, to) (str) @@ -566,45 +587,41 @@ static VALUE runtime_libruby_path(void) { #if defined _WIN32 || defined __CYGWIN__ - DWORD len, ret; -#if USE_RVARGC - len = 32; -#else - len = RSTRING_EMBED_LEN_MAX; -#endif + DWORD ret; + DWORD len = 32; VALUE path; VALUE wsopath = rb_str_new(0, len*sizeof(WCHAR)); WCHAR *wlibpath; char *libpath; while (wlibpath = (WCHAR *)RSTRING_PTR(wsopath), - ret = GetModuleFileNameW(libruby, wlibpath, len), - (ret == len)) + ret = GetModuleFileNameW(libruby, wlibpath, len), + (ret == len)) { - rb_str_modify_expand(wsopath, len*sizeof(WCHAR)); - rb_str_set_len(wsopath, (len += len)*sizeof(WCHAR)); + rb_str_modify_expand(wsopath, len*sizeof(WCHAR)); + rb_str_set_len(wsopath, (len += len)*sizeof(WCHAR)); } if (!ret || ret > len) rb_fatal("failed to get module file name"); #if defined __CYGWIN__ { - const int win_to_posix = CCP_WIN_W_TO_POSIX | CCP_RELATIVE; - size_t newsize = cygwin_conv_path(win_to_posix, wlibpath, 0, 0); - if (!newsize) rb_fatal("failed to convert module path to cygwin"); - path = rb_str_new(0, newsize); - libpath = RSTRING_PTR(path); - if (cygwin_conv_path(win_to_posix, wlibpath, libpath, newsize)) { - rb_str_resize(path, 0); - } + const int win_to_posix = CCP_WIN_W_TO_POSIX | CCP_RELATIVE; + size_t newsize = cygwin_conv_path(win_to_posix, wlibpath, 0, 0); + if (!newsize) rb_fatal("failed to convert module path to cygwin"); + path = rb_str_new(0, newsize); + libpath = RSTRING_PTR(path); + if (cygwin_conv_path(win_to_posix, wlibpath, libpath, newsize)) { + rb_str_resize(path, 0); + } } #else { - DWORD i; - for (len = ret, i = 0; i < len; ++i) { - if (wlibpath[i] == L'\\') { - wlibpath[i] = L'/'; - ret = i+1; /* chop after the last separator */ - } - } + DWORD i; + for (len = ret, i = 0; i < len; ++i) { + if (wlibpath[i] == L'\\') { + wlibpath[i] = L'/'; + ret = i+1; /* chop after the last separator */ + } + } } len = WideCharToMultiByte(CP_UTF8, 0, wlibpath, ret, NULL, 0, NULL, NULL); path = rb_utf8_str_new(0, len); @@ -619,17 +636,17 @@ runtime_libruby_path(void) const void* addr = (void *)(VALUE)expand_include_path; if (!dladdr((void *)addr, &dli)) { - return rb_str_new(0, 0); + return rb_str_new(0, 0); } #ifdef __linux__ else if (origarg.argc > 0 && origarg.argv && dli.dli_fname == origarg.argv[0]) { - fname = rb_str_new_cstr("/proc/self/exe"); - path = rb_readlink(fname, NULL); + fname = rb_str_new_cstr("/proc/self/exe"); + path = rb_readlink(fname, NULL); } #endif else { - fname = rb_str_new_cstr(dli.dli_fname); - path = rb_realpath_internal(Qnil, fname, 1); + fname = rb_str_new_cstr(dli.dli_fname); + path = rb_realpath_internal(Qnil, fname, 1); } rb_str_resize(fname, 0); return path; @@ -649,6 +666,7 @@ ruby_init_loadpath(void) VALUE load_path, archlibdir = 0; ID id_initial_load_path_mark; const char *paths = ruby_initial_load_paths; + #if defined LOAD_RELATIVE #if !defined ENABLE_MULTIARCH # define RUBY_ARCH_PATH "" @@ -667,44 +685,44 @@ ruby_init_loadpath(void) p = strrchr(libpath, '/'); if (p) { - static const char libdir[] = "/" + static const char libdir[] = "/" #ifdef LIBDIR_BASENAME - LIBDIR_BASENAME + LIBDIR_BASENAME #else - "lib" + "lib" #endif - RUBY_ARCH_PATH; - const ptrdiff_t libdir_len = (ptrdiff_t)sizeof(libdir) - - rb_strlen_lit(RUBY_ARCH_PATH) - 1; - static const char bindir[] = "/bin"; - const ptrdiff_t bindir_len = (ptrdiff_t)sizeof(bindir) - 1; + RUBY_ARCH_PATH; + const ptrdiff_t libdir_len = (ptrdiff_t)sizeof(libdir) + - rb_strlen_lit(RUBY_ARCH_PATH) - 1; + static const char bindir[] = "/bin"; + const ptrdiff_t bindir_len = (ptrdiff_t)sizeof(bindir) - 1; - const char *p2 = NULL; + const char *p2 = NULL; #ifdef ENABLE_MULTIARCH multiarch: #endif - if (p - libpath >= bindir_len && !STRNCASECMP(p - bindir_len, bindir, bindir_len)) { - p -= bindir_len; - archlibdir = rb_str_subseq(sopath, 0, p - libpath); - rb_str_cat_cstr(archlibdir, libdir); - OBJ_FREEZE_RAW(archlibdir); - } - else if (p - libpath >= libdir_len && !strncmp(p - libdir_len, libdir, libdir_len)) { - archlibdir = rb_str_subseq(sopath, 0, (p2 ? p2 : p) - libpath); - OBJ_FREEZE_RAW(archlibdir); - p -= libdir_len; - } + if (p - libpath >= bindir_len && !STRNCASECMP(p - bindir_len, bindir, bindir_len)) { + p -= bindir_len; + archlibdir = rb_str_subseq(sopath, 0, p - libpath); + rb_str_cat_cstr(archlibdir, libdir); + OBJ_FREEZE(archlibdir); + } + else if (p - libpath >= libdir_len && !strncmp(p - libdir_len, libdir, libdir_len)) { + archlibdir = rb_str_subseq(sopath, 0, (p2 ? p2 : p) - libpath); + OBJ_FREEZE(archlibdir); + p -= libdir_len; + } #ifdef ENABLE_MULTIARCH - else if (p2) { - p = p2; - } - else { - p2 = p; - p = rb_enc_path_last_separator(libpath, p, rb_ascii8bit_encoding()); - if (p) goto multiarch; - p = p2; - } + else if (p2) { + p = p2; + } + else { + p2 = p; + p = rb_enc_path_last_separator(libpath, p, rb_ascii8bit_encoding()); + if (p) goto multiarch; + p = p2; + } #endif baselen = p - libpath; } @@ -723,7 +741,7 @@ ruby_init_loadpath(void) #endif rb_gc_register_address(&ruby_prefix_path); ruby_prefix_path = PREFIX_PATH(); - OBJ_FREEZE_RAW(ruby_prefix_path); + OBJ_FREEZE(ruby_prefix_path); if (!archlibdir) archlibdir = ruby_prefix_path; rb_gc_register_address(&ruby_archlibdir_path); ruby_archlibdir_path = archlibdir; @@ -734,11 +752,11 @@ ruby_init_loadpath(void) id_initial_load_path_mark = INITIAL_LOAD_PATH_MARK; while (*paths) { - size_t len = strlen(paths); - VALUE path = RUBY_RELATIVE(paths, len); - rb_ivar_set(path, id_initial_load_path_mark, path); - rb_ary_push(load_path, path); - paths += len + 1; + size_t len = strlen(paths); + VALUE path = RUBY_RELATIVE(paths, len); + rb_ivar_set(path, id_initial_load_path_mark, path); + rb_ary_push(load_path, path); + paths += len + 1; } rb_const_set(rb_cObject, rb_intern_const("TMP_RUBY_PREFIX"), ruby_prefix_path); @@ -752,7 +770,7 @@ add_modules(VALUE *req_list, const char *mod) VALUE feature; if (!list) { - *req_list = list = rb_ary_tmp_new(0); + *req_list = list = rb_ary_hidden_new(0); } feature = rb_str_cat_cstr(rb_str_tmp_new(0), mod); rb_ary_push(list, feature); @@ -768,11 +786,11 @@ require_libraries(VALUE *req_list) CONST_ID(require, "require"); while (list && RARRAY_LEN(list) > 0) { - VALUE feature = rb_ary_shift(list); - rb_enc_associate(feature, extenc); - RBASIC_SET_CLASS_RAW(feature, rb_cString); - OBJ_FREEZE(feature); - rb_funcallv(self, require, 1, &feature); + VALUE feature = rb_ary_shift(list); + rb_enc_associate(feature, extenc); + RBASIC_SET_CLASS_RAW(feature, rb_cString); + OBJ_FREEZE(feature); + rb_funcallv(self, require, 1, &feature); } *req_list = 0; } @@ -783,68 +801,69 @@ toplevel_context(rb_binding_t *bind) return &bind->block; } -static void -process_sflag(int *sflag) -{ - if (*sflag > 0) { - long n; - const VALUE *args; - VALUE argv = rb_argv; - - n = RARRAY_LEN(argv); - args = RARRAY_CONST_PTR(argv); - while (n > 0) { - VALUE v = *args++; - char *s = StringValuePtr(v); - char *p; - int hyphen = FALSE; - - if (s[0] != '-') - break; - n--; - if (s[1] == '-' && s[2] == '\0') - break; - - v = Qtrue; - /* check if valid name before replacing - with _ */ - for (p = s + 1; *p; p++) { - if (*p == '=') { - *p++ = '\0'; - v = rb_str_new2(p); - break; - } - if (*p == '-') { - hyphen = TRUE; - } - else if (*p != '_' && !ISALNUM(*p)) { - VALUE name_error[2]; - name_error[0] = - rb_str_new2("invalid name for global variable - "); - if (!(p = strchr(p, '='))) { - rb_str_cat2(name_error[0], s); - } - else { - rb_str_cat(name_error[0], s, p - s); - } - name_error[1] = args[-1]; - rb_exc_raise(rb_class_new_instance(2, name_error, rb_eNameError)); - } - } - s[0] = '$'; - if (hyphen) { - for (p = s + 1; *p; ++p) { - if (*p == '-') - *p = '_'; - } - } - rb_gv_set(s, v); - } - n = RARRAY_LEN(argv) - n; - while (n--) { - rb_ary_shift(argv); - } - *sflag = -1; +static int +process_sflag(int sflag) +{ + if (sflag > 0) { + long n; + const VALUE *args; + VALUE argv = rb_argv; + + n = RARRAY_LEN(argv); + args = RARRAY_CONST_PTR(argv); + while (n > 0) { + VALUE v = *args++; + char *s = StringValuePtr(v); + char *p; + int hyphen = FALSE; + + if (s[0] != '-') + break; + n--; + if (s[1] == '-' && s[2] == '\0') + break; + + v = Qtrue; + /* check if valid name before replacing - with _ */ + for (p = s + 1; *p; p++) { + if (*p == '=') { + *p++ = '\0'; + v = rb_str_new2(p); + break; + } + if (*p == '-') { + hyphen = TRUE; + } + else if (*p != '_' && !ISALNUM(*p)) { + VALUE name_error[2]; + name_error[0] = + rb_str_new2("invalid name for global variable - "); + if (!(p = strchr(p, '='))) { + rb_str_cat2(name_error[0], s); + } + else { + rb_str_cat(name_error[0], s, p - s); + } + name_error[1] = args[-1]; + rb_exc_raise(rb_class_new_instance(2, name_error, rb_eNameError)); + } + } + s[0] = '$'; + if (hyphen) { + for (p = s + 1; *p; ++p) { + if (*p == '-') + *p = '_'; + } + } + rb_gv_set(s, v); + } + n = RARRAY_LEN(argv) - n; + while (n--) { + rb_ary_shift(argv); + } + return -1; } + return sflag; } static long proc_options(long argc, char **argv, ruby_cmdline_options_t *opt, int envopt); @@ -858,23 +877,35 @@ moreswitches(const char *s, ruby_cmdline_options_t *opt, int envopt) VALUE argstr, argary; void *ptr; + VALUE src_enc_name = opt->src.enc.name; + VALUE ext_enc_name = opt->ext.enc.name; + VALUE int_enc_name = opt->intern.enc.name; + ruby_features_t feat = opt->features; + ruby_features_t warn = opt->warn; + long backtrace_length_limit = opt->backtrace_length_limit; + const char *crash_report = opt->crash_report; + while (ISSPACE(*s)) s++; if (!*s) return; - argstr = rb_str_tmp_new((len = strlen(s)) + (envopt!=0)); + + opt->src.enc.name = opt->ext.enc.name = opt->intern.enc.name = 0; + + const int hyphen = *s != '-'; + argstr = rb_str_tmp_new((len = strlen(s)) + hyphen); argary = rb_str_tmp_new(0); p = RSTRING_PTR(argstr); - if (envopt) *p++ = ' '; - memcpy(p, s, len + 1); + if (hyphen) *p = '-'; + memcpy(p + hyphen, s, len + 1); ap = 0; rb_str_cat(argary, (char *)&ap, sizeof(ap)); while (*p) { - ap = p; - rb_str_cat(argary, (char *)&ap, sizeof(ap)); - while (*p && !ISSPACE(*p)) ++p; - if (!*p) break; - *p++ = '\0'; - while (ISSPACE(*p)) ++p; + ap = p; + rb_str_cat(argary, (char *)&ap, sizeof(ap)); + while (*p && !ISSPACE(*p)) ++p; + if (!*p) break; + *p++ = '\0'; + while (ISSPACE(*p)) ++p; } argc = RSTRING_LEN(argary) / sizeof(ap); ap = 0; @@ -883,14 +914,32 @@ moreswitches(const char *s, ruby_cmdline_options_t *opt, int envopt) MEMMOVE(argv, RSTRING_PTR(argary), char *, argc); while ((i = proc_options(argc, argv, opt, envopt)) > 1 && envopt && (argc -= i) > 0) { - argv += i; - if (**argv != '-') { - *--*argv = '-'; - } - if ((*argv)[1]) { - ++argc; - --argv; - } + argv += i; + if (**argv != '-') { + *--*argv = '-'; + } + if ((*argv)[1]) { + ++argc; + --argv; + } + } + + if (src_enc_name) { + opt->src.enc.name = src_enc_name; + } + if (ext_enc_name) { + opt->ext.enc.name = ext_enc_name; + } + if (int_enc_name) { + opt->intern.enc.name = int_enc_name; + } + FEATURE_SET_RESTORE(opt->features, feat); + FEATURE_SET_RESTORE(opt->warn, warn); + if (BACKTRACE_LENGTH_LIMIT_VALID_P(backtrace_length_limit)) { + opt->backtrace_length_limit = backtrace_length_limit; + } + if (crash_report) { + opt->crash_report = crash_report; } ruby_xfree(ptr); @@ -904,15 +953,17 @@ name_match_p(const char *name, const char *str, size_t len) { if (len == 0) return 0; while (1) { - while (TOLOWER(*str) == *name) { - if (!--len || !*++str) return 1; - ++name; - } - if (*str != '-' && *str != '_') return 0; - while (ISALNUM(*name)) name++; - if (*name != '-' && *name != '_') return 0; - ++name; - ++str; + while (TOLOWER(*str) == *name) { + if (!--len) return 1; + ++name; + ++str; + } + if (*str != '-' && *str != '_') return 0; + while (ISALNUM(*name)) name++; + if (*name != '-' && *name != '_') return 0; + if (!*++name) return 1; + ++str; + if (--len == 0) return 1; } } @@ -921,14 +972,14 @@ name_match_p(const char *name, const char *str, size_t len) #define UNSET_WHEN(name, bit, str, len) \ if (NAME_MATCH_P((name), (str), (len))) { \ - *(unsigned int *)arg &= ~(bit); \ - return; \ + *(unsigned int *)arg &= ~(bit); \ + return; \ } #define SET_WHEN(name, bit, str, len) \ if (NAME_MATCH_P((name), (str), (len))) { \ - *(unsigned int *)arg |= (bit); \ - return; \ + *(unsigned int *)arg |= (bit); \ + return; \ } #define LITERAL_NAME_ELEMENT(name) #name @@ -949,34 +1000,34 @@ feature_option(const char *str, int len, void *arg, const unsigned int enable) #define SET_FEATURE(bit) \ if (NAME_MATCH_P(#bit, str, len)) {set |= mask = FEATURE_BIT(bit); FEATURE_FOUND;} EACH_FEATURES(SET_FEATURE, ;); + if (NAME_MATCH_P("jit", str, len)) { // This allows you to cancel --jit + set |= mask = FEATURE_BIT(jit); + goto found; + } if (NAME_MATCH_P("all", str, len)) { - // YJIT and MJIT cannot be enabled at the same time. We enable only YJIT for --enable=all. -#ifdef MJIT_FORCE_ENABLE - mask &= ~(FEATURE_BIT(yjit)); -#else - mask &= ~(FEATURE_BIT(jit)); -#endif + // YJIT and RJIT cannot be enabled at the same time. We enable only one for --enable=all. + mask &= ~feature_jit_mask | FEATURE_BIT(jit); goto found; } #if AMBIGUOUS_FEATURE_NAMES if (matched == 1) goto found; if (matched > 1) { - VALUE mesg = rb_sprintf("ambiguous feature: `%.*s' (", len, str); + VALUE mesg = rb_sprintf("ambiguous feature: '%.*s' (", len, str); #define ADD_FEATURE_NAME(bit) \ - if (FEATURE_BIT(bit) & set) { \ - rb_str_cat_cstr(mesg, #bit); \ - if (--matched) rb_str_cat_cstr(mesg, ", "); \ - } - EACH_FEATURES(ADD_FEATURE_NAME, ;); - rb_str_cat_cstr(mesg, ")"); - rb_exc_raise(rb_exc_new_str(rb_eRuntimeError, mesg)); + if (FEATURE_BIT(bit) & set) { \ + rb_str_cat_cstr(mesg, #bit); \ + if (--matched) rb_str_cat_cstr(mesg, ", "); \ + } + EACH_FEATURES(ADD_FEATURE_NAME, ;); + rb_str_cat_cstr(mesg, ")"); + rb_exc_raise(rb_exc_new_str(rb_eRuntimeError, mesg)); #undef ADD_FEATURE_NAME } #else (void)set; #endif - rb_warn("unknown argument for --%s: `%.*s'", - enable ? "enable" : "disable", len, str); + rb_warn("unknown argument for --%s: '%.*s'", + enable ? "enable" : "disable", len, str); rb_warn("features are [%.*s].", (int)strlen(list), list); return; @@ -1014,17 +1065,84 @@ debug_option(const char *str, int len, void *arg) #ifdef RUBY_DEVEL if (ruby_patchlevel < 0 && ruby_env_debug_option(str, len, 0)) return; #endif - rb_warn("unknown argument for --debug: `%.*s'", len, str); + rb_warn("unknown argument for --debug: '%.*s'", len, str); rb_warn("debug features are [%.*s].", (int)strlen(list), list); } +static int +memtermspn(const char *str, char term, int len) +{ + RUBY_ASSERT(len >= 0); + if (len <= 0) return 0; + const char *next = memchr(str, term, len); + return next ? (int)(next - str) : len; +} + +static const char additional_opt_sep = '+'; + +static unsigned int +dump_additional_option_flag(const char *str, int len, unsigned int bits, bool set) +{ +#define SET_DUMP_OPT(bit) if (NAME_MATCH_P(#bit, str, len)) { \ + return set ? (bits | DUMP_BIT(opt_ ## bit)) : (bits & ~DUMP_BIT(opt_ ## bit)); \ + } + SET_DUMP_OPT(error_tolerant); + SET_DUMP_OPT(comment); + SET_DUMP_OPT(optimize); +#undef SET_DUMP_OPT + rb_warn("don't know how to dump with%s '%.*s'", set ? "" : "out", len, str); + return bits; +} + +static unsigned int +dump_additional_option(const char *str, int len, unsigned int bits) +{ + int w; + for (; len-- > 0 && *str++ == additional_opt_sep; len -= w, str += w) { + w = memtermspn(str, additional_opt_sep, len); + bool set = true; + if (*str == '-' || *str == '+') { + set = *str++ == '+'; + --w; + } + else { + int n = memtermspn(str, '-', w); + if (str[n] == '-') { + if (NAME_MATCH_P("with", str, n)) { + str += n; + w -= n; + } + else if (NAME_MATCH_P("without", str, n)) { + set = false; + str += n; + w -= n; + } + } + } + bits = dump_additional_option_flag(str, w, bits, set); + } + return bits; +} + static void dump_option(const char *str, int len, void *arg) { static const char list[] = EACH_DUMPS(LITERAL_NAME_ELEMENT, ", "); -#define SET_WHEN_DUMP(bit) SET_WHEN(#bit, DUMP_BIT(bit), str, len) + unsigned int *bits_ptr = (unsigned int *)arg; + if (*str == '+' || *str == '-') { + bool set = *str++ == '+'; + *bits_ptr = dump_additional_option_flag(str, --len, *bits_ptr, set); + return; + } + int w = memtermspn(str, additional_opt_sep, len); + +#define SET_WHEN_DUMP(bit) \ + if (NAME_MATCH_P(#bit "-", (str), (w))) { \ + *bits_ptr = dump_additional_option(str + w, len - w, *bits_ptr | DUMP_BIT(bit)); \ + return; \ + } EACH_DUMPS(SET_WHEN_DUMP, ;); - rb_warn("don't know how to dump `%.*s',", len, str); + rb_warn("don't know how to dump '%.*s',", len, str); rb_warn("but only [%.*s].", (int)strlen(list), list); } @@ -1037,9 +1155,9 @@ set_option_encoding_once(const char *type, VALUE *name, const char *e, long elen ename = rb_str_new(e, elen); if (*name && - rb_funcall(ename, rb_intern("casecmp"), 1, *name) != INT2FIX(0)) { - rb_raise(rb_eRuntimeError, - "%s already set to %"PRIsVALUE, type, *name); + rb_funcall(ename, rb_intern("casecmp"), 1, *name) != INT2FIX(0)) { + rb_raise(rb_eRuntimeError, + "%s already set to %"PRIsVALUE, type, *name); } *name = ename; } @@ -1051,96 +1169,346 @@ set_option_encoding_once(const char *type, VALUE *name, const char *e, long elen #define set_source_encoding_once(opt, e, elen) \ set_option_encoding_once("source", &(opt)->src.enc.name, (e), (elen)) -#define opt_match(s, l, name) \ - ((((l) > rb_strlen_lit(name)) ? (s)[rb_strlen_lit(name)] == '=' : \ - (l) == rb_strlen_lit(name)) && \ - memcmp((s), name, rb_strlen_lit(name)) == 0 && \ - (((s) += rb_strlen_lit(name)), 1)) -#define opt_match_noarg(s, l, name) \ - opt_match(s, l, name) && (*(s) ? (rb_warn("argument to --jit-" name " is ignored"), 1) : 1) -#define opt_match_arg(s, l, name) \ - opt_match(s, l, name) && (*(s) ? 1 : (rb_raise(rb_eRuntimeError, "--jit-" name " needs an argument"), 0)) - #define yjit_opt_match_noarg(s, l, name) \ opt_match(s, l, name) && (*(s) ? (rb_warn("argument to --yjit-" name " is ignored"), 1) : 1) #define yjit_opt_match_arg(s, l, name) \ opt_match(s, l, name) && (*(s) && *(s+1) ? 1 : (rb_raise(rb_eRuntimeError, "--yjit-" name " needs an argument"), 0)) +#if USE_YJIT static bool -setup_yjit_options(const char *s, struct rb_yjit_options *yjit_opt) +setup_yjit_options(const char *s) { - const char prefix[] = "yjit-"; - if (strncmp(prefix, s, sizeof(prefix)-1) != 0) { - return false; - } - s += sizeof(prefix)-1; - const size_t l = strlen(s); - if (l == 0) { - return false; - } + // The option parsing is done in yjit/src/options.rs + bool rb_yjit_parse_option(const char* s); + bool success = rb_yjit_parse_option(s); - if (yjit_opt_match_arg(s, l, "exec-mem-size")) { - yjit_opt->exec_mem_size = atoi(s + 1); + if (success) { + return true; } - else if (yjit_opt_match_arg(s, l, "call-threshold")) { - yjit_opt->call_threshold = atoi(s + 1); + + rb_raise( + rb_eRuntimeError, + "invalid YJIT option '%s' (--help will show valid yjit options)", + s + ); +} +#endif + +/* + * Following proc_*_option functions are tree kinds: + * + * - with a required argument, takes also `argc` and `argv`, and + * returns the number of consumed argv including the option itself. + * + * - with a mandatory argument just after the option. + * + * - no required argument, this returns the address of + * the next character after the last consumed character. + */ + +/* optional */ +static const char * +proc_W_option(ruby_cmdline_options_t *opt, const char *s, int *warning) +{ + if (s[1] == ':') { + unsigned int bits = 0; + static const char no_prefix[] = "no-"; + int enable = strncmp(s += 2, no_prefix, sizeof(no_prefix)-1) != 0; + if (!enable) s += sizeof(no_prefix)-1; + size_t len = strlen(s); + if (NAME_MATCH_P("deprecated", s, len)) { + bits = 1U << RB_WARN_CATEGORY_DEPRECATED; + } + else if (NAME_MATCH_P("experimental", s, len)) { + bits = 1U << RB_WARN_CATEGORY_EXPERIMENTAL; + } + else if (NAME_MATCH_P("performance", s, len)) { + bits = 1U << RB_WARN_CATEGORY_PERFORMANCE; + } + else { + rb_warn("unknown warning category: '%s'", s); + } + if (bits) FEATURE_SET_TO(opt->warn, bits, enable ? bits : 0); + return 0; } - else if (yjit_opt_match_arg(s, l, "max-versions")) { - yjit_opt->max_versions = atoi(s + 1); + else { + size_t numlen; + int v = 2; /* -W as -W2 */ + + if (*++s) { + v = scan_oct(s, 1, &numlen); + if (numlen == 0) + v = 2; + s += numlen; + } + if (!opt->warning) { + switch (v) { + case 0: + ruby_verbose = Qnil; + break; + case 1: + ruby_verbose = Qfalse; + break; + default: + ruby_verbose = Qtrue; + break; + } + } + *warning = 1; + switch (v) { + case 0: + FEATURE_SET_TO(opt->warn, RB_WARN_CATEGORY_DEFAULT_BITS, 0); + break; + case 1: + FEATURE_SET_TO(opt->warn, 1U << RB_WARN_CATEGORY_DEPRECATED, 0); + break; + default: + FEATURE_SET(opt->warn, RB_WARN_CATEGORY_DEFAULT_BITS); + break; + } + return s; } - else if (yjit_opt_match_noarg(s, l, "greedy-versioning")) { - yjit_opt->greedy_versioning = true; +} + +/* required */ +static long +proc_e_option(ruby_cmdline_options_t *opt, const char *s, long argc, char **argv) +{ + long n = 1; + forbid_setid("-e"); + if (!*++s) { + if (!--argc) + rb_raise(rb_eRuntimeError, "no code specified for -e"); + s = *++argv; + n++; } - else if (yjit_opt_match_noarg(s, l, "no-type-prop")) { - yjit_opt->no_type_prop = true; + if (!opt->e_script) { + opt->e_script = rb_str_new(0, 0); + if (opt->script == 0) + opt->script = "-e"; + } + rb_str_cat2(opt->e_script, s); + rb_str_cat2(opt->e_script, "\n"); + return n; +} + +/* optional */ +static const char * +proc_K_option(ruby_cmdline_options_t *opt, const char *s) +{ + if (*++s) { + const char *enc_name = 0; + switch (*s) { + case 'E': case 'e': + enc_name = "EUC-JP"; + break; + case 'S': case 's': + enc_name = "Windows-31J"; + break; + case 'U': case 'u': + enc_name = "UTF-8"; + break; + case 'N': case 'n': case 'A': case 'a': + enc_name = "ASCII-8BIT"; + break; + } + if (enc_name) { + opt->src.enc.name = rb_str_new2(enc_name); + if (!opt->ext.enc.name) + opt->ext.enc.name = opt->src.enc.name; + } + s++; } - else if (yjit_opt_match_noarg(s, l, "stats")) { - yjit_opt->gen_stats = true; + return s; +} + +/* optional */ +static const char * +proc_0_option(ruby_cmdline_options_t *opt, const char *s) +{ + size_t numlen; + int v; + char c; + + v = scan_oct(s, 4, &numlen); + s += numlen; + if (v > 0377) + rb_rs = Qnil; + else if (v == 0 && numlen >= 2) { + rb_rs = rb_str_new2(""); } else { - rb_raise(rb_eRuntimeError, - "invalid yjit option `%s' (--help will show valid yjit options)", s); + c = v & 0xff; + rb_rs = rb_str_new(&c, 1); } - return true; + return s; } -#if USE_MJIT +/* mandatory */ static void -setup_mjit_options(const char *s, struct mjit_options *mjit_opt) -{ - if (*s != '-') return; - const size_t l = strlen(++s); - if (*s == 0) return; - else if (opt_match_noarg(s, l, "warnings")) { - mjit_opt->warnings = 1; +proc_encoding_option(ruby_cmdline_options_t *opt, const char *s, const char *opt_name) +{ + char *p; +# define set_encoding_part(type) \ + if (!(p = strchr(s, ':'))) { \ + set_##type##_encoding_once(opt, s, 0); \ + return; \ + } \ + else if (p > s) { \ + set_##type##_encoding_once(opt, s, p-s); \ + } + set_encoding_part(external); + if (!*(s = ++p)) return; + set_encoding_part(internal); + if (!*(s = ++p)) return; +#if defined ALLOW_DEFAULT_SOURCE_ENCODING && ALLOW_DEFAULT_SOURCE_ENCODING + set_encoding_part(source); + if (!*(s = ++p)) return; +#endif + rb_raise(rb_eRuntimeError, "extra argument for %s: %s", opt_name, s); +# undef set_encoding_part + UNREACHABLE; +} + +static long +proc_long_options(ruby_cmdline_options_t *opt, const char *s, long argc, char **argv, int envopt) +{ + size_t n; + long argc0 = argc; +# define is_option_end(c, allow_hyphen) \ + (!(c) || ((allow_hyphen) && (c) == '-') || (c) == '=') +# define check_envopt(name, allow_envopt) \ + (((allow_envopt) || !envopt) ? (void)0 : \ + rb_raise(rb_eRuntimeError, "invalid switch in RUBYOPT: --" name)) +# define need_argument(name, s, needs_arg, next_arg) \ + ((*(s) ? !*++(s) : (next_arg) && (argc <= 1 || !((s) = argv[1]) || (--argc, ++argv, 0))) && (needs_arg) ? \ + rb_raise(rb_eRuntimeError, "missing argument for --" name) \ + : (void)0) +# define is_option_with_arg(name, allow_hyphen, allow_envopt) \ + is_option_with_optarg(name, allow_hyphen, allow_envopt, Qtrue, Qtrue) +# define is_option_with_optarg(name, allow_hyphen, allow_envopt, needs_arg, next_arg) \ + (strncmp((name), s, n = sizeof(name) - 1) == 0 && is_option_end(s[n], (allow_hyphen)) && \ + (s[n] != '-' || (s[n] && s[n+1])) ? \ + (check_envopt(name, (allow_envopt)), s += n, \ + need_argument(name, s, needs_arg, next_arg), 1) : 0) + + if (strcmp("copyright", s) == 0) { + if (envopt) goto noenvopt_long; + opt->dump |= DUMP_BIT(copyright); + } + else if (is_option_with_optarg("debug", Qtrue, Qtrue, Qfalse, Qfalse)) { + if (s && *s) { + ruby_each_words(s, debug_option, &opt->features); + } + else { + ruby_debug = Qtrue; + ruby_verbose = Qtrue; + } + } + else if (is_option_with_arg("enable", Qtrue, Qtrue)) { + ruby_each_words(s, enable_option, &opt->features); + } + else if (is_option_with_arg("disable", Qtrue, Qtrue)) { + ruby_each_words(s, disable_option, &opt->features); } - else if (opt_match(s, l, "debug")) { - if (*s) - mjit_opt->debug_flags = strdup(s + 1); - else - mjit_opt->debug = 1; + else if (is_option_with_arg("encoding", Qfalse, Qtrue)) { + proc_encoding_option(opt, s, "--encoding"); } - else if (opt_match_noarg(s, l, "wait")) { - mjit_opt->wait = 1; + else if (is_option_with_arg("internal-encoding", Qfalse, Qtrue)) { + set_internal_encoding_once(opt, s, 0); + } + else if (is_option_with_arg("external-encoding", Qfalse, Qtrue)) { + set_external_encoding_once(opt, s, 0); + } + else if (is_option_with_arg("parser", Qfalse, Qtrue)) { + if (strcmp("prism", s) == 0) { + (*rb_ruby_prism_ptr()) = true; + } + else if (strcmp("parse.y", s) == 0) { + // default behavior + } + else { + rb_raise(rb_eRuntimeError, "unknown parser %s", s); + } } - else if (opt_match_noarg(s, l, "save-temps")) { - mjit_opt->save_temps = 1; +#if defined ALLOW_DEFAULT_SOURCE_ENCODING && ALLOW_DEFAULT_SOURCE_ENCODING + else if (is_option_with_arg("source-encoding", Qfalse, Qtrue)) { + set_source_encoding_once(opt, s, 0); } - else if (opt_match(s, l, "verbose")) { - mjit_opt->verbose = *s ? atoi(s + 1) : 1; +#endif + else if (strcmp("version", s) == 0) { + if (envopt) goto noenvopt_long; + opt->dump |= DUMP_BIT(version); } - else if (opt_match_arg(s, l, "max-cache")) { - mjit_opt->max_cache_size = atoi(s + 1); + else if (strcmp("verbose", s) == 0) { + opt->verbose = 1; + ruby_verbose = Qtrue; } - else if (opt_match_arg(s, l, "min-calls")) { - mjit_opt->min_calls = atoi(s + 1); + else if (strcmp("jit", s) == 0) { +#if USE_YJIT || USE_RJIT + FEATURE_SET(opt->features, FEATURE_BIT(jit)); +#else + rb_warn("Ruby was built without JIT support"); +#endif + } + else if (is_option_with_optarg("rjit", '-', true, false, false)) { +#if USE_RJIT + extern void rb_rjit_setup_options(const char *s, struct rb_rjit_options *rjit_opt); + FEATURE_SET(opt->features, FEATURE_BIT(rjit)); + rb_rjit_setup_options(s, &opt->rjit); +#else + rb_warn("RJIT support is disabled."); +#endif + } + else if (is_option_with_optarg("yjit", '-', true, false, false)) { +#if USE_YJIT + FEATURE_SET(opt->features, FEATURE_BIT(yjit)); + setup_yjit_options(s); +#else + rb_warn("Ruby was built without YJIT support." + " You may need to install rustc to build Ruby with YJIT."); +#endif + } + else if (strcmp("yydebug", s) == 0) { + if (envopt) goto noenvopt_long; + opt->dump |= DUMP_BIT(yydebug); + } + else if (is_option_with_arg("dump", Qfalse, Qfalse)) { + ruby_each_words(s, dump_option, &opt->dump); + } + else if (strcmp("help", s) == 0) { + if (envopt) goto noenvopt_long; + opt->dump |= DUMP_BIT(help); + return 0; + } + else if (is_option_with_arg("backtrace-limit", Qfalse, Qtrue)) { + char *e; + long n = strtol(s, &e, 10); + if (errno == ERANGE || !BACKTRACE_LENGTH_LIMIT_VALID_P(n) || *e) { + rb_raise(rb_eRuntimeError, "wrong limit for backtrace length"); + } + else { + opt->backtrace_length_limit = n; + } + } + else if (is_option_with_arg("crash-report", true, true)) { + opt->crash_report = s; } else { rb_raise(rb_eRuntimeError, - "invalid MJIT option `%s' (--help will show valid MJIT options)", s); + "invalid option --%s (-h will show valid options)", s); } + return argc0 - argc + 1; + + noenvopt_long: + rb_raise(rb_eRuntimeError, "invalid switch in RUBYOPT: --%s", s); +# undef is_option_end +# undef check_envopt +# undef need_argument +# undef is_option_with_arg +# undef is_option_with_optarg + UNREACHABLE_RETURN(0); } -#endif static long proc_options(long argc, char **argv, ruby_cmdline_options_t *opt, int envopt) @@ -1150,433 +1518,211 @@ proc_options(long argc, char **argv, ruby_cmdline_options_t *opt, int envopt) int warning = opt->warning; if (argc <= 0 || !argv) - return 0; + return 0; for (argc--, argv++; argc > 0; argc--, argv++) { - const char *const arg = argv[0]; - if (!arg || arg[0] != '-' || !arg[1]) - break; + const char *const arg = argv[0]; + if (!arg || arg[0] != '-' || !arg[1]) + break; - s = arg + 1; + s = arg + 1; reswitch: - switch (*s) { - case 'a': - if (envopt) goto noenvopt; - opt->do_split = TRUE; - s++; - goto reswitch; - - case 'p': - if (envopt) goto noenvopt; - opt->do_print = TRUE; - /* through */ - case 'n': - if (envopt) goto noenvopt; - opt->do_loop = TRUE; - s++; - goto reswitch; - - case 'd': - ruby_debug = Qtrue; - ruby_verbose = Qtrue; - s++; - goto reswitch; - - case 'y': - if (envopt) goto noenvopt; - opt->dump |= DUMP_BIT(yydebug); - s++; - goto reswitch; - - case 'v': - if (opt->verbose) { - s++; - goto reswitch; - } - opt->dump |= DUMP_BIT(version_v); - opt->verbose = 1; - case 'w': - if (!opt->warning) { - warning = 1; - ruby_verbose = Qtrue; - } - FEATURE_SET(opt->warn, RB_WARN_CATEGORY_ALL_BITS); - s++; - goto reswitch; - - case 'W': - if (s[1] == ':') { - unsigned int bits = 0; - static const char no_prefix[] = "no-"; - int enable = strncmp(s += 2, no_prefix, sizeof(no_prefix)-1) != 0; - if (!enable) s += sizeof(no_prefix)-1; - size_t len = strlen(s); - if (NAME_MATCH_P("deprecated", s, len)) { - bits = 1U << RB_WARN_CATEGORY_DEPRECATED; - } - else if (NAME_MATCH_P("experimental", s, len)) { - bits = 1U << RB_WARN_CATEGORY_EXPERIMENTAL; - } - else { - rb_warn("unknown warning category: `%s'", s); - } - if (bits) FEATURE_SET_TO(opt->warn, bits, enable ? bits : 0); - break; + switch (*s) { + case 'a': + if (envopt) goto noenvopt; + opt->do_split = TRUE; + s++; + goto reswitch; + + case 'p': + if (envopt) goto noenvopt; + opt->do_print = TRUE; + /* through */ + case 'n': + if (envopt) goto noenvopt; + opt->do_loop = TRUE; + s++; + goto reswitch; + + case 'd': + ruby_debug = Qtrue; + ruby_verbose = Qtrue; + s++; + goto reswitch; + + case 'y': + if (envopt) goto noenvopt; + opt->dump |= DUMP_BIT(yydebug); + s++; + goto reswitch; + + case 'v': + if (opt->verbose) { + s++; + goto reswitch; } - { - size_t numlen; - int v = 2; /* -W as -W2 */ - - if (*++s) { - v = scan_oct(s, 1, &numlen); - if (numlen == 0) - v = 2; - s += numlen; - } - if (!opt->warning) { - switch (v) { - case 0: - ruby_verbose = Qnil; - break; - case 1: - ruby_verbose = Qfalse; - break; - default: - ruby_verbose = Qtrue; - break; - } - } - warning = 1; - switch (v) { - case 0: - FEATURE_SET_TO(opt->warn, RB_WARN_CATEGORY_ALL_BITS, 0); - break; - case 1: - FEATURE_SET_TO(opt->warn, 1U << RB_WARN_CATEGORY_DEPRECATED, 0); - break; - default: - FEATURE_SET(opt->warn, RB_WARN_CATEGORY_ALL_BITS); - break; - } - } - goto reswitch; - - case 'c': - if (envopt) goto noenvopt; - opt->dump |= DUMP_BIT(syntax); - s++; - goto reswitch; - - case 's': - if (envopt) goto noenvopt; - forbid_setid("-s"); - if (!opt->sflag) opt->sflag = 1; - s++; - goto reswitch; - - case 'h': - if (envopt) goto noenvopt; - opt->dump |= DUMP_BIT(usage); - goto switch_end; - - case 'l': - if (envopt) goto noenvopt; - opt->do_line = TRUE; - rb_output_rs = rb_rs; - s++; - goto reswitch; - - case 'S': - if (envopt) goto noenvopt; - forbid_setid("-S"); - opt->do_search = TRUE; - s++; - goto reswitch; - - case 'e': - if (envopt) goto noenvopt; - forbid_setid("-e"); - if (!*++s) { - if (!--argc) - rb_raise(rb_eRuntimeError, "no code specified for -e"); - s = *++argv; - } - if (!opt->e_script) { - opt->e_script = rb_str_new(0, 0); - if (opt->script == 0) - opt->script = "-e"; - } - rb_str_cat2(opt->e_script, s); - rb_str_cat2(opt->e_script, "\n"); - break; - - case 'r': - forbid_setid("-r"); - if (*++s) { - add_modules(&opt->req_list, s); - } - else if (argc > 1) { - add_modules(&opt->req_list, argv[1]); - argc--, argv++; - } - break; - - case 'i': - if (envopt) goto noenvopt; - forbid_setid("-i"); - ruby_set_inplace_mode(s + 1); - break; - - case 'x': - if (envopt) goto noenvopt; - forbid_setid("-x"); - opt->xflag = TRUE; - s++; - if (*s && chdir(s) < 0) { - rb_fatal("Can't chdir to %s", s); - } - break; - - case 'C': - case 'X': - if (envopt) goto noenvopt; - if (!*++s && (!--argc || !(s = *++argv) || !*s)) { - rb_fatal("Can't chdir"); - } - if (chdir(s) < 0) { - rb_fatal("Can't chdir to %s", s); - } - break; - - case 'F': - if (envopt) goto noenvopt; - if (*++s) { - rb_fs = rb_reg_new(s, strlen(s), 0); - } - break; - - case 'E': - if (!*++s && (!--argc || !(s = *++argv))) { - rb_raise(rb_eRuntimeError, "missing argument for -E"); - } - goto encoding; - - case 'U': - set_internal_encoding_once(opt, "UTF-8", 0); - ++s; - goto reswitch; - - case 'K': - if (*++s) { - const char *enc_name = 0; - switch (*s) { - case 'E': case 'e': - enc_name = "EUC-JP"; - break; - case 'S': case 's': - enc_name = "Windows-31J"; - break; - case 'U': case 'u': - enc_name = "UTF-8"; - break; - case 'N': case 'n': case 'A': case 'a': - enc_name = "ASCII-8BIT"; - break; - } - if (enc_name) { - opt->src.enc.name = rb_str_new2(enc_name); - if (!opt->ext.enc.name) - opt->ext.enc.name = opt->src.enc.name; - } - s++; - } - goto reswitch; - - case 'I': - forbid_setid("-I"); - if (*++s) - ruby_incpush_expand(s); - else if (argc > 1) { - ruby_incpush_expand(argv[1]); - argc--, argv++; - } - break; - - case '0': - if (envopt) goto noenvopt; - { - size_t numlen; - int v; - char c; - - v = scan_oct(s, 4, &numlen); - s += numlen; - if (v > 0377) - rb_rs = Qnil; - else if (v == 0 && numlen >= 2) { - rb_rs = rb_str_new2(""); - } - else { - c = v & 0xff; - rb_rs = rb_str_new(&c, 1); - } - } - goto reswitch; - - case '-': - if (!s[1] || (s[1] == '\r' && !s[2])) { - argc--, argv++; - goto switch_end; - } - s++; - -# define is_option_end(c, allow_hyphen) \ - (!(c) || ((allow_hyphen) && (c) == '-') || (c) == '=') -# define check_envopt(name, allow_envopt) \ - (((allow_envopt) || !envopt) ? (void)0 : \ - rb_raise(rb_eRuntimeError, "invalid switch in RUBYOPT: --" name)) -# define need_argument(name, s, needs_arg, next_arg) \ - ((*(s) ? !*++(s) : (next_arg) && (!argc || !((s) = argv[1]) || (--argc, ++argv, 0))) && (needs_arg) ? \ - rb_raise(rb_eRuntimeError, "missing argument for --" name) \ - : (void)0) -# define is_option_with_arg(name, allow_hyphen, allow_envopt) \ - is_option_with_optarg(name, allow_hyphen, allow_envopt, Qtrue, Qtrue) -# define is_option_with_optarg(name, allow_hyphen, allow_envopt, needs_arg, next_arg) \ - (strncmp((name), s, n = sizeof(name) - 1) == 0 && is_option_end(s[n], (allow_hyphen)) ? \ - (check_envopt(name, (allow_envopt)), s += n, \ - need_argument(name, s, needs_arg, next_arg), 1) : 0) - - if (strcmp("copyright", s) == 0) { - if (envopt) goto noenvopt_long; - opt->dump |= DUMP_BIT(copyright); - } - else if (is_option_with_optarg("debug", Qtrue, Qtrue, Qfalse, Qfalse)) { - if (s && *s) { - ruby_each_words(s, debug_option, &opt->features); - } - else { - ruby_debug = Qtrue; - ruby_verbose = Qtrue; - } + opt->dump |= DUMP_BIT(version_v); + opt->verbose = 1; + case 'w': + if (!opt->warning) { + warning = 1; + ruby_verbose = Qtrue; } - else if (is_option_with_arg("enable", Qtrue, Qtrue)) { - ruby_each_words(s, enable_option, &opt->features); - } - else if (is_option_with_arg("disable", Qtrue, Qtrue)) { - ruby_each_words(s, disable_option, &opt->features); - } - else if (is_option_with_arg("encoding", Qfalse, Qtrue)) { - char *p; - encoding: - do { -# define set_encoding_part(type) \ - if (!(p = strchr(s, ':'))) { \ - set_##type##_encoding_once(opt, s, 0); \ - break; \ - } \ - else if (p > s) { \ - set_##type##_encoding_once(opt, s, p-s); \ - } - set_encoding_part(external); - if (!*(s = ++p)) break; - set_encoding_part(internal); - if (!*(s = ++p)) break; -#if defined ALLOW_DEFAULT_SOURCE_ENCODING && ALLOW_DEFAULT_SOURCE_ENCODING - set_encoding_part(source); - if (!*(s = ++p)) break; -#endif - rb_raise(rb_eRuntimeError, "extra argument for %s: %s", - (arg[1] == '-' ? "--encoding" : "-E"), s); -# undef set_encoding_part - } while (0); - } - else if (is_option_with_arg("internal-encoding", Qfalse, Qtrue)) { - set_internal_encoding_once(opt, s, 0); - } - else if (is_option_with_arg("external-encoding", Qfalse, Qtrue)) { - set_external_encoding_once(opt, s, 0); - } -#if defined ALLOW_DEFAULT_SOURCE_ENCODING && ALLOW_DEFAULT_SOURCE_ENCODING - else if (is_option_with_arg("source-encoding", Qfalse, Qtrue)) { - set_source_encoding_once(opt, s, 0); - } -#endif - else if (strcmp("version", s) == 0) { - if (envopt) goto noenvopt_long; - opt->dump |= DUMP_BIT(version); - } - else if (strcmp("verbose", s) == 0) { - opt->verbose = 1; - ruby_verbose = Qtrue; - } - else if (strncmp("jit", s, 3) == 0) { -#if USE_MJIT - FEATURE_SET(opt->features, FEATURE_BIT(jit)); - setup_mjit_options(s + 3, &opt->mjit); -#else - rb_warn("MJIT support is disabled."); -#endif + FEATURE_SET(opt->warn, RB_WARN_CATEGORY_DEFAULT_BITS); + s++; + goto reswitch; + + case 'W': + if (!(s = proc_W_option(opt, s, &warning))) break; + goto reswitch; + + case 'c': + if (envopt) goto noenvopt; + opt->dump |= DUMP_BIT(syntax); + s++; + goto reswitch; + + case 's': + if (envopt) goto noenvopt; + forbid_setid("-s"); + if (!opt->sflag) opt->sflag = 1; + s++; + goto reswitch; + + case 'h': + if (envopt) goto noenvopt; + opt->dump |= DUMP_BIT(usage); + goto switch_end; + + case 'l': + if (envopt) goto noenvopt; + opt->do_line = TRUE; + rb_output_rs = rb_rs; + s++; + goto reswitch; + + case 'S': + if (envopt) goto noenvopt; + forbid_setid("-S"); + opt->do_search = TRUE; + s++; + goto reswitch; + + case 'e': + if (envopt) goto noenvopt; + if (!(n = proc_e_option(opt, s, argc, argv))) break; + --n; + argc -= n; + argv += n; + break; + + case 'r': + forbid_setid("-r"); + if (*++s) { + add_modules(&opt->req_list, s); } - else if (strcmp("yjit", s) == 0 || setup_yjit_options(s, &opt->yjit)) { -#if USE_MJIT - FEATURE_SET(opt->features, FEATURE_BIT(yjit)); -#else - rb_warn("Ruby was built without JIT support"); -#endif + else if (argc > 1) { + add_modules(&opt->req_list, argv[1]); + argc--, argv++; + } + break; + + case 'i': + if (envopt) goto noenvopt; + forbid_setid("-i"); + ruby_set_inplace_mode(s + 1); + break; + + case 'x': + if (envopt) goto noenvopt; + forbid_setid("-x"); + opt->xflag = TRUE; + s++; + if (*s && chdir(s) < 0) { + rb_fatal("Can't chdir to %s", s); + } + break; + + case 'C': + case 'X': + if (envopt) goto noenvopt; + if (!*++s && (!--argc || !(s = *++argv) || !*s)) { + rb_fatal("Can't chdir"); } - else if (strcmp("yydebug", s) == 0) { - if (envopt) goto noenvopt_long; - opt->dump |= DUMP_BIT(yydebug); - } - else if (is_option_with_arg("dump", Qfalse, Qfalse)) { - ruby_each_words(s, dump_option, &opt->dump); - } - else if (strcmp("help", s) == 0) { - if (envopt) goto noenvopt_long; - opt->dump |= DUMP_BIT(help); - goto switch_end; - } - else if (is_option_with_arg("backtrace-limit", Qfalse, Qfalse)) { - char *e; - long n = strtol(s, &e, 10); - if (errno == ERANGE || n < 0 || *e) rb_raise(rb_eRuntimeError, "wrong limit for backtrace length"); - rb_backtrace_length_limit = n; + if (chdir(s) < 0) { + rb_fatal("Can't chdir to %s", s); + } + break; + + case 'F': + if (envopt) goto noenvopt; + if (*++s) { + rb_fs = rb_reg_new(s, strlen(s), 0); + } + break; + + case 'E': + if (!*++s && (!--argc || !(s = *++argv))) { + rb_raise(rb_eRuntimeError, "missing argument for -E"); } - else { - rb_raise(rb_eRuntimeError, - "invalid option --%s (-h will show valid options)", s); - } - break; - - case '\r': - if (!s[1]) - break; - - default: - { - rb_raise(rb_eRuntimeError, - "invalid option -%c (-h will show valid options)", - (int)(unsigned char)*s); - } - goto switch_end; - - noenvopt: - /* "EIdvwWrKU" only */ - rb_raise(rb_eRuntimeError, "invalid switch in RUBYOPT: -%c", *s); - break; - - noenvopt_long: - rb_raise(rb_eRuntimeError, "invalid switch in RUBYOPT: --%s", s); - break; - - case 0: - break; -# undef is_option_end -# undef check_envopt -# undef need_argument -# undef is_option_with_arg -# undef is_option_with_optarg - } + proc_encoding_option(opt, s, "-E"); + break; + + case 'U': + set_internal_encoding_once(opt, "UTF-8", 0); + ++s; + goto reswitch; + + case 'K': + if (!(s = proc_K_option(opt, s))) break; + goto reswitch; + + case 'I': + forbid_setid("-I"); + if (*++s) + ruby_incpush_expand(s); + else if (argc > 1) { + ruby_incpush_expand(argv[1]); + argc--, argv++; + } + break; + + case '0': + if (envopt) goto noenvopt; + if (!(s = proc_0_option(opt, s))) break; + goto reswitch; + + case '-': + if (!s[1] || (s[1] == '\r' && !s[2])) { + argc--, argv++; + goto switch_end; + } + s++; + + if (!(n = proc_long_options(opt, s, argc, argv, envopt))) goto switch_end; + --n; + argc -= n; + argv += n; + break; + + case '\r': + if (!s[1]) + break; + + default: + rb_raise(rb_eRuntimeError, + "invalid option -%c (-h will show valid options)", + (int)(unsigned char)*s); + goto switch_end; + + noenvopt: + /* "EIdvwWrKU" only */ + rb_raise(rb_eRuntimeError, "invalid switch in RUBYOPT: -%c", *s); + break; + + case 0: + break; + } } switch_end: @@ -1595,12 +1741,24 @@ ruby_init_prelude(void) void rb_call_builtin_inits(void); +// Initialize extra optional exts linked statically. +// This empty definition will be replaced with the actual strong symbol by linker. +#if RBIMPL_HAS_ATTRIBUTE(weak) +__attribute__((weak)) +#endif +void +Init_extra_exts(void) +{ +} + static void ruby_opt_init(ruby_cmdline_options_t *opt) { + rb_warning_category_update(opt->warn.mask, opt->warn.set); + if (opt->dump & dump_exit_bits) return; - if (opt->features.set & FEATURE_BIT(gems)) { + if (FEATURE_SET_P(opt->features, gems)) { rb_define_module("Gem"); if (opt->features.set & FEATURE_BIT(error_highlight)) { rb_define_module("ErrorHighlight"); @@ -1608,14 +1766,53 @@ ruby_opt_init(ruby_cmdline_options_t *opt) if (opt->features.set & FEATURE_BIT(did_you_mean)) { rb_define_module("DidYouMean"); } + if (opt->features.set & FEATURE_BIT(syntax_suggest)) { + rb_define_module("SyntaxSuggest"); + } } - rb_warning_category_update(opt->warn.mask, opt->warn.set); + /* [Feature #19785] Warning for removed GC environment variable. + * Remove this in Ruby 3.4. */ + if (getenv("RUBY_GC_HEAP_INIT_SLOTS")) { + rb_warn_deprecated("The environment variable RUBY_GC_HEAP_INIT_SLOTS", + "environment variables RUBY_GC_HEAP_%d_INIT_SLOTS"); + } + + if (getenv("RUBY_FREE_AT_EXIT")) { + rb_warn("Free at exit is experimental and may be unstable"); + rb_free_at_exit = true; + } + +#if USE_RJIT + // rb_call_builtin_inits depends on RubyVM::RJIT.enabled? + if (opt->rjit.on) + rb_rjit_enabled = true; + if (opt->rjit.stats) + rb_rjit_stats_enabled = true; + if (opt->rjit.trace_exits) + rb_rjit_trace_exits_enabled = true; +#endif Init_ext(); /* load statically linked extensions before rubygems */ + Init_extra_exts(); + + GET_VM()->running = 0; rb_call_builtin_inits(); + GET_VM()->running = 1; + memset(ruby_vm_redefined_flag, 0, sizeof(ruby_vm_redefined_flag)); + ruby_init_prelude(); + // Initialize JITs after prelude because JITing prelude is typically not optimal. +#if USE_RJIT + // Also, rb_rjit_init is safe only after rb_call_builtin_inits() defines RubyVM::RJIT::Compiler. + if (opt->rjit.on) + rb_rjit_init(&opt->rjit); +#endif +#if USE_YJIT + rb_yjit_init(opt->yjit); +#endif + ruby_set_script_name(opt->script_name); require_libraries(&opt->req_list); } @@ -1627,10 +1824,10 @@ opt_enc_index(VALUE enc_name) int i = rb_enc_find_index(s); if (i < 0) { - rb_raise(rb_eRuntimeError, "unknown encoding name - %s", s); + rb_raise(rb_eRuntimeError, "unknown encoding name - %s", s); } else if (rb_enc_dummy_p(rb_enc_from_index(i))) { - rb_raise(rb_eRuntimeError, "dummy encoding is not acceptable - %s ", s); + rb_raise(rb_eRuntimeError, "dummy encoding is not acceptable - %s ", s); } return i; } @@ -1662,8 +1859,8 @@ uscore_get(void) line = rb_lastline_get(); if (!RB_TYPE_P(line, T_STRING)) { - rb_raise(rb_eTypeError, "$_ value need to be String (%s given)", - NIL_P(line) ? "nil" : rb_obj_classname(line)); + rb_raise(rb_eTypeError, "$_ value need to be String (%s given)", + NIL_P(line) ? "nil" : rb_obj_classname(line)); } return line; } @@ -1746,7 +1943,10 @@ rb_f_chomp(int argc, VALUE *argv, VALUE _) static void setup_pager_env(void) { - if (!getenv("LESS")) ruby_setenv("LESS", "-R"); // Output "raw" control characters. + if (!getenv("LESS")) { + // Output "raw" control characters, and move per sections. + ruby_setenv("LESS", "-R +/^[A-Z].*"); + } } #ifdef _WIN32 @@ -1777,12 +1977,273 @@ copy_str(VALUE str, rb_encoding *enc, bool intern) return rb_enc_interned_str(RSTRING_PTR(str), RSTRING_LEN(str), enc); } +#if USE_YJIT +// Check that an environment variable is set to a truthy value +static bool +env_var_truthy(const char *name) +{ + const char *value = getenv(name); + + if (!value) + return false; + if (strcmp(value, "1") == 0) + return true; + if (strcmp(value, "true") == 0) + return true; + if (strcmp(value, "yes") == 0) + return true; + + return false; +} +#endif + +rb_pid_t rb_fork_ruby(int *status); + +static void +show_help(const char *progname, int help) +{ + int tty = isatty(1); + int columns = 0; + if (help && tty) { + const char *pager_env = getenv("RUBY_PAGER"); + if (!pager_env) pager_env = getenv("PAGER"); + if (pager_env && *pager_env && isatty(0)) { + const char *columns_env = getenv("COLUMNS"); + if (columns_env) columns = atoi(columns_env); + VALUE pager = rb_str_new_cstr(pager_env); +#ifdef HAVE_WORKING_FORK + int fds[2]; + if (rb_pipe(fds) == 0) { + rb_pid_t pid = rb_fork_ruby(NULL); + if (pid > 0) { + /* exec PAGER with reading from child */ + dup2(fds[0], 0); + } + else if (pid == 0) { + /* send the help message to the parent PAGER */ + dup2(fds[1], 1); + dup2(fds[1], 2); + } + close(fds[0]); + close(fds[1]); + if (pid > 0) { + setup_pager_env(); + rb_f_exec(1, &pager); + kill(SIGTERM, pid); + rb_waitpid(pid, 0, 0); + } + } +#else + setup_pager_env(); + VALUE port = rb_io_popen(pager, rb_str_new_lit("w"), Qnil, Qnil); + if (!NIL_P(port)) { + int oldout = dup(1); + int olderr = dup(2); + int fd = RFILE(port)->fptr->fd; + tty = tty_enabled(); + dup2(fd, 1); + dup2(fd, 2); + usage(progname, 1, tty, columns); + fflush(stdout); + dup2(oldout, 1); + dup2(olderr, 2); + rb_io_close(port); + return; + } +#endif + } + } + usage(progname, help, tty, columns); +} + +static rb_ast_t * +process_script(ruby_cmdline_options_t *opt) +{ + rb_ast_t *ast; + VALUE parser = rb_parser_new(); + const unsigned int dump = opt->dump; + + if (dump & DUMP_BIT(yydebug)) { + rb_parser_set_yydebug(parser, Qtrue); + } + + if ((dump & dump_exit_bits) && (dump & DUMP_BIT(opt_error_tolerant))) { + rb_parser_error_tolerant(parser); + } + + if (opt->e_script) { + VALUE progname = rb_progname; + rb_parser_set_context(parser, 0, TRUE); + + ruby_opt_init(opt); + ruby_set_script_name(progname); + rb_parser_set_options(parser, opt->do_print, opt->do_loop, + opt->do_line, opt->do_split); + ast = rb_parser_compile_string(parser, opt->script, opt->e_script, 1); + } + else { + VALUE f; + int xflag = opt->xflag; + f = open_load_file(opt->script_name, &xflag); + opt->xflag = xflag != 0; + rb_parser_set_context(parser, 0, f == rb_stdin); + ast = load_file(parser, opt->script_name, f, 1, opt); + } + if (!ast->body.root) { + rb_ast_dispose(ast); + return NULL; + } + return ast; +} + +/** + * Call ruby_opt_init to set up the global state based on the command line + * options, and then warn if prism is enabled and the experimental warning + * category is enabled. + */ +static void +prism_opt_init(ruby_cmdline_options_t *opt) +{ + ruby_opt_init(opt); + + if (rb_warning_category_enabled_p(RB_WARN_CATEGORY_EXPERIMENTAL)) { + rb_category_warn( + RB_WARN_CATEGORY_EXPERIMENTAL, + "The compiler based on the Prism parser is currently experimental " + "and compatibility with the compiler based on parse.y is not yet " + "complete. Please report any issues you find on the `ruby/prism` " + "issue tracker." + ); + } +} + +/** + * Process the command line options and parse the script into the given result. + * Raise an error if the script cannot be parsed. + */ +static void +prism_script(ruby_cmdline_options_t *opt, pm_parse_result_t *result) +{ + memset(result, 0, sizeof(pm_parse_result_t)); + + pm_options_t *options = &result->options; + pm_options_line_set(options, 1); + + if (opt->ext.enc.name != 0) { + pm_options_encoding_set(options, StringValueCStr(opt->ext.enc.name)); + } + + uint8_t command_line = 0; + if (opt->do_split) command_line |= PM_OPTIONS_COMMAND_LINE_A; + if (opt->do_line) command_line |= PM_OPTIONS_COMMAND_LINE_L; + if (opt->do_loop) command_line |= PM_OPTIONS_COMMAND_LINE_N; + if (opt->do_print) command_line |= PM_OPTIONS_COMMAND_LINE_P; + if (opt->xflag) command_line |= PM_OPTIONS_COMMAND_LINE_X; + + VALUE error; + if (strcmp(opt->script, "-") == 0) { + pm_options_command_line_set(options, command_line); + pm_options_filepath_set(options, "-"); + + prism_opt_init(opt); + error = pm_parse_stdin(result); + } + else if (opt->e_script) { + command_line |= PM_OPTIONS_COMMAND_LINE_E; + pm_options_command_line_set(options, command_line); + + prism_opt_init(opt); + error = pm_parse_string(result, opt->e_script, rb_str_new2("-e")); + } + else { + pm_options_command_line_set(options, command_line); + error = pm_load_file(result, opt->script_name); + + // If reading the file did not error, at that point we load the command + // line options. We do it in this order so that if the main script fails + // to load, it doesn't require files required by -r. + if (NIL_P(error)) { + prism_opt_init(opt); + error = pm_parse_file(result, opt->script_name); + } + + // If we found an __END__ marker, then we're going to define a global + // DATA constant that is a file object that can be read to read the + // contents after the marker. + if (NIL_P(error) && result->parser.data_loc.start != NULL) { + int xflag = opt->xflag; + VALUE file = open_load_file(opt->script_name, &xflag); + + const pm_parser_t *parser = &result->parser; + size_t offset = parser->data_loc.start - parser->start + 7; + + if ((parser->start + offset < parser->end) && parser->start[offset] == '\r') offset++; + if ((parser->start + offset < parser->end) && parser->start[offset] == '\n') offset++; + + rb_funcall(file, rb_intern_const("seek"), 2, SIZET2NUM(offset), INT2FIX(SEEK_SET)); + rb_define_global_const("DATA", file); + } + } + + if (!NIL_P(error)) { + pm_parse_result_free(result); + rb_exc_raise(error); + } +} + +static VALUE +prism_dump_tree(pm_parse_result_t *result) +{ + pm_buffer_t output_buffer = { 0 }; + + pm_prettyprint(&output_buffer, &result->parser, result->node.ast_node); + VALUE tree = rb_str_new(output_buffer.value, output_buffer.length); + pm_buffer_free(&output_buffer); + return tree; +} + +static void +process_options_global_setup(const ruby_cmdline_options_t *opt, const rb_iseq_t *iseq) +{ + if (OPT_BACKTRACE_LENGTH_LIMIT_VALID_P(opt)) { + rb_backtrace_length_limit = opt->backtrace_length_limit; + } + + if (opt->do_loop) { + rb_define_global_function("sub", rb_f_sub, -1); + rb_define_global_function("gsub", rb_f_gsub, -1); + rb_define_global_function("chop", rb_f_chop, 0); + rb_define_global_function("chomp", rb_f_chomp, -1); + } + + rb_define_readonly_boolean("$-p", opt->do_print); + rb_define_readonly_boolean("$-l", opt->do_line); + rb_define_readonly_boolean("$-a", opt->do_split); + + rb_gvar_ractor_local("$-p"); + rb_gvar_ractor_local("$-l"); + rb_gvar_ractor_local("$-a"); + + if ((rb_e_script = opt->e_script) != 0) { + rb_str_freeze(rb_e_script); + rb_vm_register_global_object(opt->e_script); + } + + rb_execution_context_t *ec = GET_EC(); + VALUE script = (opt->e_script ? opt->e_script : Qnil); + rb_exec_event_hook_script_compiled(ec, iseq, script); +} + static VALUE process_options(int argc, char **argv, ruby_cmdline_options_t *opt) { - rb_ast_t *ast = 0; - VALUE parser; - VALUE script_name; + struct { + rb_ast_t *ast; + pm_parse_result_t prism; + } result = {0}; +#define dispose_result() \ + (result.ast ? rb_ast_dispose(result.ast) : pm_parse_result_free(&result.prism)) + const rb_iseq_t *iseq; rb_encoding *enc, *lenc; #if UTF8_PATH @@ -1797,85 +2258,19 @@ process_options(int argc, char **argv, ruby_cmdline_options_t *opt) const long loaded_before_enc = RARRAY_LEN(vm->loaded_features); if (opt->dump & (DUMP_BIT(usage)|DUMP_BIT(help))) { - int tty = isatty(1); - const char *const progname = - (argc > 0 && argv && argv[0] ? argv[0] : - origarg.argc > 0 && origarg.argv && origarg.argv[0] ? origarg.argv[0] : - ruby_engine); - int columns = 0; - if ((opt->dump & DUMP_BIT(help)) && tty) { - const char *pager_env = getenv("RUBY_PAGER"); - if (!pager_env) pager_env = getenv("PAGER"); - if (pager_env && *pager_env && isatty(0)) { - const char *columns_env = getenv("COLUMNS"); - if (columns_env) columns = atoi(columns_env); - VALUE pager = rb_str_new_cstr(pager_env); -#ifdef HAVE_WORKING_FORK - int fds[2]; - if (rb_pipe(fds) == 0) { - rb_pid_t pid = rb_fork(); - if (pid > 0) { - /* exec PAGER with reading from child */ - dup2(fds[0], 0); - } - else if (pid == 0) { - /* send the help message to the parent PAGER */ - dup2(fds[1], 1); - dup2(fds[1], 2); - } - close(fds[0]); - close(fds[1]); - if (pid > 0) { - setup_pager_env(); - rb_f_exec(1, &pager); - kill(SIGTERM, pid); - rb_waitpid(pid, 0, 0); - } - } -#else - setup_pager_env(); - VALUE port = rb_io_popen(pager, rb_str_new_lit("w"), Qnil, Qnil); - if (!NIL_P(port)) { - int oldout = dup(1); - int olderr = dup(2); - int fd = RFILE(port)->fptr->fd; - tty = tty_enabled(); - dup2(fd, 1); - dup2(fd, 2); - usage(progname, 1, tty, columns); - fflush(stdout); - dup2(oldout, 1); - dup2(olderr, 2); - rb_io_close(port); - return Qtrue; - } -#endif - } - } - usage(progname, (opt->dump & DUMP_BIT(help)), tty, columns); - return Qtrue; + const char *const progname = + (argc > 0 && argv && argv[0] ? argv[0] : + origarg.argc > 0 && origarg.argv && origarg.argv[0] ? origarg.argv[0] : + ruby_engine); + show_help(progname, (opt->dump & DUMP_BIT(help))); + return Qtrue; } argc -= i; argv += i; - if ((opt->features.set & FEATURE_BIT(rubyopt)) && (s = getenv("RUBYOPT"))) { - VALUE src_enc_name = opt->src.enc.name; - VALUE ext_enc_name = opt->ext.enc.name; - VALUE int_enc_name = opt->intern.enc.name; - ruby_features_t feat = opt->features; - ruby_features_t warn = opt->warn; - - opt->src.enc.name = opt->ext.enc.name = opt->intern.enc.name = 0; - moreswitches(s, opt, 1); - if (src_enc_name) - opt->src.enc.name = src_enc_name; - if (ext_enc_name) - opt->ext.enc.name = ext_enc_name; - if (int_enc_name) - opt->intern.enc.name = int_enc_name; - FEATURE_SET_RESTORE(opt->features, feat); - FEATURE_SET_RESTORE(opt->warn, warn); + if (FEATURE_SET_P(opt->features, rubyopt) && (s = getenv("RUBYOPT"))) { + moreswitches(s, opt, 1); } if (opt->src.enc.name) @@ -1884,68 +2279,71 @@ process_options(int argc, char **argv, ruby_cmdline_options_t *opt) */ rb_warning("-K is specified; it is for 1.8 compatibility and may cause odd behavior"); -#ifdef __OpenBSD__ - /* Disable yjit on OpenBSD, stops --enable-all from failing with: - mmap call failed: Not supported */ - opt->features.set &= ~FEATURE_BIT(yjit); + if (!(FEATURE_SET_BITS(opt->features) & feature_jit_mask)) { +#if USE_YJIT + if (!FEATURE_USED_P(opt->features, yjit) && env_var_truthy("RUBY_YJIT_ENABLE")) { + FEATURE_SET(opt->features, FEATURE_BIT(yjit)); + } #endif + } + if (MULTI_BITS_P(FEATURE_SET_BITS(opt->features) & feature_jit_mask)) { + rb_warn("RJIT and YJIT cannot both be enabled at the same time. Exiting"); + return Qfalse; + } -#if USE_MJIT - if (opt->features.set & FEATURE_BIT(jit)) { - opt->mjit.on = TRUE; /* set mjit.on for ruby_show_version() API and check to call mjit_init() */ +#if USE_RJIT + if (FEATURE_SET_P(opt->features, rjit)) { + opt->rjit.on = true; // set opt->rjit.on for Init_ruby_description() and calling rb_rjit_init() } #endif - if (opt->features.set & FEATURE_BIT(yjit)) { -#if USE_MJIT - if (opt->mjit.on) { - rb_warn("MJIT and YJIT cannot both be enabled at the same time. Exiting"); - exit(1); - } -#endif - rb_yjit_init(&opt->yjit); +#if USE_YJIT + if (FEATURE_SET_P(opt->features, yjit)) { + opt->yjit = true; // set opt->yjit for Init_ruby_description() and calling rb_yjit_init() } - if (opt->dump & (DUMP_BIT(version) | DUMP_BIT(version_v))) { -#if USE_MJIT - mjit_opts.on = opt->mjit.on; /* used by ruby_show_version(). mjit_init() still can't be called here. */ #endif - ruby_show_version(); - if (opt->dump & DUMP_BIT(version)) return Qtrue; + + ruby_mn_threads_params(); + Init_ruby_description(opt); + + if (opt->dump & (DUMP_BIT(version) | DUMP_BIT(version_v))) { + ruby_show_version(); + if (opt->dump & DUMP_BIT(version)) return Qtrue; } if (opt->dump & DUMP_BIT(copyright)) { - ruby_show_copyright(); - return Qtrue; + ruby_show_copyright(); + return Qtrue; } if (!opt->e_script) { - if (argc <= 0) { /* no more args */ - if (opt->verbose) - return Qtrue; - opt->script = "-"; - } - else { - opt->script = argv[0]; - if (!opt->script || opt->script[0] == '\0') { - opt->script = "-"; - } - else if (opt->do_search) { - const char *path = getenv("RUBYPATH"); - - opt->script = 0; - if (path) { - opt->script = dln_find_file_r(argv[0], path, fbuf, sizeof(fbuf)); - } - if (!opt->script) { - opt->script = dln_find_file_r(argv[0], getenv(PATH_ENV), fbuf, sizeof(fbuf)); - } - if (!opt->script) - opt->script = argv[0]; - } - argc--; - argv++; - } - if (opt->script[0] == '-' && !opt->script[1]) { - forbid_setid("program input from stdin"); - } + if (argc <= 0) { /* no more args */ + if (opt->verbose) + return Qtrue; + opt->script = "-"; + } + else { + opt->script = argv[0]; + if (!opt->script || opt->script[0] == '\0') { + opt->script = "-"; + } + else if (opt->do_search) { + const char *path = getenv("RUBYPATH"); + + opt->script = 0; + if (path) { + opt->script = dln_find_file_r(argv[0], path, fbuf, sizeof(fbuf)); + } + if (!opt->script) { + opt->script = dln_find_file_r(argv[0], getenv(PATH_ENV), fbuf, sizeof(fbuf)); + } + if (!opt->script) + opt->script = argv[0]; + } + argc--; + argv++; + } + if (opt->script[0] == '-' && !opt->script[1]) { + forbid_setid("program input from stdin"); + } } opt->script_name = rb_str_new_cstr(opt->script); @@ -1953,86 +2351,72 @@ process_options(int argc, char **argv, ruby_cmdline_options_t *opt) #ifdef _WIN32 translit_char_bin(RSTRING_PTR(opt->script_name), '\\', '/'); -#elif defined DOSISH - translit_char(RSTRING_PTR(opt->script_name), '\\', '/'); #endif ruby_gc_set_params(); ruby_init_loadpath(); -#if USE_MJIT - if (opt->mjit.on) - /* Using TMP_RUBY_PREFIX created by ruby_init_loadpath(). */ - mjit_init(&opt->mjit); -#endif - - Init_ruby_description(); Init_enc(); lenc = rb_locale_encoding(); rb_enc_associate(rb_progname, lenc); rb_obj_freeze(rb_progname); - parser = rb_parser_new(); - if (opt->dump & DUMP_BIT(yydebug)) { - rb_parser_set_yydebug(parser, Qtrue); - } if (opt->ext.enc.name != 0) { - opt->ext.enc.index = opt_enc_index(opt->ext.enc.name); + opt->ext.enc.index = opt_enc_index(opt->ext.enc.name); } if (opt->intern.enc.name != 0) { - opt->intern.enc.index = opt_enc_index(opt->intern.enc.name); + opt->intern.enc.index = opt_enc_index(opt->intern.enc.name); } if (opt->src.enc.name != 0) { - opt->src.enc.index = opt_enc_index(opt->src.enc.name); - src_encoding_index = opt->src.enc.index; + opt->src.enc.index = opt_enc_index(opt->src.enc.name); + src_encoding_index = opt->src.enc.index; } if (opt->ext.enc.index >= 0) { - enc = rb_enc_from_index(opt->ext.enc.index); + enc = rb_enc_from_index(opt->ext.enc.index); } else { - enc = IF_UTF8_PATH(uenc, lenc); + enc = IF_UTF8_PATH(uenc, lenc); } rb_enc_set_default_external(rb_enc_from_encoding(enc)); if (opt->intern.enc.index >= 0) { - enc = rb_enc_from_index(opt->intern.enc.index); - rb_enc_set_default_internal(rb_enc_from_encoding(enc)); - opt->intern.enc.index = -1; + enc = rb_enc_from_index(opt->intern.enc.index); + rb_enc_set_default_internal(rb_enc_from_encoding(enc)); + opt->intern.enc.index = -1; #if UTF8_PATH - ienc = enc; + ienc = enc; #endif } - script_name = opt->script_name; rb_enc_associate(opt->script_name, IF_UTF8_PATH(uenc, lenc)); #if UTF8_PATH if (uenc != lenc) { - opt->script_name = str_conv_enc(opt->script_name, uenc, lenc); - opt->script = RSTRING_PTR(opt->script_name); + opt->script_name = str_conv_enc(opt->script_name, uenc, lenc); + opt->script = RSTRING_PTR(opt->script_name); } #endif rb_obj_freeze(opt->script_name); if (IF_UTF8_PATH(uenc != lenc, 1)) { - long i; + long i; VALUE load_path = vm->load_path; - const ID id_initial_load_path_mark = INITIAL_LOAD_PATH_MARK; + const ID id_initial_load_path_mark = INITIAL_LOAD_PATH_MARK; int modifiable = FALSE; rb_get_expanded_load_path(); - for (i = 0; i < RARRAY_LEN(load_path); ++i) { - VALUE path = RARRAY_AREF(load_path, i); - int mark = rb_attr_get(path, id_initial_load_path_mark) == path; + for (i = 0; i < RARRAY_LEN(load_path); ++i) { + VALUE path = RARRAY_AREF(load_path, i); + int mark = rb_attr_get(path, id_initial_load_path_mark) == path; #if UTF8_PATH - VALUE newpath = rb_str_conv_enc(path, uenc, lenc); - if (newpath == path) continue; - path = newpath; + VALUE newpath = rb_str_conv_enc(path, uenc, lenc); + if (newpath == path) continue; + path = newpath; #else - if (!(path = copy_str(path, lenc, !mark))) continue; + if (!(path = copy_str(path, lenc, !mark))) continue; #endif - if (mark) rb_ivar_set(path, id_initial_load_path_mark, path); + if (mark) rb_ivar_set(path, id_initial_load_path_mark, path); if (!modifiable) { rb_ary_modify(load_path); modifiable = TRUE; } - RARRAY_ASET(load_path, i, path); - } + RARRAY_ASET(load_path, i, path); + } if (modifiable) { rb_ary_replace(vm->load_path_snapshot, load_path); } @@ -2041,10 +2425,13 @@ process_options(int argc, char **argv, ruby_cmdline_options_t *opt) VALUE loaded_features = vm->loaded_features; bool modified = false; for (long i = loaded_before_enc; i < RARRAY_LEN(loaded_features); ++i) { - VALUE path = RARRAY_AREF(loaded_features, i); + VALUE path = RARRAY_AREF(loaded_features, i); if (!(path = copy_str(path, IF_UTF8_PATH(uenc, lenc), true))) continue; - modified = true; - RARRAY_ASET(loaded_features, i, path); + if (!modified) { + rb_ary_modify(loaded_features); + modified = true; + } + RARRAY_ASET(loaded_features, i, path); } if (modified) { rb_ary_replace(vm->loaded_features_snapshot, loaded_features); @@ -2052,158 +2439,134 @@ process_options(int argc, char **argv, ruby_cmdline_options_t *opt) } if (opt->features.mask & COMPILATION_FEATURES) { - VALUE option = rb_hash_new(); + VALUE option = rb_hash_new(); #define SET_COMPILE_OPTION(h, o, name) \ - rb_hash_aset((h), ID2SYM(rb_intern_const(#name)), \ - RBOOL(FEATURE_SET_P(o->features, FEATURE_BIT(name)))); - SET_COMPILE_OPTION(option, opt, frozen_string_literal); - SET_COMPILE_OPTION(option, opt, debug_frozen_string_literal); - rb_funcallv(rb_cISeq, rb_intern_const("compile_option="), 1, &option); + rb_hash_aset((h), ID2SYM(rb_intern_const(#name)), \ + RBOOL(FEATURE_SET_P(o->features, name))) + SET_COMPILE_OPTION(option, opt, frozen_string_literal); + SET_COMPILE_OPTION(option, opt, debug_frozen_string_literal); + rb_funcallv(rb_cISeq, rb_intern_const("compile_option="), 1, &option); #undef SET_COMPILE_OPTION } ruby_set_argv(argc, argv); - process_sflag(&opt->sflag); - - rb_parser_set_context(parser, 0, TRUE); + opt->sflag = process_sflag(opt->sflag); if (opt->e_script) { - VALUE progname = rb_progname; - rb_encoding *eenc; - if (opt->src.enc.index >= 0) { - eenc = rb_enc_from_index(opt->src.enc.index); - } - else { - eenc = lenc; + rb_encoding *eenc; + if (opt->src.enc.index >= 0) { + eenc = rb_enc_from_index(opt->src.enc.index); + } + else { + eenc = lenc; #if UTF8_PATH - if (ienc) eenc = ienc; + if (ienc) eenc = ienc; #endif - } + } #if UTF8_PATH - if (eenc != uenc) { - opt->e_script = str_conv_enc(opt->e_script, uenc, eenc); - } + if (eenc != uenc) { + opt->e_script = str_conv_enc(opt->e_script, uenc, eenc); + } #endif - rb_enc_associate(opt->e_script, eenc); - ruby_opt_init(opt); - ruby_set_script_name(progname); - rb_parser_set_options(parser, opt->do_print, opt->do_loop, - opt->do_line, opt->do_split); - ast = rb_parser_compile_string(parser, opt->script, opt->e_script, 1); + rb_enc_associate(opt->e_script, eenc); + } + + if (!(*rb_ruby_prism_ptr())) { + if (!(result.ast = process_script(opt))) return Qfalse; } else { - VALUE f; - f = open_load_file(script_name, &opt->xflag); - ast = load_file(parser, opt->script_name, f, 1, opt); + prism_script(opt, &result.prism); } ruby_set_script_name(opt->script_name); - if (dump & DUMP_BIT(yydebug)) { - dump &= ~DUMP_BIT(yydebug); - if (!dump) return Qtrue; + if ((dump & DUMP_BIT(yydebug)) && !(dump &= ~DUMP_BIT(yydebug))) { + dispose_result(); + return Qtrue; } if (opt->ext.enc.index >= 0) { - enc = rb_enc_from_index(opt->ext.enc.index); + enc = rb_enc_from_index(opt->ext.enc.index); } else { - enc = IF_UTF8_PATH(uenc, lenc); + enc = IF_UTF8_PATH(uenc, lenc); } rb_enc_set_default_external(rb_enc_from_encoding(enc)); if (opt->intern.enc.index >= 0) { - /* Set in the shebang line */ - enc = rb_enc_from_index(opt->intern.enc.index); - rb_enc_set_default_internal(rb_enc_from_encoding(enc)); + /* Set in the shebang line */ + enc = rb_enc_from_index(opt->intern.enc.index); + rb_enc_set_default_internal(rb_enc_from_encoding(enc)); } else if (!rb_default_internal_encoding()) - /* Freeze default_internal */ - rb_enc_set_default_internal(Qnil); + /* Freeze default_internal */ + rb_enc_set_default_internal(Qnil); rb_stdio_set_default_encoding(); - if (!ast->body.root) { - rb_ast_dispose(ast); - return Qfalse; - } - - process_sflag(&opt->sflag); + opt->sflag = process_sflag(opt->sflag); opt->xflag = 0; if (dump & DUMP_BIT(syntax)) { - printf("Syntax OK\n"); - dump &= ~DUMP_BIT(syntax); - if (!dump) return Qtrue; + printf("Syntax OK\n"); + dump &= ~DUMP_BIT(syntax); + if (!dump) return Qtrue; } - if (opt->do_loop) { - rb_define_global_function("sub", rb_f_sub, -1); - rb_define_global_function("gsub", rb_f_gsub, -1); - rb_define_global_function("chop", rb_f_chop, 0); - rb_define_global_function("chomp", rb_f_chomp, -1); - } - - if (dump & (DUMP_BIT(parsetree)|DUMP_BIT(parsetree_with_comment))) { - rb_io_write(rb_stdout, rb_parser_dump_tree(ast->body.root, dump & DUMP_BIT(parsetree_with_comment))); - rb_io_flush(rb_stdout); - dump &= ~DUMP_BIT(parsetree)&~DUMP_BIT(parsetree_with_comment); - if (!dump) { - rb_ast_dispose(ast); - return Qtrue; - } + if (dump & DUMP_BIT(parsetree)) { + VALUE tree; + if (result.ast) { + int comment = opt->dump & DUMP_BIT(opt_comment); + tree = rb_parser_dump_tree(result.ast->body.root, comment); + } + else { + tree = prism_dump_tree(&result.prism); + } + rb_io_write(rb_stdout, tree); + rb_io_flush(rb_stdout); + dump &= ~DUMP_BIT(parsetree); + if (!dump) { + dispose_result(); + return Qtrue; + } } { - VALUE path = Qnil; - if (!opt->e_script && strcmp(opt->script, "-")) { - path = rb_realpath_internal(Qnil, script_name, 1); + VALUE path = Qnil; + if (!opt->e_script && strcmp(opt->script, "-")) { + path = rb_realpath_internal(Qnil, opt->script_name, 1); #if UTF8_PATH - if (uenc != lenc) { - path = str_conv_enc(path, uenc, lenc); - } + if (uenc != lenc) { + path = str_conv_enc(path, uenc, lenc); + } #endif - if (!ENCODING_GET(path)) { /* ASCII-8BIT */ - rb_enc_copy(path, opt->script_name); - } - } + if (!ENCODING_GET(path)) { /* ASCII-8BIT */ + rb_enc_copy(path, opt->script_name); + } + } rb_binding_t *toplevel_binding; - GetBindingPtr(rb_const_get(rb_cObject, rb_intern("TOPLEVEL_BINDING")), - toplevel_binding); + GetBindingPtr(rb_const_get(rb_cObject, rb_intern("TOPLEVEL_BINDING")), toplevel_binding); const struct rb_block *base_block = toplevel_context(toplevel_binding); - iseq = rb_iseq_new_main(&ast->body, opt->script_name, path, vm_block_iseq(base_block)); - rb_ast_dispose(ast); + const rb_iseq_t *parent = vm_block_iseq(base_block); + bool optimize = (opt->dump & DUMP_BIT(opt_optimize)) != 0; + + if (!result.ast) { + pm_parse_result_t *pm = &result.prism; + iseq = pm_iseq_new_main(&pm->node, opt->script_name, path, parent, optimize); + pm_parse_result_free(pm); + } + else { + rb_ast_t *ast = result.ast; + iseq = rb_iseq_new_main(&ast->body, opt->script_name, path, parent, optimize); + rb_ast_dispose(ast); + } } if (dump & DUMP_BIT(insns)) { - rb_io_write(rb_stdout, rb_iseq_disasm((const rb_iseq_t *)iseq)); - rb_io_flush(rb_stdout); - dump &= ~DUMP_BIT(insns); - if (!dump) return Qtrue; + rb_io_write(rb_stdout, rb_iseq_disasm((const rb_iseq_t *)iseq)); + rb_io_flush(rb_stdout); + dump &= ~DUMP_BIT(insns); + if (!dump) return Qtrue; } if (opt->dump & dump_exit_bits) return Qtrue; - rb_define_readonly_boolean("$-p", opt->do_print); - rb_define_readonly_boolean("$-l", opt->do_line); - rb_define_readonly_boolean("$-a", opt->do_split); - - rb_gvar_ractor_local("$-p"); - rb_gvar_ractor_local("$-l"); - rb_gvar_ractor_local("$-a"); - - if ((rb_e_script = opt->e_script) != 0) { - rb_str_freeze(rb_e_script); - rb_gc_register_mark_object(opt->e_script); - } - - { - rb_execution_context_t *ec = GET_EC(); - - if (opt->e_script) { - /* -e */ - rb_exec_event_hook_script_compiled(ec, iseq, opt->e_script); - } - else { - /* file */ - rb_exec_event_hook_script_compiled(ec, iseq, Qnil); - } - } + process_options_global_setup(opt, iseq); return (VALUE)iseq; } @@ -2211,14 +2574,16 @@ process_options(int argc, char **argv, ruby_cmdline_options_t *opt) static void warn_cr_in_shebang(const char *str, long len) { - if (str[len-1] == '\n' && str[len-2] == '\r') { - rb_warn("shebang line ending with \\r may cause problems"); + if (len > 1 && str[len-1] == '\n' && str[len-2] == '\r') { + rb_warn("shebang line ending with \\r may cause problems"); } } #else #define warn_cr_in_shebang(str, len) (void)0 #endif +void rb_reset_argf_lineno(long n); + struct load_file_arg { VALUE parser; VALUE fname; @@ -2227,6 +2592,8 @@ struct load_file_arg { VALUE f; }; +void rb_set_script_lines_for(VALUE vparser, VALUE path); + static VALUE load_file_internal(VALUE argp_v) { @@ -2243,114 +2610,118 @@ load_file_internal(VALUE argp_v) CONST_ID(set_encoding, "set_encoding"); if (script) { - VALUE c = 1; /* something not nil */ - VALUE line; - char *p, *str; - long len; - int no_src_enc = !opt->src.enc.name; - int no_ext_enc = !opt->ext.enc.name; - int no_int_enc = !opt->intern.enc.name; - - enc = rb_ascii8bit_encoding(); - rb_funcall(f, set_encoding, 1, rb_enc_from_encoding(enc)); - - if (opt->xflag) { - line_start--; - search_shebang: - while (!NIL_P(line = rb_io_gets(f))) { - line_start++; - RSTRING_GETMEM(line, str, len); - if (len > 2 && str[0] == '#' && str[1] == '!') { - if (line_start == 1) warn_cr_in_shebang(str, len); - if ((p = strstr(str+2, ruby_engine)) != 0) { - goto start_read; - } - } - } - rb_loaderror("no Ruby script found in input"); - } - - c = rb_io_getbyte(f); - if (c == INT2FIX('#')) { - c = rb_io_getbyte(f); + VALUE c = 1; /* something not nil */ + VALUE line; + char *p, *str; + long len; + int no_src_enc = !opt->src.enc.name; + int no_ext_enc = !opt->ext.enc.name; + int no_int_enc = !opt->intern.enc.name; + + enc = rb_ascii8bit_encoding(); + rb_funcall(f, set_encoding, 1, rb_enc_from_encoding(enc)); + + if (opt->xflag) { + line_start--; + search_shebang: + while (!NIL_P(line = rb_io_gets(f))) { + line_start++; + RSTRING_GETMEM(line, str, len); + if (len > 2 && str[0] == '#' && str[1] == '!') { + if (line_start == 1) warn_cr_in_shebang(str, len); + if ((p = strstr(str+2, ruby_engine)) != 0) { + goto start_read; + } + } + } + rb_loaderror("no Ruby script found in input"); + } + + c = rb_io_getbyte(f); + if (c == INT2FIX('#')) { + c = rb_io_getbyte(f); if (c == INT2FIX('!') && !NIL_P(line = rb_io_gets(f))) { - RSTRING_GETMEM(line, str, len); - warn_cr_in_shebang(str, len); - if ((p = strstr(str, ruby_engine)) == 0) { - /* not ruby script, assume -x flag */ - goto search_shebang; - } - - start_read: - str += len - 1; - if (*str == '\n') *str-- = '\0'; - if (*str == '\r') *str-- = '\0'; - /* ruby_engine should not contain a space */ - if ((p = strstr(p, " -")) != 0) { - opt->warning = 0; - moreswitches(p + 1, opt, 0); - } - - /* push back shebang for pragma may exist in next line */ - rb_io_ungetbyte(f, rb_str_new2("!\n")); - } - else if (!NIL_P(c)) { - rb_io_ungetbyte(f, c); - } - rb_io_ungetbyte(f, INT2FIX('#')); - if (no_src_enc && opt->src.enc.name) { - opt->src.enc.index = opt_enc_index(opt->src.enc.name); - src_encoding_index = opt->src.enc.index; - } - if (no_ext_enc && opt->ext.enc.name) { - opt->ext.enc.index = opt_enc_index(opt->ext.enc.name); - } - if (no_int_enc && opt->intern.enc.name) { - opt->intern.enc.index = opt_enc_index(opt->intern.enc.name); - } - } - else if (!NIL_P(c)) { - rb_io_ungetbyte(f, c); - } + RSTRING_GETMEM(line, str, len); + warn_cr_in_shebang(str, len); + if ((p = strstr(str, ruby_engine)) == 0) { + /* not ruby script, assume -x flag */ + goto search_shebang; + } + + start_read: + str += len - 1; + if (*str == '\n') *str-- = '\0'; + if (*str == '\r') *str-- = '\0'; + /* ruby_engine should not contain a space */ + if ((p = strstr(p, " -")) != 0) { + opt->warning = 0; + moreswitches(p + 1, opt, 0); + } + + /* push back shebang for pragma may exist in next line */ + rb_io_ungetbyte(f, rb_str_new2("!\n")); + } + else if (!NIL_P(c)) { + rb_io_ungetbyte(f, c); + } + rb_io_ungetbyte(f, INT2FIX('#')); + if (no_src_enc && opt->src.enc.name) { + opt->src.enc.index = opt_enc_index(opt->src.enc.name); + src_encoding_index = opt->src.enc.index; + } + if (no_ext_enc && opt->ext.enc.name) { + opt->ext.enc.index = opt_enc_index(opt->ext.enc.name); + } + if (no_int_enc && opt->intern.enc.name) { + opt->intern.enc.index = opt_enc_index(opt->intern.enc.name); + } + } + else if (!NIL_P(c)) { + rb_io_ungetbyte(f, c); + } if (NIL_P(c)) { - argp->f = f = Qnil; - } + argp->f = f = Qnil; + } + rb_reset_argf_lineno(0); ruby_opt_init(opt); } if (opt->src.enc.index >= 0) { - enc = rb_enc_from_index(opt->src.enc.index); + enc = rb_enc_from_index(opt->src.enc.index); } else if (f == rb_stdin) { - enc = rb_locale_encoding(); + enc = rb_locale_encoding(); } else { - enc = rb_utf8_encoding(); + enc = rb_utf8_encoding(); } rb_parser_set_options(parser, opt->do_print, opt->do_loop, - opt->do_line, opt->do_split); + opt->do_line, opt->do_split); + + rb_set_script_lines_for(parser, orig_fname); + if (NIL_P(f)) { - f = rb_str_new(0, 0); - rb_enc_associate(f, enc); - return (VALUE)rb_parser_compile_string_path(parser, orig_fname, f, line_start); + f = rb_str_new(0, 0); + rb_enc_associate(f, enc); + return (VALUE)rb_parser_compile_string_path(parser, orig_fname, f, line_start); } rb_funcall(f, set_encoding, 2, rb_enc_from_encoding(enc), rb_str_new_cstr("-")); ast = rb_parser_compile_file_path(parser, orig_fname, f, line_start); rb_funcall(f, set_encoding, 1, rb_parser_encoding(parser)); if (script && rb_parser_end_seen_p(parser)) { - /* - * DATA is a File that contains the data section of the executed file. - * To create a data section use <tt>__END__</tt>: - * - * $ cat t.rb - * puts DATA.gets - * __END__ - * hello world! - * - * $ ruby t.rb - * hello world! - */ - rb_define_global_const("DATA", f); - argp->f = Qnil; + /* + * DATA is a File that contains the data section of the executed file. + * To create a data section use <tt>__END__</tt>: + * + * $ cat t.rb + * puts DATA.gets + * __END__ + * hello world! + * + * $ ruby t.rb + * hello world! + */ + rb_define_global_const("DATA", f); + argp->f = Qnil; } return (VALUE)ast; } @@ -2379,71 +2750,72 @@ static VALUE open_load_file(VALUE fname_v, int *xflag) { const char *fname = (fname_v = rb_str_encode_ospath(fname_v), - StringValueCStr(fname_v)); + StringValueCStr(fname_v)); long flen = RSTRING_LEN(fname_v); VALUE f; int e; if (flen == 1 && fname[0] == '-') { - f = rb_stdin; + f = rb_stdin; } else { - int fd; - /* open(2) may block if fname is point to FIFO and it's empty. Let's - use O_NONBLOCK. */ -#if defined O_NONBLOCK && HAVE_FCNTL && !(O_NONBLOCK & O_ACCMODE) - /* TODO: fix conflicting O_NONBLOCK in ruby/win32.h */ -# define MODE_TO_LOAD (O_RDONLY | O_NONBLOCK) -#elif defined O_NDELAY && HAVE_FCNTL && !(O_NDELAY & O_ACCMODE) -# define MODE_TO_LOAD (O_RDONLY | O_NDELAY) -#else -# define MODE_TO_LOAD (O_RDONLY) + int fd; + /* open(2) may block if fname is point to FIFO and it's empty. Let's + use O_NONBLOCK. */ + const int MODE_TO_LOAD = O_RDONLY | ( +#if defined O_NONBLOCK && HAVE_FCNTL + /* TODO: fix conflicting O_NONBLOCK in ruby/win32.h */ + !(O_NONBLOCK & O_ACCMODE) ? O_NONBLOCK : +#endif +#if defined O_NDELAY && HAVE_FCNTL + !(O_NDELAY & O_ACCMODE) ? O_NDELAY : #endif - int mode = MODE_TO_LOAD; + 0); + int mode = MODE_TO_LOAD; #if defined DOSISH || defined __CYGWIN__ # define isdirsep(x) ((x) == '/' || (x) == '\\') - { - static const char exeext[] = ".exe"; - enum {extlen = sizeof(exeext)-1}; - if (flen > extlen && !isdirsep(fname[flen-extlen-1]) && - STRNCASECMP(fname+flen-extlen, exeext, extlen) == 0) { - mode |= O_BINARY; - *xflag = 1; - } - } -#endif - - if ((fd = rb_cloexec_open(fname, mode, 0)) < 0) { - e = errno; - if (!rb_gc_for_fd(e)) { - rb_load_fail(fname_v, strerror(e)); - } - if ((fd = rb_cloexec_open(fname, mode, 0)) < 0) { - rb_load_fail(fname_v, strerror(errno)); - } - } - rb_update_max_fd(fd); - - if (MODE_TO_LOAD != O_RDONLY && (e = disable_nonblock(fd)) != 0) { - (void)close(fd); - rb_load_fail(fname_v, strerror(e)); - } - - e = ruby_is_fd_loadable(fd); - if (!e) { - e = errno; - (void)close(fd); - rb_load_fail(fname_v, strerror(e)); - } - - f = rb_io_fdopen(fd, mode, fname); - if (e < 0) { - /* - We need to wait if FIFO is empty. It's FIFO's semantics. - rb_thread_wait_fd() release GVL. So, it's safe. - */ - rb_io_wait(f, RB_INT2NUM(RUBY_IO_READABLE), Qnil); - } + { + static const char exeext[] = ".exe"; + enum {extlen = sizeof(exeext)-1}; + if (flen > extlen && !isdirsep(fname[flen-extlen-1]) && + STRNCASECMP(fname+flen-extlen, exeext, extlen) == 0) { + mode |= O_BINARY; + *xflag = 1; + } + } +#endif + + if ((fd = rb_cloexec_open(fname, mode, 0)) < 0) { + e = errno; + if (!rb_gc_for_fd(e)) { + rb_load_fail(fname_v, strerror(e)); + } + if ((fd = rb_cloexec_open(fname, mode, 0)) < 0) { + rb_load_fail(fname_v, strerror(errno)); + } + } + rb_update_max_fd(fd); + + if (MODE_TO_LOAD != O_RDONLY && (e = disable_nonblock(fd)) != 0) { + (void)close(fd); + rb_load_fail(fname_v, strerror(e)); + } + + e = ruby_is_fd_loadable(fd); + if (!e) { + e = errno; + (void)close(fd); + rb_load_fail(fname_v, strerror(e)); + } + + f = rb_io_fdopen(fd, mode, fname); + if (e < 0) { + /* + We need to wait if FIFO is empty. It's FIFO's semantics. + rb_thread_wait_fd() release GVL. So, it's safe. + */ + rb_io_wait(f, RB_INT2NUM(RUBY_IO_READABLE), Qnil); + } } return f; } @@ -2455,7 +2827,7 @@ restore_load_file(VALUE arg) VALUE f = argp->f; if (!NIL_P(f) && f != rb_stdin) { - rb_io_close(f); + rb_io_close(f); } return Qnil; } @@ -2470,7 +2842,7 @@ load_file(VALUE parser, VALUE fname, VALUE f, int script, ruby_cmdline_options_t arg.opt = opt; arg.f = f; return (rb_ast_t *)rb_ensure(load_file_internal, (VALUE)&arg, - restore_load_file, (VALUE)&arg); + restore_load_file, (VALUE)&arg); } void * @@ -2490,7 +2862,9 @@ void * rb_parser_load_file(VALUE parser, VALUE fname_v) { ruby_cmdline_options_t opt; - VALUE f = open_load_file(fname_v, &cmdline_options_init(&opt)->xflag); + int xflag = 0; + VALUE f = open_load_file(fname_v, &xflag); + cmdline_options_init(&opt)->xflag = xflag != 0; return load_file(parser, fname_v, f, 0, &opt); } @@ -2548,7 +2922,7 @@ static void set_arg0(VALUE val, ID id, VALUE *_) { if (origarg.argv == 0) - rb_raise(rb_eRuntimeError, "$0 not initialized"); + rb_raise(rb_eRuntimeError, "$0 not initialized"); rb_progname = rb_str_new_frozen(ruby_setproctitle(val)); } @@ -2565,12 +2939,18 @@ external_str_new_cstr(const char *p) #endif } +static void +set_progname(VALUE name) +{ + rb_orig_progname = rb_progname = name; + rb_vm_set_progname(rb_progname); +} + void ruby_script(const char *name) { if (name) { - rb_orig_progname = rb_progname = external_str_new_cstr(name); - rb_vm_set_progname(rb_progname); + set_progname(rb_str_freeze(external_str_new_cstr(name))); } } @@ -2581,8 +2961,7 @@ ruby_script(const char *name) void ruby_set_script_name(VALUE name) { - rb_orig_progname = rb_progname = rb_str_dup(name); - rb_vm_set_progname(rb_progname); + set_progname(rb_str_new_frozen(name)); } static void @@ -2626,13 +3005,13 @@ opt_W_getter(ID id, VALUE *dmy) switch (v) { case Qnil: - return INT2FIX(0); + return INT2FIX(0); case Qfalse: - return INT2FIX(1); + return INT2FIX(1); case Qtrue: - return INT2FIX(2); + return INT2FIX(2); default: - return Qnil; + return Qnil; } } @@ -2688,10 +3067,10 @@ ruby_set_argv(int argc, char **argv) rb_ary_clear(av); for (i = 0; i < argc; i++) { - VALUE arg = external_str_new_cstr(argv[i]); + VALUE arg = external_str_new_cstr(argv[i]); - OBJ_FREEZE(arg); - rb_ary_push(av, arg); + OBJ_FREEZE(arg); + rb_ary_push(av, arg); } } @@ -2702,19 +3081,26 @@ ruby_process_options(int argc, char **argv) VALUE iseq; const char *script_name = (argc > 0 && argv[0]) ? argv[0] : ruby_engine; + (*rb_ruby_prism_ptr()) = false; + if (!origarg.argv || origarg.argc <= 0) { - origarg.argc = argc; - origarg.argv = argv; + origarg.argc = argc; + origarg.argv = argv; } - ruby_script(script_name); /* for the time being */ + set_progname(external_str_new_cstr(script_name)); /* for the time being */ rb_argv0 = rb_str_new4(rb_progname); - rb_gc_register_mark_object(rb_argv0); - iseq = process_options(argc, argv, cmdline_options_init(&opt)); + rb_vm_register_global_object(rb_argv0); #ifndef HAVE_SETPROCTITLE ruby_init_setproctitle(argc, argv); #endif + iseq = process_options(argc, argv, cmdline_options_init(&opt)); + + if (opt.crash_report && *opt.crash_report) { + void ruby_set_crash_report(const char *template); + ruby_set_crash_report(opt.crash_report); + } return (void*)(struct RData*)iseq; } @@ -2755,8 +3141,8 @@ ruby_sysinit(int *argc, char ***argv) rb_w32_sysinit(argc, argv); #endif if (*argc >= 0 && *argv) { - origarg.argc = *argc; - origarg.argv = *argv; + origarg.argc = *argc; + origarg.argv = *argv; } fill_standard_fds(); } |