Put RRI into code.

9 files changed, 24 insertions, 191 deletions

diff --git a/ChangeLog b/ChangeLog
index 8507b985..6e655835 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,18 @@
+Fri Jun 17 10:55:27 2011  Arnold D. Robbins  <arnold@skeeve.com>
+
+	Implement Rational Range Interpretation (RRI) directly in code.
+
+	* regex.h [RE_RANGES_IGNORE_LOCALES]: Remove macro and its use.
+	* dfa.c (parse_bracket_exp): Remove use of RE_RANGES_IGNORE_LOCALES
+	and just do it in code.
+	(hard-locale.h): Remove include.
+	(hard_LC_COLLATE): Remove variable and its uses.
+	* re.c (resetup): Remove use of RE_RANGES_IGNORE_LOCALES.
+	* regcomp.c (build_range_exp): Remove use of RE_RANGES_IGNORE_LOCALES
+	and just do it in code. Remove cmp_buf array; it's no longer needed.
+	* Makefile.am (base_sources): Remove hard_locale.h and hard_locale.c.
+	* hard_locale.h, hard_locale.c: Removed from dist.
+
 Sun Jun 12 23:43:06 2011  Arnold D. Robbins  <arnold@skeeve.com>
 
 	* re.c (resetup): Always turn on RE_RANGES_IGNORE_LOCALES.
diff --git a/Makefile.am b/Makefile.am
index c89f6423..d08ff172 100644
--- a/Makefile.am
+++ b/Makefile.am
@@ -96,8 +96,6 @@ base_sources = \
 	getopt1.c \
 	getopt_int.h \
 	gettext.h \
-	hard-locale.h \
-	hard-locale.c \
 	io.c \
 	mbsupport.h \
 	main.c \
diff --git a/Makefile.in b/Makefile.in
index 6348d56b..4feab746 100644
--- a/Makefile.in
+++ b/Makefile.in
@@ -90,10 +90,9 @@ PROGRAMS = $(bin_PROGRAMS)
 am__objects_1 = array.$(OBJEXT) awkgram.$(OBJEXT) builtin.$(OBJEXT) \
 	dfa.$(OBJEXT) ext.$(OBJEXT) field.$(OBJEXT) \
 	floatcomp.$(OBJEXT) gawkmisc.$(OBJEXT) getopt.$(OBJEXT) \
-	getopt1.$(OBJEXT) hard-locale.$(OBJEXT) io.$(OBJEXT) \
-	main.$(OBJEXT) msg.$(OBJEXT) node.$(OBJEXT) random.$(OBJEXT) \
-	re.$(OBJEXT) regex.$(OBJEXT) replace.$(OBJEXT) \
-	version.$(OBJEXT)
+	getopt1.$(OBJEXT) io.$(OBJEXT) main.$(OBJEXT) msg.$(OBJEXT) \
+	node.$(OBJEXT) random.$(OBJEXT) re.$(OBJEXT) regex.$(OBJEXT) \
+	replace.$(OBJEXT) version.$(OBJEXT)
 am_dgawk_OBJECTS = $(am__objects_1) eval_d.$(OBJEXT) profile.$(OBJEXT) \
 	command.$(OBJEXT) debug.$(OBJEXT)
 dgawk_OBJECTS = $(am_dgawk_OBJECTS)
@@ -372,8 +371,6 @@ base_sources = \
 	getopt1.c \
 	getopt_int.h \
 	gettext.h \
-	hard-locale.h \
-	hard-locale.c \
 	io.c \
 	mbsupport.h \
 	main.c \
@@ -530,7 +527,6 @@ distclean-compile:
 @AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/gawkmisc.Po@am__quote@
 @AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/getopt.Po@am__quote@
 @AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/getopt1.Po@am__quote@
-@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/hard-locale.Po@am__quote@
 @AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/io.Po@am__quote@
 @AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/main.Po@am__quote@
 @AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/msg.Po@am__quote@
diff --git a/dfa.c b/dfa.c
index 2042dc34..f1557423 100644
--- a/dfa.c
+++ b/dfa.c
@@ -64,7 +64,6 @@
 #endif
 #endif
 
-/* need this before include of hard-locale.h */
 #ifdef GAWK
 #define bool int
 #define true (1)
@@ -73,7 +72,6 @@
 
 #include "regex.h"
 #include "dfa.h"
-#include "hard-locale.h"
 #include "xalloc.h"
 
 #ifdef GAWK
@@ -650,7 +648,6 @@ static int laststart;		/* True if we're separated from beginning or (, |
                                    only by zero-width characters. */
 static int parens;		/* Count of outstanding left parens. */
 static int minrep, maxrep;	/* Repeat counts for {m,n}. */
-static int hard_LC_COLLATE;	/* Nonzero if LC_COLLATE is hard.  */
 
 static int cur_mb_len = 1;	/* Length of the multibyte representation of
                                    wctok.  */
@@ -1007,29 +1004,8 @@ parse_bracket_exp (void)
                   c1 = tolower (c1);
                   c2 = tolower (c2);
                 }
-              if (!hard_LC_COLLATE
-                  || (syntax_bits & RE_RANGES_IGNORE_LOCALES))
-                for (c = c1; c <= c2; c++)
-                  setbit_case_fold_c (c, ccl);
-              else
-                {
-                  /* Defer to the system regex library about the meaning
-                     of range expressions.  */
-                  regex_t re;
-                  char pattern[6] = { '[', 0, '-', 0, ']', 0 };
-                  char subject[2] = { 0, 0 };
-		  pattern[1] = c1;
-		  pattern[3] = c2;
-                  regcomp (&re, pattern, REG_NOSUB);
-                  for (c = 0; c < NOTCHAR; ++c)
-                    {
-                      subject[0] = c;
-                      if (!(case_fold && isupper (c))
-                          && regexec (&re, subject, 0, NULL, 0) != REG_NOMATCH)
-                        setbit_case_fold_c (c, ccl);
-                    }
-                  regfree (&re);
-                }
+              for (c = c1; c <= c2; c++)
+                setbit_case_fold_c (c, ccl);
             }
 
           colon_warning_state |= 8;
@@ -1821,9 +1797,6 @@ dfaparse (char const *s, size_t len, struct dfa *d)
   lasttok = END;
   laststart = 1;
   parens = 0;
-#ifdef LC_COLLATE
-  hard_LC_COLLATE = hard_locale (LC_COLLATE);
-#endif
 #if MBS_SUPPORT
   if (MB_CUR_MAX > 1)
     {
diff --git a/hard-locale.c b/hard-locale.c
deleted file mode 100644
index 8b7353bb..00000000
--- a/hard-locale.c
+++ /dev/null
@@ -1,83 +0,0 @@
-/* hard-locale.c -- Determine whether a locale is hard.
-
-   Copyright (C) 1997, 1998, 1999, 2002, 2003, 2004, 2006, 2007, 2009, 2010,
-   2011,
-   Free Software Foundation, Inc.
-
-   This program is free software: you can redistribute it and/or modify
-   it under the terms of the GNU General Public License as published by
-   the Free Software Foundation; either version 3 of the License, or
-   (at your option) any later version.
-
-   This program is distributed in the hope that it will be useful,
-   but WITHOUT ANY WARRANTY; without even the implied warranty of
-   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-   GNU General Public License for more details.
-
-   You should have received a copy of the GNU General Public License
-   along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
-
-#include <config.h>
-
-#ifdef GAWK
-#define bool int
-#define true (1)
-#define false (0)
-#endif
-
-#include "hard-locale.h"
-
-#ifdef HAVE_LOCALE_H
-#include <locale.h>
-#endif
-#include <stdlib.h>
-#include <string.h>
-
-#ifdef __GLIBC__
-# define GLIBC_VERSION __GLIBC__
-#else
-# define GLIBC_VERSION 0
-#endif
-
-/* Return true if the current CATEGORY locale is hard, i.e. if you
-   can't get away with assuming traditional C or POSIX behavior.  */
-bool
-hard_locale (int category)
-{
-#if ! (defined ENABLE_NLS && HAVE_SETLOCALE)
-  return 0;
-#else
-  bool hard = true;
-  char const *p = setlocale (category, NULL);
-
-  if (p)
-    {
-      if (2 <= GLIBC_VERSION)
-        {
-          if (strcmp (p, "C") == 0 || strcmp (p, "POSIX") == 0)
-            hard = false;
-        }
-      else
-        {
-          char *locale = strdup (p);
-          if (locale)
-            {
-              /* Temporarily set the locale to the "C" and "POSIX" locales
-                 to find their names, so that we can determine whether one
-                 or the other is the caller's locale.  */
-              if (((p = setlocale (category, "C"))
-                   && strcmp (p, locale) == 0)
-                  || ((p = setlocale (category, "POSIX"))
-                      && strcmp (p, locale) == 0))
-                hard = false;
-
-              /* Restore the caller's locale.  */
-              setlocale (category, locale);
-              free (locale);
-            }
-        }
-    }
-
-  return hard;
-#endif
-}
diff --git a/hard-locale.h b/hard-locale.h
deleted file mode 100644
index 160d5443..00000000
--- a/hard-locale.h
+++ /dev/null
@@ -1,27 +0,0 @@
-/* Determine whether a locale is hard.
-
-   Copyright (C) 1999, 2003, 2004, 2009, 2010, 2011 Free Software Foundation, Inc.
-
-   This program is free software: you can redistribute it and/or modify
-   it under the terms of the GNU General Public License as published by
-   the Free Software Foundation; either version 3 of the License, or
-   (at your option) any later version.
-
-   This program is distributed in the hope that it will be useful,
-   but WITHOUT ANY WARRANTY; without even the implied warranty of
-   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-   GNU General Public License for more details.
-
-   You should have received a copy of the GNU General Public License
-   along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
-
-#ifndef HARD_LOCALE_H_
-# define HARD_LOCALE_H_ 1
-
-#ifndef GAWK
-# include <stdbool.h>
-#endif
-
-bool hard_locale (int);
-
-#endif /* HARD_LOCALE_H_ */
diff --git a/re.c b/re.c
index 2e1a37e7..234384b7 100644
--- a/re.c
+++ b/re.c
@@ -388,20 +388,6 @@ resetup()
 		syn = RE_SYNTAX_GNU_AWK;	/* POSIX re's + GNU ops */
 
 	/*
-	 * As of POSIX 1003.1-2008 (see rule 7 of 
-	 * http://pubs.opengroup.org/onlinepubs/9699919799/basedefs/V1_chap09.html#tag_09_03_05
-	 * and the rationale, at http://pubs.opengroup.org/onlinepubs/9699919799/xrat/V4_xbd_chap09.html#tag_21_09_03_05)
-	 * POSIX changed ranges outside the POSIX locale from requiring
-	 * Collation Element Order to being "undefined". This gives an
-	 * implementation, like gawk, the freedom to do ranges as it
-	 * pleases.
-	 *
-	 * We very much please to always use numeric ordering, as
-	 * the Good Lord intended.
-	 */
-	syn |= RE_RANGES_IGNORE_LOCALES;
-
-	/*
 	 * Interval expressions are now on by default, as POSIX is
 	 * wide-spread enough that people want it. The do_intervals
 	 * variable remains for use with --traditional.
diff --git a/regcomp.c b/regcomp.c
index 22c79cdb..a181d63f 100644
--- a/regcomp.c
+++ b/regcomp.c
@@ -2651,7 +2651,6 @@ build_range_exp (reg_syntax_t syntax, bitset_t sbcset,
 # endif /* not RE_ENABLE_I18N */
 {
   unsigned int start_ch, end_ch;
-  int ignore_locales = (syntax & RE_RANGES_IGNORE_LOCALES) != 0;
 
   /* Equivalence Classes and Character Classes can't be a range start/end.  */
   if (BE (start_elem->type == EQUIV_CLASS || start_elem->type == CHAR_CLASS
@@ -2672,7 +2671,6 @@ build_range_exp (reg_syntax_t syntax, bitset_t sbcset,
     wchar_t wc;
     wint_t start_wc;
     wint_t end_wc;
-    wchar_t cmp_buf[6] = {L'\0', L'\0', L'\0', L'\0', L'\0', L'\0'};
 
     start_ch = ((start_elem->type == SB_CHAR) ? start_elem->opr.ch
 		: ((start_elem->type == COLL_SYM) ? start_elem->opr.name[0]
@@ -2698,12 +2696,7 @@ build_range_exp (reg_syntax_t syntax, bitset_t sbcset,
 #endif
     if (start_wc == WEOF || end_wc == WEOF)
       return REG_ECOLLATE;
-    cmp_buf[0] = start_wc;
-    cmp_buf[4] = end_wc;
-    if (ignore_locales && start_wc > end_wc)
-      return REG_ERANGE;
-    else if ((syntax & RE_NO_EMPTY_RANGES)
-             && wcscoll (cmp_buf, cmp_buf + 4) > 0)
+    else if ((syntax & RE_NO_EMPTY_RANGES) && start_wc > end_wc)
       return REG_ERANGE;
 
     /* Got valid collation sequence values, add them as a new entry.
@@ -2742,23 +2735,10 @@ build_range_exp (reg_syntax_t syntax, bitset_t sbcset,
       }
 
     /* Build the table for single byte characters.  */
-    if (ignore_locales)
-      {
-        for (wc = 0; wc < SBC_MAX; ++wc)
-          {
-    	     if (start_wc <= wc && wc <= end_wc)
-    	       bitset_set (sbcset, wc);
-          }
-      }
-    else
+    for (wc = 0; wc < SBC_MAX; ++wc)
       {
-        for (wc = 0; wc < SBC_MAX; ++wc)
-          {
-    	     cmp_buf[2] = wc;
-    	     if (wcscoll (cmp_buf, cmp_buf + 2) <= 0
-    	         && wcscoll (cmp_buf + 2, cmp_buf + 4) <= 0)
-    	       bitset_set (sbcset, wc);
-          }
+         if (start_wc <= wc && wc <= end_wc)
+           bitset_set (sbcset, wc);
       }
   }
 # else /* not RE_ENABLE_I18N */
diff --git a/regex.h b/regex.h
index a2d120f9..6bc503b2 100644
--- a/regex.h
+++ b/regex.h
@@ -184,10 +184,6 @@ typedef unsigned long int reg_syntax_t;
 /* If this bit is set, then no_sub will be set to 1 during
    re_compile_pattern.  */
 # define RE_NO_SUB (RE_CONTEXT_INVALID_DUP << 1)
-
-/* If this bit is set, then ranges act like they are in
-   the "C" locale.  */
-# define RE_RANGES_IGNORE_LOCALES (RE_NO_SUB << 1)
 #endif
 
 /* This global variable defines the particular regexp syntax to use (for
@@ -213,7 +209,6 @@ extern reg_syntax_t re_syntax_options;
 
 #define RE_SYNTAX_GNU_AWK						\
   ((RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS		\
-    | RE_RANGES_IGNORE_LOCALES                                          \
     | RE_INVALID_INTERVAL_ORD)						\
    & ~(RE_DOT_NOT_NULL | RE_CONTEXT_INDEP_OPS				\
       | RE_CONTEXT_INVALID_OPS ))