Merge branch 'master' into select

1 files changed, 307 insertions, 281 deletions

diff --git a/main.c b/main.c
index 44021e50..93f94998 100644
--- a/main.c
+++ b/main.c
@@ -142,17 +142,20 @@ static bool disallow_var_assigns = false;	/* true for --exec */
 
 static void add_preassign(enum assign_type type, char *val);
 
+static void parse_args(int argc, char **argv);
+static void set_locale_stuff(void);
+static bool stopped_early = false;
+
 int do_flags = false;
 bool do_optimize = false;		/* apply default optimizations */
 static int do_nostalgia = false;	/* provide a blast from the past */
 static int do_binary = false;		/* hands off my data! */
 static int do_version = false;		/* print version info */
+static const char *locale = "";		/* default value to setlocale */
 
 int use_lc_numeric = false;	/* obey locale for decimal point */
 
-#if MBS_SUPPORT
 int gawk_mb_cur_max;		/* MB_CUR_MAX value, see comment in main() */
-#endif
 
 FILE *output_fp;		/* default gawk output, can be redirected in the debugger */
 bool output_is_tty = false;	/* control flushing of output */
@@ -186,6 +189,9 @@ static const struct option optab[] = {
 	{ "lint",		optional_argument,	NULL,	'L' },
 	{ "lint-old",		no_argument,		NULL,	't' },
 	{ "load",		required_argument,	NULL,	'l' },
+#if defined(LOCALEDEBUG)
+	{ "locale",		required_argument,	NULL,	'Z' },
+#endif
 	{ "non-decimal-data",	no_argument,		NULL,	'n' },
 	{ "nostalgia",		no_argument,		& do_nostalgia,	1 },
 	{ "optimize",		no_argument,		NULL,	'O' },
@@ -209,15 +215,7 @@ static const struct option optab[] = {
 int
 main(int argc, char **argv)
 {
-	/*
-	 * The + on the front tells GNU getopt not to rearrange argv.
-	 */
-	const char *optlist = "+F:f:v:W;bcCd::D::e:E:ghi:l:L:nNo::Op::MPrStVY";
-	bool stopped_early = false;
-	int old_optind;
 	int i;
-	int c;
-	char *scan, *src;
 	char *extra_stack;
 	int have_srcfile = 0;
 	SRCFILE *s;
@@ -233,49 +231,11 @@ main(int argc, char **argv)
 #endif /* HAVE_MTRACE */
 #endif /* HAVE_MCHECK_H */
 
-#if defined(LC_CTYPE)
-	setlocale(LC_CTYPE, "");
-#endif
-#if defined(LC_COLLATE)
-	setlocale(LC_COLLATE, "");
-#endif
-#if defined(LC_MESSAGES)
-	setlocale(LC_MESSAGES, "");
-#endif
-#if defined(LC_NUMERIC) && defined(HAVE_LOCALE_H)
-	/*
-	 * Force the issue here.  According to POSIX 2001, decimal
-	 * point is used for parsing source code and for command-line
-	 * assignments and the locale value for processing input,
-	 * number to string conversion, and printing output.
-	 *
-	 * 10/2005 --- see below also; we now only use the locale's
-	 * decimal point if do_posix in effect.
-	 *
-	 * 9/2007:
-	 * This is a mess. We need to get the locale's numeric info for
-	 * the thousands separator for the %'d flag.
-	 */
-	setlocale(LC_NUMERIC, "");
-	init_locale(& loc);
-	setlocale(LC_NUMERIC, "C");
-#endif
-#if defined(LC_TIME)
-	setlocale(LC_TIME, "");
-#endif
-
-#if MBS_SUPPORT
-	/*
-	 * In glibc, MB_CUR_MAX is actually a function.  This value is
-	 * tested *a lot* in many speed-critical places in gawk. Caching
-	 * this value once makes a speed difference.
-	 */
-	gawk_mb_cur_max = MB_CUR_MAX;
-	/* Without MBS_SUPPORT, gawk_mb_cur_max is 1. */
+	myname = gawk_name(argv[0]);
+	os_arg_fixup(&argc, &argv); /* emulate redirection, expand wildcards */
 
-	/* init the cache for checking bytes if they're characters */
-	init_btowc_cache();
-#endif
+	if (argc < 2)
+		usage(EXIT_FAILURE, stderr);
 
 	(void) bindtextdomain(PACKAGE, LOCALEDIR);
 	(void) textdomain(PACKAGE);
@@ -307,12 +267,6 @@ main(int argc, char **argv)
 	(void) stackoverflow_install_handler(catchstackoverflow, extra_stack, STACK_SIZE);
 #undef STACK_SIZE
 
-	myname = gawk_name(argv[0]);
-	os_arg_fixup(&argc, &argv); /* emulate redirection, expand wildcards */
-
-	if (argc < 2)
-		usage(EXIT_FAILURE, stderr);
-
 	/* initialize the null string */
 	Nnull_string = make_string("", 0);
 
@@ -327,230 +281,22 @@ main(int argc, char **argv)
 
 	output_fp = stdout;
 
-	/* we do error messages ourselves on invalid options */
-	opterr = false;
-
-	/* copy argv before getopt gets to it; used to restart the debugger */  
-	save_argv(argc, argv);
-
 	/* initialize global (main) execution context */
 	push_context(new_context());
 
-	/* option processing. ready, set, go! */
-	for (optopt = 0, old_optind = 1;
-	     (c = getopt_long(argc, argv, optlist, optab, NULL)) != EOF;
-	     optopt = 0, old_optind = optind) {
-		if (do_posix)
-			opterr = true;
-
-		switch (c) {
-		case 'F':
-			add_preassign(PRE_ASSIGN_FS, optarg);
-			break;
-
-		case 'E':
-			disallow_var_assigns = true;
-			/* fall through */
-		case 'f':
-			/*
-			 * Allow multiple -f options.
-			 * This makes function libraries real easy.
-			 * Most of the magic is in the scanner.
-			 *
-			 * The following is to allow for whitespace at the end
-			 * of a #! /bin/gawk line in an executable file
-			 */
-			scan = optarg;
-			if (argv[optind-1] != optarg)
-				while (isspace((unsigned char) *scan))
-					scan++;
-			src = (*scan == '\0' ? argv[optind++] : optarg);
-			(void) add_srcfile((src && src[0] == '-' && src[1] == '\0') ?
-					SRC_STDIN : SRC_FILE,
-					src, srcfiles, NULL, NULL);
-
-			break;
-
-		case 'v':
-			add_preassign(PRE_ASSIGN, optarg);
-			break;
-
-		case 'b':
-			do_binary = true;
-			break;
-
-		case 'c':
-			do_flags |= DO_TRADITIONAL;
-			break;
-
-		case 'C':
-			copyleft();
-			break;
-
-		case 'd':
-			do_flags |= DO_DUMP_VARS;
-			if (optarg != NULL && optarg[0] != '\0')
-				varfile = optarg;
-			break;
-
-		case 'D':
-			do_flags |= DO_DEBUG;
-			if (optarg != NULL && optarg[0] != '\0')
-				command_file = optarg;
-			break;
-
-		case 'e':
-			if (optarg[0] == '\0')
-				warning(_("empty argument to `-e/--source' ignored"));
-			else
-				(void) add_srcfile(SRC_CMDLINE, optarg, srcfiles, NULL, NULL);
-			break;
-
-		case 'g':
-			do_flags |= DO_INTL;
-			break;
-
-		case 'h':
-			/* write usage to stdout, per GNU coding stds */
-			usage(EXIT_SUCCESS, stdout);
-			break;
-
-		case 'i':
-			(void) add_srcfile(SRC_INC, optarg, srcfiles, NULL, NULL);
-			break;
+	parse_args(argc, argv);
 
-		case 'l':
-			(void) add_srcfile(SRC_EXTLIB, optarg, srcfiles, NULL, NULL);
-			break;
+	set_locale_stuff();
 
-#ifndef NO_LINT
-		case 'L':
-			do_flags |= DO_LINT_ALL;
-			if (optarg != NULL) {
-				if (strcmp(optarg, "fatal") == 0)
-					lintfunc = r_fatal;
-				else if (strcmp(optarg, "invalid") == 0) {
-					do_flags &= ~DO_LINT_ALL;
-					do_flags |= DO_LINT_INVALID;
-				}
-			}
-			break;
-
-		case 't':
-			do_flags |= DO_LINT_OLD;
-			break;
-#else
-		case 'L':
-		case 't':
-			break;
-#endif
-
-		case 'n':
-			do_flags |= DO_NON_DEC_DATA;
-			break;
-
-		case 'N':
-			use_lc_numeric = true;
-			break;
-
-		case 'O':
-			do_optimize = true;
-			break;
-
-		case 'p':
-			do_flags |= DO_PROFILE;
-			/* fall through */
-		case 'o':
-			do_flags |= DO_PRETTY_PRINT;
-			if (optarg != NULL)
-				set_prof_file(optarg);
-			else
-				set_prof_file(DEFAULT_PROFILE);
-			break;
-
-		case 'M':
-#ifdef HAVE_MPFR
-			do_flags |= DO_MPFR;
-#else
-			warning(_("-M ignored: MPFR/GMP support not compiled in"));
-#endif
-			break;
-
-		case 'P':
-			do_flags |= DO_POSIX;
-			break;
-
-		case 'r':
-			do_flags |= DO_INTERVALS;
- 			break;
- 
-		case 'S':
-			do_flags |= DO_SANDBOX;
-  			break;
-
-		case 'V':
-			do_version = true;
-			break;
-
-		case 'W':       /* gawk specific options - now in getopt_long */
-			fprintf(stderr, _("%s: option `-W %s' unrecognized, ignored\n"),
-				argv[0], optarg);
-			break;
-
-		case 0:
-			/*
-			 * getopt_long found an option that sets a variable
-			 * instead of returning a letter. Do nothing, just
-			 * cycle around for the next one.
-			 */
-			break;
+	/*
+	 * In glibc, MB_CUR_MAX is actually a function.  This value is
+	 * tested *a lot* in many speed-critical places in gawk. Caching
+	 * this value once makes a speed difference.
+	 */
+	gawk_mb_cur_max = MB_CUR_MAX;
 
-		case 'Y':
-#if defined(YYDEBUG) || defined(GAWKDEBUG)
-			if (c == 'Y') {
-				yydebug = 2;
-				break;
-			}
-#endif
-			/* if not debugging, fall through */
-		case '?':
-		default:
-			/*
-			 * If not posix, an unrecognized option stops argument
-			 * processing so that it can go into ARGV for the awk
-			 * program to see. This makes use of ``#! /bin/gawk -f''
-			 * easier.
-			 *
-			 * However, it's never simple. If optopt is set,
-			 * an option that requires an argument didn't get the
-			 * argument. We care because if opterr is 0, then
-			 * getopt_long won't print the error message for us.
-			 */
-			if (! do_posix
-			    && (optopt == '\0' || strchr(optlist, optopt) == NULL)) {
-				/*
-				 * can't just do optind--. In case of an
-				 * option with >= 2 letters, getopt_long
-				 * won't have incremented optind.
-				 */
-				optind = old_optind;
-				stopped_early = true;
-				goto out;
-			} else if (optopt != '\0') {
-				/* Use POSIX required message format */
-				fprintf(stderr,
-					_("%s: option requires an argument -- %c\n"),
-					myname, optopt);
-				usage(EXIT_FAILURE, stderr);
-			}
-			/* else
-				let getopt print error message for us */
-			break;
-		}
-		if (c == 'E')	/* --exec ends option processing */
-			break;
-	}
-out:
+	/* init the cache for checking bytes if they're characters */
+	init_btowc_cache();
 
 	if (do_nostalgia)
 		nostalgia();
@@ -583,7 +329,6 @@ out:
 	if (do_lint && os_is_setuid())
 		warning(_("running %s setuid root may be a security problem"), myname);
 
-#if MBS_SUPPORT
 	if (do_binary) {
 		if (do_posix)
 			warning(_("`--posix' overrides `--characters-as-bytes'"));
@@ -593,7 +338,6 @@ out:
 		setlocale(LC_ALL, "C");
 #endif
 	}
-#endif
 
 	if (do_debug)	/* Need to register the debugger pre-exec hook before any other */
 		init_debug();
@@ -734,9 +478,9 @@ out:
 	 * data using the local decimal point.
 	 */
 	if (use_lc_numeric)
-		setlocale(LC_NUMERIC, "");
+		setlocale(LC_NUMERIC, locale);
 #endif
-	
+
 	init_io();
 	output_fp = stdout;
 
@@ -1379,7 +1123,7 @@ arg_assign(char *arg, bool initing)
 			setlocale(LC_NUMERIC, "C");
 		(void) force_number(it);
 		if (do_posix)
-			setlocale(LC_NUMERIC, "");
+			setlocale(LC_NUMERIC, locale);
 #endif /* LC_NUMERIC */
 
 		/*
@@ -1643,3 +1387,285 @@ getenv_long(const char *name)
 	}
 	return -1;
 }
+
+/* parse_args --- do the getopt_long thing */
+
+static void
+parse_args(int argc, char **argv)
+{
+	/*
+	 * The + on the front tells GNU getopt not to rearrange argv.
+	 */
+	const char *optlist = "+F:f:v:W;bcCd::D::e:E:ghi:l:L:nNo::Op::MPrStVYZ:";
+	int old_optind;
+	int c;
+	char *scan;
+	char *src;
+
+	/* we do error messages ourselves on invalid options */
+	opterr = false;
+
+	/* copy argv before getopt gets to it; used to restart the debugger */  
+	save_argv(argc, argv);
+
+	/* option processing. ready, set, go! */
+	for (optopt = 0, old_optind = 1;
+	     (c = getopt_long(argc, argv, optlist, optab, NULL)) != EOF;
+	     optopt = 0, old_optind = optind) {
+		if (do_posix)
+			opterr = true;
+
+		switch (c) {
+		case 'F':
+			add_preassign(PRE_ASSIGN_FS, optarg);
+			break;
+
+		case 'E':
+			disallow_var_assigns = true;
+			/* fall through */
+		case 'f':
+			/*
+			 * Allow multiple -f options.
+			 * This makes function libraries real easy.
+			 * Most of the magic is in the scanner.
+			 *
+			 * The following is to allow for whitespace at the end
+			 * of a #! /bin/gawk line in an executable file
+			 */
+			scan = optarg;
+			if (argv[optind-1] != optarg)
+				while (isspace((unsigned char) *scan))
+					scan++;
+			src = (*scan == '\0' ? argv[optind++] : optarg);
+			(void) add_srcfile((src && src[0] == '-' && src[1] == '\0') ?
+					SRC_STDIN : SRC_FILE,
+					src, srcfiles, NULL, NULL);
+
+			break;
+
+		case 'v':
+			add_preassign(PRE_ASSIGN, optarg);
+			break;
+
+		case 'b':
+			do_binary = true;
+			break;
+
+		case 'c':
+			do_flags |= DO_TRADITIONAL;
+			break;
+
+		case 'C':
+			copyleft();
+			break;
+
+		case 'd':
+			do_flags |= DO_DUMP_VARS;
+			if (optarg != NULL && optarg[0] != '\0')
+				varfile = optarg;
+			break;
+
+		case 'D':
+			do_flags |= DO_DEBUG;
+			if (optarg != NULL && optarg[0] != '\0')
+				command_file = optarg;
+			break;
+
+		case 'e':
+			if (optarg[0] == '\0')
+				warning(_("empty argument to `-e/--source' ignored"));
+			else
+				(void) add_srcfile(SRC_CMDLINE, optarg, srcfiles, NULL, NULL);
+			break;
+
+		case 'g':
+			do_flags |= DO_INTL;
+			break;
+
+		case 'h':
+			/* write usage to stdout, per GNU coding stds */
+			usage(EXIT_SUCCESS, stdout);
+			break;
+
+		case 'i':
+			(void) add_srcfile(SRC_INC, optarg, srcfiles, NULL, NULL);
+			break;
+
+		case 'l':
+			(void) add_srcfile(SRC_EXTLIB, optarg, srcfiles, NULL, NULL);
+			break;
+
+#ifndef NO_LINT
+		case 'L':
+			do_flags |= DO_LINT_ALL;
+			if (optarg != NULL) {
+				if (strcmp(optarg, "fatal") == 0)
+					lintfunc = r_fatal;
+				else if (strcmp(optarg, "invalid") == 0) {
+					do_flags &= ~DO_LINT_ALL;
+					do_flags |= DO_LINT_INVALID;
+				}
+			}
+			break;
+
+		case 't':
+			do_flags |= DO_LINT_OLD;
+			break;
+#else
+		case 'L':
+		case 't':
+			break;
+#endif
+
+		case 'n':
+			do_flags |= DO_NON_DEC_DATA;
+			break;
+
+		case 'N':
+			use_lc_numeric = true;
+			break;
+
+		case 'O':
+			do_optimize = true;
+			break;
+
+		case 'p':
+			do_flags |= DO_PROFILE;
+			/* fall through */
+		case 'o':
+			do_flags |= DO_PRETTY_PRINT;
+			if (optarg != NULL)
+				set_prof_file(optarg);
+			else
+				set_prof_file(DEFAULT_PROFILE);
+			break;
+
+		case 'M':
+#ifdef HAVE_MPFR
+			do_flags |= DO_MPFR;
+#else
+			warning(_("-M ignored: MPFR/GMP support not compiled in"));
+#endif
+			break;
+
+		case 'P':
+			do_flags |= DO_POSIX;
+			break;
+
+		case 'r':
+			do_flags |= DO_INTERVALS;
+ 			break;
+ 
+		case 'S':
+			do_flags |= DO_SANDBOX;
+  			break;
+
+		case 'V':
+			do_version = true;
+			break;
+
+		case 'W':       /* gawk specific options - now in getopt_long */
+			fprintf(stderr, _("%s: option `-W %s' unrecognized, ignored\n"),
+				argv[0], optarg);
+			break;
+
+		case 0:
+			/*
+			 * getopt_long found an option that sets a variable
+			 * instead of returning a letter. Do nothing, just
+			 * cycle around for the next one.
+			 */
+			break;
+
+		case 'Y':
+		case 'Z':
+#if defined(YYDEBUG) || defined(GAWKDEBUG)
+			if (c == 'Y') {
+				yydebug = 2;
+				break;
+			}
+#endif
+#if defined(LOCALEDEBUG)
+			if (c == 'Z') {
+				locale = optarg;
+				break;
+			}
+#endif
+			/* if not debugging, fall through */
+		case '?':
+		default:
+			/*
+			 * If not posix, an unrecognized option stops argument
+			 * processing so that it can go into ARGV for the awk
+			 * program to see. This makes use of ``#! /bin/gawk -f''
+			 * easier.
+			 *
+			 * However, it's never simple. If optopt is set,
+			 * an option that requires an argument didn't get the
+			 * argument. We care because if opterr is 0, then
+			 * getopt_long won't print the error message for us.
+			 */
+			if (! do_posix
+			    && (optopt == '\0' || strchr(optlist, optopt) == NULL)) {
+				/*
+				 * can't just do optind--. In case of an
+				 * option with >= 2 letters, getopt_long
+				 * won't have incremented optind.
+				 */
+				optind = old_optind;
+				stopped_early = true;
+				goto out;
+			} else if (optopt != '\0') {
+				/* Use POSIX required message format */
+				fprintf(stderr,
+					_("%s: option requires an argument -- %c\n"),
+					myname, optopt);
+				usage(EXIT_FAILURE, stderr);
+			}
+			/* else
+				let getopt print error message for us */
+			break;
+		}
+		if (c == 'E')	/* --exec ends option processing */
+			break;
+	}
+out:
+	return;
+}
+
+/* set_locale_stuff --- setup the locale stuff */
+
+static void
+set_locale_stuff(void)
+{
+#if defined(LC_CTYPE)
+	setlocale(LC_CTYPE, locale);
+#endif
+#if defined(LC_COLLATE)
+	setlocale(LC_COLLATE, locale);
+#endif
+#if defined(LC_MESSAGES)
+	setlocale(LC_MESSAGES, locale);
+#endif
+#if defined(LC_NUMERIC) && defined(HAVE_LOCALE_H)
+	/*
+	 * Force the issue here.  According to POSIX 2001, decimal
+	 * point is used for parsing source code and for command-line
+	 * assignments and the locale value for processing input,
+	 * number to string conversion, and printing output.
+	 *
+	 * 10/2005 --- see below also; we now only use the locale's
+	 * decimal point if do_posix in effect.
+	 *
+	 * 9/2007:
+	 * This is a mess. We need to get the locale's numeric info for
+	 * the thousands separator for the %'d flag.
+	 */
+	setlocale(LC_NUMERIC, locale);
+	init_locale(& loc);
+	setlocale(LC_NUMERIC, "C");
+#endif
+#if defined(LC_TIME)
+	setlocale(LC_TIME, locale);
+#endif
+}