From: Pekka Riikonen Date: Sun, 24 Apr 2005 14:29:47 +0000 (+0000) Subject: Removed getopt routines. X-Git-Tag: silc.server.0.9.20~24 X-Git-Url: http://git.silcnet.org/gitweb/?p=silc.git;a=commitdiff_plain;h=87f1d52fe18602dfac152e95f37eafdf8683af83 Removed getopt routines. Removed GNU regex and added free GNU compatible regex with POSIX regex API. --- diff --git a/lib/contrib/Makefile.am b/lib/contrib/Makefile.am index 74faaf72..26ff5751 100644 --- a/lib/contrib/Makefile.am +++ b/lib/contrib/Makefile.am @@ -19,16 +19,10 @@ AUTOMAKE_OPTIONS = 1.0 no-dependencies foreign noinst_LTLIBRARIES = libcontrib.la -if HAVE_GETOPT_LONG -GETOPT = -else -GETOPT = getopt.c getopt1.c -endif - if HAVE_REGEX REGEX = else -REGEX = regex.c +REGEX = regexpr.c endif STRINGPREP = nfkc.c rfc3454.c stringprep.c @@ -36,7 +30,7 @@ STRINGPREP = nfkc.c rfc3454.c stringprep.c if SILC_WIN32 libcontrib_la_SOURCES = $(STRINGPREP) else -libcontrib_la_SOURCES = $(STRINGPREP) $(GETOPT) $(REGEX) +libcontrib_la_SOURCES = $(STRINGPREP) $(REGEX) endif EXTRA_DIST = *.c *.h diff --git a/lib/contrib/getopt.c b/lib/contrib/getopt.c deleted file mode 100644 index 0b47ea7e..00000000 --- a/lib/contrib/getopt.c +++ /dev/null @@ -1,748 +0,0 @@ -/* Getopt for GNU. - NOTE: getopt is now part of the C library, so if you don't know what - "Keep this file name-space clean" means, talk to roland@gnu.ai.mit.edu - before changing it! - - Copyright (C) 1987, 88, 89, 90, 91, 92, 1993 - 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 2, 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, write to the Free Software - Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ - - -#ifndef __STDC__ -/* This is a separate conditional since some stdc systems - reject `defined (const)'. */ -#ifndef const -#define const -#endif -#endif - -/* This tells Alpha OSF/1 not to define a getopt prototype in . */ -#ifndef _NO_PROTO -#define _NO_PROTO -#endif - -#include -#include - -/* Comment out all this code if we are using the GNU C Library, and are not - actually compiling the library itself. This code is part of the GNU C - Library, but also included in many other GNU distributions. Compiling - and linking in this code is a waste when using the GNU C library - (especially if it is a shared library). Rather than having every GNU - program understand `configure --with-gnu-libc' and omit the object files, - it is simpler to just do this in the source for each such file. */ - -#if defined (_LIBC) || !defined (__GNU_LIBRARY__) - - -/* This needs to come after some library #include - to get __GNU_LIBRARY__ defined. */ -#ifdef __GNU_LIBRARY__ -/* Don't include stdlib.h for non-GNU C libraries because some of them - contain conflicting prototypes for getopt. */ -#include -#endif /* GNU C library. */ - -/* If GETOPT_COMPAT is defined, `+' as well as `--' can introduce a - long-named option. Because this is not POSIX.2 compliant, it is - being phased out. */ -/* #define GETOPT_COMPAT */ - -/* This version of `getopt' appears to the caller like standard Unix `getopt' - but it behaves differently for the user, since it allows the user - to intersperse the options with the other arguments. - - As `getopt' works, it permutes the elements of ARGV so that, - when it is done, all the options precede everything else. Thus - all application programs are extended to handle flexible argument order. - - Setting the environment variable POSIXLY_CORRECT disables permutation. - Then the behavior is completely standard. - - GNU application programs can use a third alternative mode in which - they can distinguish the relative order of options and other arguments. */ - -#include "getopt.h" - -/* For communication from `getopt' to the caller. - When `getopt' finds an option that takes an argument, - the argument value is returned here. - Also, when `ordering' is RETURN_IN_ORDER, - each non-option ARGV-element is returned here. */ - -char *optarg = 0; - -/* Index in ARGV of the next element to be scanned. - This is used for communication to and from the caller - and for communication between successive calls to `getopt'. - - On entry to `getopt', zero means this is the first call; initialize. - - When `getopt' returns EOF, this is the index of the first of the - non-option elements that the caller should itself scan. - - Otherwise, `optind' communicates from one call to the next - how much of ARGV has been scanned so far. */ - -/* XXX 1003.2 says this must be 1 before any call. */ -int optind = 0; - -/* The next char to be scanned in the option-element - in which the last option character we returned was found. - This allows us to pick up the scan where we left off. - - If this is zero, or a null string, it means resume the scan - by advancing to the next ARGV-element. */ - -static char *nextchar; - -/* Callers store zero here to inhibit the error message - for unrecognized options. */ - -int opterr = 1; - -/* Set to an option character which was unrecognized. - This must be initialized on some systems to avoid linking in the - system's own getopt implementation. */ - -int optopt = '?'; - -/* Describe how to deal with options that follow non-option ARGV-elements. - - If the caller did not specify anything, - the default is REQUIRE_ORDER if the environment variable - POSIXLY_CORRECT is defined, PERMUTE otherwise. - - REQUIRE_ORDER means don't recognize them as options; - stop option processing when the first non-option is seen. - This is what Unix does. - This mode of operation is selected by either setting the environment - variable POSIXLY_CORRECT, or using `+' as the first character - of the list of option characters. - - PERMUTE is the default. We permute the contents of ARGV as we scan, - so that eventually all the non-options are at the end. This allows options - to be given in any order, even with programs that were not written to - expect this. - - RETURN_IN_ORDER is an option available to programs that were written - to expect options and other ARGV-elements in any order and that care about - the ordering of the two. We describe each non-option ARGV-element - as if it were the argument of an option with character code 1. - Using `-' as the first character of the list of option characters - selects this mode of operation. - - The special argument `--' forces an end of option-scanning regardless - of the value of `ordering'. In the case of RETURN_IN_ORDER, only - `--' can cause `getopt' to return EOF with `optind' != ARGC. */ - -static enum -{ - REQUIRE_ORDER, PERMUTE, RETURN_IN_ORDER -} ordering; - -#ifdef __GNU_LIBRARY__ -/* We want to avoid inclusion of string.h with non-GNU libraries - because there are many ways it can cause trouble. - On some systems, it contains special magic macros that don't work - in GCC. */ -#include -#define my_index strchr -#else - -/* Avoid depending on library functions or files - whose names are inconsistent. */ - -char *getenv (); - -static char * -my_index (str, chr) - const char *str; - int chr; -{ - while (*str) - { - if (*str == chr) - return (char *) str; - str++; - } - return 0; -} - -/* If using GCC, we can safely declare strlen this way. - If not using GCC, it is ok not to declare it. - (Supposedly there are some machines where it might get a warning, - but changing this conditional to __STDC__ is too risky.) */ -#ifdef __GNUC__ -#ifdef IN_GCC -#include "gstddef.h" -#else -#include -#endif -extern size_t strlen (const char *); -#endif - -#endif /* GNU C library. */ - -/* Handle permutation of arguments. */ - -/* Describe the part of ARGV that contains non-options that have - been skipped. `first_nonopt' is the index in ARGV of the first of them; - `last_nonopt' is the index after the last of them. */ - -static int first_nonopt; -static int last_nonopt; - -/* Exchange two adjacent subsequences of ARGV. - One subsequence is elements [first_nonopt,last_nonopt) - which contains all the non-options that have been skipped so far. - The other is elements [last_nonopt,optind), which contains all - the options processed since those non-options were skipped. - - `first_nonopt' and `last_nonopt' are relocated so that they describe - the new indices of the non-options in ARGV after they are moved. */ - -static void -exchange (argv) - char **argv; -{ - int bottom = first_nonopt; - int middle = last_nonopt; - int top = optind; - char *tem; - - /* Exchange the shorter segment with the far end of the longer segment. - That puts the shorter segment into the right place. - It leaves the longer segment in the right place overall, - but it consists of two parts that need to be swapped next. */ - - while (top > middle && middle > bottom) - { - if (top - middle > middle - bottom) - { - /* Bottom segment is the short one. */ - int len = middle - bottom; - register int i; - - /* Swap it with the top part of the top segment. */ - for (i = 0; i < len; i++) - { - tem = argv[bottom + i]; - argv[bottom + i] = argv[top - (middle - bottom) + i]; - argv[top - (middle - bottom) + i] = tem; - } - /* Exclude the moved bottom segment from further swapping. */ - top -= len; - } - else - { - /* Top segment is the short one. */ - int len = top - middle; - register int i; - - /* Swap it with the bottom part of the bottom segment. */ - for (i = 0; i < len; i++) - { - tem = argv[bottom + i]; - argv[bottom + i] = argv[middle + i]; - argv[middle + i] = tem; - } - /* Exclude the moved top segment from further swapping. */ - bottom += len; - } - } - - /* Update records for the slots the non-options now occupy. */ - - first_nonopt += (optind - last_nonopt); - last_nonopt = optind; -} - -/* Scan elements of ARGV (whose length is ARGC) for option characters - given in OPTSTRING. - - If an element of ARGV starts with '-', and is not exactly "-" or "--", - then it is an option element. The characters of this element - (aside from the initial '-') are option characters. If `getopt' - is called repeatedly, it returns successively each of the option characters - from each of the option elements. - - If `getopt' finds another option character, it returns that character, - updating `optind' and `nextchar' so that the next call to `getopt' can - resume the scan with the following option character or ARGV-element. - - If there are no more option characters, `getopt' returns `EOF'. - Then `optind' is the index in ARGV of the first ARGV-element - that is not an option. (The ARGV-elements have been permuted - so that those that are not options now come last.) - - OPTSTRING is a string containing the legitimate option characters. - If an option character is seen that is not listed in OPTSTRING, - return '?' after printing an error message. If you set `opterr' to - zero, the error message is suppressed but we still return '?'. - - If a char in OPTSTRING is followed by a colon, that means it wants an arg, - so the following text in the same ARGV-element, or the text of the following - ARGV-element, is returned in `optarg'. Two colons mean an option that - wants an optional arg; if there is text in the current ARGV-element, - it is returned in `optarg', otherwise `optarg' is set to zero. - - If OPTSTRING starts with `-' or `+', it requests different methods of - handling the non-option ARGV-elements. - See the comments about RETURN_IN_ORDER and REQUIRE_ORDER, above. - - Long-named options begin with `--' instead of `-'. - Their names may be abbreviated as long as the abbreviation is unique - or is an exact match for some defined option. If they have an - argument, it follows the option name in the same ARGV-element, separated - from the option name by a `=', or else the in next ARGV-element. - When `getopt' finds a long-named option, it returns 0 if that option's - `flag' field is nonzero, the value of the option's `val' field - if the `flag' field is zero. - - The elements of ARGV aren't really const, because we permute them. - But we pretend they're const in the prototype to be compatible - with other systems. - - LONGOPTS is a vector of `struct option' terminated by an - element containing a name which is zero. - - LONGIND returns the index in LONGOPT of the long-named option found. - It is only valid when a long-named option has been found by the most - recent call. - - If LONG_ONLY is nonzero, '-' as well as '--' can introduce - long-named options. */ - -int -_getopt_internal (argc, argv, optstring, longopts, longind, long_only) - int argc; - char *const *argv; - const char *optstring; - const struct option *longopts; - int *longind; - int long_only; -{ - int option_index; - - optarg = 0; - - /* Initialize the internal data when the first call is made. - Start processing options with ARGV-element 1 (since ARGV-element 0 - is the program name); the sequence of previously skipped - non-option ARGV-elements is empty. */ - - if (optind == 0) - { - first_nonopt = last_nonopt = optind = 1; - - nextchar = NULL; - - /* Determine how to handle the ordering of options and nonoptions. */ - - if (optstring[0] == '-') - { - ordering = RETURN_IN_ORDER; - ++optstring; - } - else if (optstring[0] == '+') - { - ordering = REQUIRE_ORDER; - ++optstring; - } - else if (getenv ("POSIXLY_CORRECT") != NULL) - ordering = REQUIRE_ORDER; - else - ordering = PERMUTE; - } - - if (nextchar == NULL || *nextchar == '\0') - { - if (ordering == PERMUTE) - { - /* If we have just processed some options following some non-options, - exchange them so that the options come first. */ - - if (first_nonopt != last_nonopt && last_nonopt != optind) - exchange ((char **) argv); - else if (last_nonopt != optind) - first_nonopt = optind; - - /* Now skip any additional non-options - and extend the range of non-options previously skipped. */ - - while (optind < argc - && (argv[optind][0] != '-' || argv[optind][1] == '\0') -#ifdef GETOPT_COMPAT - && (longopts == NULL - || argv[optind][0] != '+' || argv[optind][1] == '\0') -#endif /* GETOPT_COMPAT */ - ) - optind++; - last_nonopt = optind; - } - - /* Special ARGV-element `--' means premature end of options. - Skip it like a null option, - then exchange with previous non-options as if it were an option, - then skip everything else like a non-option. */ - - if (optind != argc && !strcmp (argv[optind], "--")) - { - optind++; - - if (first_nonopt != last_nonopt && last_nonopt != optind) - exchange ((char **) argv); - else if (first_nonopt == last_nonopt) - first_nonopt = optind; - last_nonopt = argc; - - optind = argc; - } - - /* If we have done all the ARGV-elements, stop the scan - and back over any non-options that we skipped and permuted. */ - - if (optind == argc) - { - /* Set the next-arg-index to point at the non-options - that we previously skipped, so the caller will digest them. */ - if (first_nonopt != last_nonopt) - optind = first_nonopt; - return EOF; - } - - /* If we have come to a non-option and did not permute it, - either stop the scan or describe it to the caller and pass it by. */ - - if ((argv[optind][0] != '-' || argv[optind][1] == '\0') -#ifdef GETOPT_COMPAT - && (longopts == NULL - || argv[optind][0] != '+' || argv[optind][1] == '\0') -#endif /* GETOPT_COMPAT */ - ) - { - if (ordering == REQUIRE_ORDER) - return EOF; - optarg = argv[optind++]; - return 1; - } - - /* We have found another option-ARGV-element. - Start decoding its characters. */ - - nextchar = (argv[optind] + 1 - + (longopts != NULL && argv[optind][1] == '-')); - } - - if (longopts != NULL - && ((argv[optind][0] == '-' - && (argv[optind][1] == '-' || long_only)) -#ifdef GETOPT_COMPAT - || argv[optind][0] == '+' -#endif /* GETOPT_COMPAT */ - )) - { - const struct option *p; - char *s = nextchar; - int exact = 0; - int ambig = 0; - const struct option *pfound = NULL; - int indfound = 0; - - while (*s && *s != '=') - s++; - - /* Test all options for either exact match or abbreviated matches. */ - for (p = longopts, option_index = 0; p->name; - p++, option_index++) - if (!strncmp (p->name, nextchar, s - nextchar)) - { - if (s - nextchar == strlen (p->name)) - { - /* Exact match found. */ - pfound = p; - indfound = option_index; - exact = 1; - break; - } - else if (pfound == NULL) - { - /* First nonexact match found. */ - pfound = p; - indfound = option_index; - } - else - /* Second nonexact match found. */ - ambig = 1; - } - - if (ambig && !exact) - { - if (opterr) - fprintf (stderr, "%s: option `%s' is ambiguous\n", - argv[0], argv[optind]); - nextchar += strlen (nextchar); - optind++; - return '?'; - } - - if (pfound != NULL) - { - option_index = indfound; - optind++; - if (*s) - { - /* Don't test has_arg with >, because some C compilers don't - allow it to be used on enums. */ - if (pfound->has_arg) - optarg = s + 1; - else - { - if (opterr) - { - if (argv[optind - 1][1] == '-') - /* --option */ - fprintf (stderr, - "%s: option `--%s' doesn't allow an argument\n", - argv[0], pfound->name); - else - /* +option or -option */ - fprintf (stderr, - "%s: option `%c%s' doesn't allow an argument\n", - argv[0], argv[optind - 1][0], pfound->name); - } - nextchar += strlen (nextchar); - return '?'; - } - } - else if (pfound->has_arg == 1) - { - if (optind < argc) - optarg = argv[optind++]; - else - { - if (opterr) - fprintf (stderr, "%s: option `%s' requires an argument\n", - argv[0], argv[optind - 1]); - nextchar += strlen (nextchar); - return optstring[0] == ':' ? ':' : '?'; - } - } - nextchar += strlen (nextchar); - if (longind != NULL) - *longind = option_index; - if (pfound->flag) - { - *(pfound->flag) = pfound->val; - return 0; - } - return pfound->val; - } - /* Can't find it as a long option. If this is not getopt_long_only, - or the option starts with '--' or is not a valid short - option, then it's an error. - Otherwise interpret it as a short option. */ - if (!long_only || argv[optind][1] == '-' -#ifdef GETOPT_COMPAT - || argv[optind][0] == '+' -#endif /* GETOPT_COMPAT */ - || my_index (optstring, *nextchar) == NULL) - { - if (opterr) - { - if (argv[optind][1] == '-') - /* --option */ - fprintf (stderr, "%s: unrecognized option `--%s'\n", - argv[0], nextchar); - else - /* +option or -option */ - fprintf (stderr, "%s: unrecognized option `%c%s'\n", - argv[0], argv[optind][0], nextchar); - } - nextchar = (char *) ""; - optind++; - return '?'; - } - } - - /* Look at and handle the next option-character. */ - - { - char c = *nextchar++; - char *temp = my_index (optstring, c); - - /* Increment `optind' when we start to process its last character. */ - if (*nextchar == '\0') - ++optind; - - if (temp == NULL || c == ':') - { - if (opterr) - { -#if 0 - if (c < 040 || c >= 0177) - fprintf (stderr, "%s: unrecognized option, character code 0%o\n", - argv[0], c); - else - fprintf (stderr, "%s: unrecognized option `-%c'\n", argv[0], c); -#else - /* 1003.2 specifies the format of this message. */ - fprintf (stderr, "%s: illegal option -- %c\n", argv[0], c); -#endif - } - optopt = c; - return '?'; - } - if (temp[1] == ':') - { - if (temp[2] == ':') - { - /* This is an option that accepts an argument optionally. */ - if (*nextchar != '\0') - { - optarg = nextchar; - optind++; - } - else - optarg = 0; - nextchar = NULL; - } - else - { - /* This is an option that requires an argument. */ - if (*nextchar != '\0') - { - optarg = nextchar; - /* If we end this ARGV-element by taking the rest as an arg, - we must advance to the next element now. */ - optind++; - } - else if (optind == argc) - { - if (opterr) - { -#if 0 - fprintf (stderr, "%s: option `-%c' requires an argument\n", - argv[0], c); -#else - /* 1003.2 specifies the format of this message. */ - fprintf (stderr, "%s: option requires an argument -- %c\n", - argv[0], c); -#endif - } - optopt = c; - if (optstring[0] == ':') - c = ':'; - else - c = '?'; - } - else - /* We already incremented `optind' once; - increment it again when taking next ARGV-elt as argument. */ - optarg = argv[optind++]; - nextchar = NULL; - } - } - return c; - } -} - -int -getopt (argc, argv, optstring) - int argc; - char *const *argv; - const char *optstring; -{ - return _getopt_internal (argc, argv, optstring, - (const struct option *) 0, - (int *) 0, - 0); -} - -#endif /* _LIBC or not __GNU_LIBRARY__. */ - -#ifdef TEST - -/* Compile with -DTEST to make an executable for use in testing - the above definition of `getopt'. */ - -int -main (argc, argv) - int argc; - char **argv; -{ - int c; - int digit_optind = 0; - - while (1) - { - int this_option_optind = optind ? optind : 1; - - c = getopt (argc, argv, "abc:d:0123456789"); - if (c == EOF) - break; - - switch (c) - { - case '0': - case '1': - case '2': - case '3': - case '4': - case '5': - case '6': - case '7': - case '8': - case '9': - if (digit_optind != 0 && digit_optind != this_option_optind) - printf ("digits occur in two different argv-elements.\n"); - digit_optind = this_option_optind; - printf ("option %c\n", c); - break; - - case 'a': - printf ("option a\n"); - break; - - case 'b': - printf ("option b\n"); - break; - - case 'c': - printf ("option c with value `%s'\n", optarg); - break; - - case '?': - break; - - default: - printf ("?? getopt returned character code 0%o ??\n", c); - } - } - - if (optind < argc) - { - printf ("non-option ARGV-elements: "); - while (optind < argc) - printf ("%s ", argv[optind++]); - printf ("\n"); - } - - exit (0); -} - -#endif /* TEST */ diff --git a/lib/contrib/getopt.h b/lib/contrib/getopt.h deleted file mode 100644 index 45541f5a..00000000 --- a/lib/contrib/getopt.h +++ /dev/null @@ -1,129 +0,0 @@ -/* Declarations for getopt. - Copyright (C) 1989, 1990, 1991, 1992, 1993 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 2, 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, write to the Free Software - Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ - -#ifndef _GETOPT_H -#define _GETOPT_H 1 - -#ifdef __cplusplus -extern "C" { -#endif - -/* For communication from `getopt' to the caller. - When `getopt' finds an option that takes an argument, - the argument value is returned here. - Also, when `ordering' is RETURN_IN_ORDER, - each non-option ARGV-element is returned here. */ - -extern char *optarg; - -/* Index in ARGV of the next element to be scanned. - This is used for communication to and from the caller - and for communication between successive calls to `getopt'. - - On entry to `getopt', zero means this is the first call; initialize. - - When `getopt' returns EOF, this is the index of the first of the - non-option elements that the caller should itself scan. - - Otherwise, `optind' communicates from one call to the next - how much of ARGV has been scanned so far. */ - -extern int optind; - -/* Callers store zero here to inhibit the error message `getopt' prints - for unrecognized options. */ - -extern int opterr; - -/* Set to an option character which was unrecognized. */ - -extern int optopt; - -/* Describe the long-named options requested by the application. - The LONG_OPTIONS argument to getopt_long or getopt_long_only is a vector - of `struct option' terminated by an element containing a name which is - zero. - - The field `has_arg' is: - no_argument (or 0) if the option does not take an argument, - required_argument (or 1) if the option requires an argument, - optional_argument (or 2) if the option takes an optional argument. - - If the field `flag' is not NULL, it points to a variable that is set - to the value given in the field `val' when the option is found, but - left unchanged if the option is not found. - - To have a long-named option do something other than set an `int' to - a compiled-in constant, such as set a value from `optarg', set the - option's `flag' field to zero and its `val' field to a nonzero - value (the equivalent single-letter option character, if there is - one). For long options that have a zero `flag' field, `getopt' - returns the contents of the `val' field. */ - -struct option -{ -#if __STDC__ - const char *name; -#else - char *name; -#endif - /* has_arg can't be an enum because some compilers complain about - type mismatches in all the code that assumes it is an int. */ - int has_arg; - int *flag; - int val; -}; - -/* Names for the values of the `has_arg' field of `struct option'. */ - -#define no_argument 0 -#define required_argument 1 -#define optional_argument 2 - -#if __STDC__ -#if defined(__GNU_LIBRARY__) -/* Many other libraries have conflicting prototypes for getopt, with - differences in the consts, in stdlib.h. To avoid compilation - errors, only prototype getopt for the GNU C library. */ -extern int getopt (int argc, char *const *argv, const char *shortopts); -#else /* not __GNU_LIBRARY__ */ -extern int getopt (); -#endif /* not __GNU_LIBRARY__ */ -extern int getopt_long (int argc, char *const *argv, const char *shortopts, - const struct option *longopts, int *longind); -extern int getopt_long_only (int argc, char *const *argv, - const char *shortopts, - const struct option *longopts, int *longind); - -/* Internal only. Users should not call this directly. */ -extern int _getopt_internal (int argc, char *const *argv, - const char *shortopts, - const struct option *longopts, int *longind, - int long_only); -#else /* not __STDC__ */ -extern int getopt (); -extern int getopt_long (); -extern int getopt_long_only (); - -extern int _getopt_internal (); -#endif /* not __STDC__ */ - -#ifdef __cplusplus -} -#endif - -#endif /* _GETOPT_H */ diff --git a/lib/contrib/getopt1.c b/lib/contrib/getopt1.c deleted file mode 100644 index fe6bdd99..00000000 --- a/lib/contrib/getopt1.c +++ /dev/null @@ -1,177 +0,0 @@ -/* getopt_long and getopt_long_only entry points for GNU getopt. - Copyright (C) 1987, 88, 89, 90, 91, 92, 1993 - 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 2, 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, write to the Free Software - Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ - - -#include "getopt.h" - -#ifndef __STDC__ -/* This is a separate conditional since some stdc systems - reject `defined (const)'. */ -#ifndef const -#define const -#endif -#endif - -#include - -/* Comment out all this code if we are using the GNU C Library, and are not - actually compiling the library itself. This code is part of the GNU C - Library, but also included in many other GNU distributions. Compiling - and linking in this code is a waste when using the GNU C library - (especially if it is a shared library). Rather than having every GNU - program understand `configure --with-gnu-libc' and omit the object files, - it is simpler to just do this in the source for each such file. */ - -#if defined (_LIBC) || !defined (__GNU_LIBRARY__) - - -/* This needs to come after some library #include - to get __GNU_LIBRARY__ defined. */ -#ifdef __GNU_LIBRARY__ -#include -#else -char *getenv (); -#endif - -#ifndef NULL -#define NULL 0 -#endif - -int -getopt_long (argc, argv, options, long_options, opt_index) - int argc; - char *const *argv; - const char *options; - const struct option *long_options; - int *opt_index; -{ - return _getopt_internal (argc, argv, options, long_options, opt_index, 0); -} - -/* Like getopt_long, but '-' as well as '--' can indicate a long option. - If an option that starts with '-' (not '--') doesn't match a long option, - but does match a short option, it is parsed as a short option - instead. */ - -int -getopt_long_only (argc, argv, options, long_options, opt_index) - int argc; - char *const *argv; - const char *options; - const struct option *long_options; - int *opt_index; -{ - return _getopt_internal (argc, argv, options, long_options, opt_index, 1); -} - - -#endif /* _LIBC or not __GNU_LIBRARY__. */ - -#ifdef TEST - -#include - -int -main (argc, argv) - int argc; - char **argv; -{ - int c; - int digit_optind = 0; - - while (1) - { - int this_option_optind = optind ? optind : 1; - int option_index = 0; - static struct option long_options[] = - { - {"add", 1, 0, 0}, - {"append", 0, 0, 0}, - {"delete", 1, 0, 0}, - {"verbose", 0, 0, 0}, - {"create", 0, 0, 0}, - {"file", 1, 0, 0}, - {0, 0, 0, 0} - }; - - c = getopt_long (argc, argv, "abc:d:0123456789", - long_options, &option_index); - if (c == EOF) - break; - - switch (c) - { - case 0: - printf ("option %s", long_options[option_index].name); - if (optarg) - printf (" with arg %s", optarg); - printf ("\n"); - break; - - case '0': - case '1': - case '2': - case '3': - case '4': - case '5': - case '6': - case '7': - case '8': - case '9': - if (digit_optind != 0 && digit_optind != this_option_optind) - printf ("digits occur in two different argv-elements.\n"); - digit_optind = this_option_optind; - printf ("option %c\n", c); - break; - - case 'a': - printf ("option a\n"); - break; - - case 'b': - printf ("option b\n"); - break; - - case 'c': - printf ("option c with value `%s'\n", optarg); - break; - - case 'd': - printf ("option d with value `%s'\n", optarg); - break; - - case '?': - break; - - default: - printf ("?? getopt returned character code 0%o ??\n", c); - } - } - - if (optind < argc) - { - printf ("non-option ARGV-elements: "); - while (optind < argc) - printf ("%s ", argv[optind++]); - printf ("\n"); - } - - exit (0); -} - -#endif /* TEST */ diff --git a/lib/contrib/regex.c b/lib/contrib/regex.c deleted file mode 100644 index 6aec4348..00000000 --- a/lib/contrib/regex.c +++ /dev/null @@ -1,4955 +0,0 @@ -/* Extended regular expression matching and search library, - version 0.12. - (Implements POSIX draft P10003.2/D11.2, except for - internationalization features.) - - Copyright (C) 1993 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 2, 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, write to the Free Software - Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. */ - -/* AIX requires this to be the first thing in the file. */ -#if defined (_AIX) && !defined (REGEX_MALLOC) - #pragma alloca -#endif - -/* -#ifndef _GNU_SOURCE -#define _GNU_SOURCE -#endif -*/ - -/* We need this for `regex.h', and perhaps for the Emacs include files. */ -#include -#include -#include -#include - -/* -ifdef HAVE_CONFIG_H -##include "config.h" -#endif -*/ - -/* The `emacs' switch turns on certain matching commands - that make sense only in Emacs. */ -#ifdef emacs - -#include "lisp.h" -#include "buffer.h" -#include "syntax.h" - -/* Emacs uses `NULL' as a predicate. */ -#undef NULL - -#else /* not emacs */ - -/* We used to test for `BSTRING' here, but only GCC and Emacs define - `BSTRING', as far as I know, and neither of them use this code. */ -#if HAVE_STRING_H || STDC_HEADERS -#include -#ifndef bcmp -#define bcmp(s1, s2, n) memcmp ((s1), (s2), (n)) -#endif -#ifndef bcopy -#define bcopy(s, d, n) memcpy ((d), (s), (n)) -#endif -#ifndef bzero -#define bzero(s, n) memset ((s), 0, (n)) -#endif -#else -#include -#endif - -#include - -/* Define the syntax stuff for \<, \>, etc. */ - -/* This must be nonzero for the wordchar and notwordchar pattern - commands in re_match_2. */ -#ifndef Sword -#define Sword 1 -#endif - -#ifdef SYNTAX_TABLE - -extern char *re_syntax_table; - -#else /* not SYNTAX_TABLE */ - -/* How many characters in the character set. */ -#define CHAR_SET_SIZE 256 - -static char re_syntax_table[CHAR_SET_SIZE]; - -static void -init_syntax_once () -{ - register int c; - static int done = 0; - - if (done) - return; - - bzero (re_syntax_table, sizeof re_syntax_table); - - for (c = 'a'; c <= 'z'; c++) - re_syntax_table[c] = Sword; - - for (c = 'A'; c <= 'Z'; c++) - re_syntax_table[c] = Sword; - - for (c = '0'; c <= '9'; c++) - re_syntax_table[c] = Sword; - - re_syntax_table['_'] = Sword; - - done = 1; -} - -#endif /* not SYNTAX_TABLE */ - -#define SYNTAX(c) re_syntax_table[c] - -#endif /* not emacs */ - -/* Get the interface, including the syntax bits. */ -#include "regex.h" - -/* isalpha etc. are used for the character classes. */ -#include - -#ifndef isascii -#define isascii(c) 1 -#endif - -#ifdef isblank -#define ISBLANK(c) (isascii (c) && isblank (c)) -#else -#define ISBLANK(c) ((c) == ' ' || (c) == '\t') -#endif -#ifdef isgraph -#define ISGRAPH(c) (isascii (c) && isgraph (c)) -#else -#define ISGRAPH(c) (isascii (c) && isprint (c) && !isspace (c)) -#endif - -#define ISPRINT(c) (isascii (c) && isprint (c)) -#define ISDIGIT(c) (isascii (c) && isdigit (c)) -#define ISALNUM(c) (isascii (c) && isalnum (c)) -#define ISALPHA(c) (isascii (c) && isalpha (c)) -#define ISCNTRL(c) (isascii (c) && iscntrl (c)) -#define ISLOWER(c) (isascii (c) && islower (c)) -#define ISPUNCT(c) (isascii (c) && ispunct (c)) -#define ISSPACE(c) (isascii (c) && isspace (c)) -#define ISUPPER(c) (isascii (c) && isupper (c)) -#define ISXDIGIT(c) (isascii (c) && isxdigit (c)) - -#ifndef NULL -#define NULL 0 -#endif - -/* We remove any previous definition of `SIGN_EXTEND_CHAR', - since ours (we hope) works properly with all combinations of - machines, compilers, `char' and `unsigned char' argument types. - (Per Bothner suggested the basic approach.) */ -#undef SIGN_EXTEND_CHAR -#if __STDC__ -#define SIGN_EXTEND_CHAR(c) ((signed char) (c)) -#else /* not __STDC__ */ -/* As in Harbison and Steele. */ -#define SIGN_EXTEND_CHAR(c) ((((unsigned char) (c)) ^ 128) - 128) -#endif - -/* Should we use malloc or alloca? If REGEX_MALLOC is not defined, we - use `alloca' instead of `malloc'. This is because using malloc in - re_search* or re_match* could cause memory leaks when C-g is used in - Emacs; also, malloc is slower and causes storage fragmentation. On - the other hand, malloc is more portable, and easier to debug. - - Because we sometimes use alloca, some routines have to be macros, - not functions -- `alloca'-allocated space disappears at the end of the - function it is called in. */ - -#ifdef REGEX_MALLOC - -#define REGEX_ALLOCATE malloc -#define REGEX_REALLOCATE(source, osize, nsize) realloc (source, nsize) - -#else /* not REGEX_MALLOC */ - -/* Emacs already defines alloca, sometimes. */ -#ifndef alloca - -/* Make alloca work the best possible way. */ -#ifdef __GNUC__ -#define alloca __builtin_alloca -#else /* not __GNUC__ */ -#if HAVE_ALLOCA_H -#include -#else /* not __GNUC__ or HAVE_ALLOCA_H */ -#ifndef _AIX /* Already did AIX, up at the top. */ -char *alloca (); -#endif /* not _AIX */ -#endif /* not HAVE_ALLOCA_H */ -#endif /* not __GNUC__ */ - -#endif /* not alloca */ - -#define REGEX_ALLOCATE alloca - -/* Assumes a `char *destination' variable. */ -#define REGEX_REALLOCATE(source, osize, nsize) \ - (destination = (char *) alloca (nsize), \ - bcopy (source, destination, osize), \ - destination) - -#endif /* not REGEX_MALLOC */ - - -/* True if `size1' is non-NULL and PTR is pointing anywhere inside - `string1' or just past its end. This works if PTR is NULL, which is - a good thing. */ -#define FIRST_STRING_P(ptr) \ - (size1 && string1 <= (ptr) && (ptr) <= string1 + size1) - -/* (Re)Allocate N items of type T using malloc, or fail. */ -#define TALLOC(n, t) ((t *) malloc ((n) * sizeof (t))) -#define RETALLOC(addr, n, t) ((addr) = (t *) realloc (addr, (n) * sizeof (t))) -#define REGEX_TALLOC(n, t) ((t *) REGEX_ALLOCATE ((n) * sizeof (t))) - -#define BYTEWIDTH 8 /* In bits. */ - -#define STREQ(s1, s2) ((strcmp (s1, s2) == 0)) - -#define MAX(a, b) ((a) > (b) ? (a) : (b)) -#define MIN(a, b) ((a) < (b) ? (a) : (b)) - -typedef char boolean; -#define false 0 -#define true 1 - -/* These are the command codes that appear in compiled regular - expressions. Some opcodes are followed by argument bytes. A - command code can specify any interpretation whatsoever for its - arguments. Zero bytes may appear in the compiled regular expression. - - The value of `exactn' is needed in search.c (search_buffer) in Emacs. - So regex.h defines a symbol `RE_EXACTN_VALUE' to be 1; the value of - `exactn' we use here must also be 1. */ - -typedef enum -{ - no_op = 0, - - /* Followed by one byte giving n, then by n literal bytes. */ - exactn = 1, - - /* Matches any (more or less) character. */ - anychar, - - /* Matches any one char belonging to specified set. First - following byte is number of bitmap bytes. Then come bytes - for a bitmap saying which chars are in. Bits in each byte - are ordered low-bit-first. A character is in the set if its - bit is 1. A character too large to have a bit in the map is - automatically not in the set. */ - charset, - - /* Same parameters as charset, but match any character that is - not one of those specified. */ - charset_not, - - /* Start remembering the text that is matched, for storing in a - register. Followed by one byte with the register number, in - the range 0 to one less than the pattern buffer's re_nsub - field. Then followed by one byte with the number of groups - inner to this one. (This last has to be part of the - start_memory only because we need it in the on_failure_jump - of re_match_2.) */ - start_memory, - - /* Stop remembering the text that is matched and store it in a - memory register. Followed by one byte with the register - number, in the range 0 to one less than `re_nsub' in the - pattern buffer, and one byte with the number of inner groups, - just like `start_memory'. (We need the number of inner - groups here because we don't have any easy way of finding the - corresponding start_memory when we're at a stop_memory.) */ - stop_memory, - - /* Match a duplicate of something remembered. Followed by one - byte containing the register number. */ - duplicate, - - /* Fail unless at beginning of line. */ - begline, - - /* Fail unless at end of line. */ - endline, - - /* Succeeds if at beginning of buffer (if emacs) or at beginning - of string to be matched (if not). */ - begbuf, - - /* Analogously, for end of buffer/string. */ - endbuf, - - /* Followed by two byte relative address to which to jump. */ - jump, - - /* Same as jump, but marks the end of an alternative. */ - jump_past_alt, - - /* Followed by two-byte relative address of place to resume at - in case of failure. */ - on_failure_jump, - - /* Like on_failure_jump, but pushes a placeholder instead of the - current string position when executed. */ - on_failure_keep_string_jump, - - /* Throw away latest failure point and then jump to following - two-byte relative address. */ - pop_failure_jump, - - /* Change to pop_failure_jump if know won't have to backtrack to - match; otherwise change to jump. This is used to jump - back to the beginning of a repeat. If what follows this jump - clearly won't match what the repeat does, such that we can be - sure that there is no use backtracking out of repetitions - already matched, then we change it to a pop_failure_jump. - Followed by two-byte address. */ - maybe_pop_jump, - - /* Jump to following two-byte address, and push a dummy failure - point. This failure point will be thrown away if an attempt - is made to use it for a failure. A `+' construct makes this - before the first repeat. Also used as an intermediary kind - of jump when compiling an alternative. */ - dummy_failure_jump, - - /* Push a dummy failure point and continue. Used at the end of - alternatives. */ - push_dummy_failure, - - /* Followed by two-byte relative address and two-byte number n. - After matching N times, jump to the address upon failure. */ - succeed_n, - - /* Followed by two-byte relative address, and two-byte number n. - Jump to the address N times, then fail. */ - jump_n, - - /* Set the following two-byte relative address to the - subsequent two-byte number. The address *includes* the two - bytes of number. */ - set_number_at, - - wordchar, /* Matches any word-constituent character. */ - notwordchar, /* Matches any char that is not a word-constituent. */ - - wordbeg, /* Succeeds if at word beginning. */ - wordend, /* Succeeds if at word end. */ - - wordbound, /* Succeeds if at a word boundary. */ - notwordbound /* Succeeds if not at a word boundary. */ - -#ifdef emacs - ,before_dot, /* Succeeds if before point. */ - at_dot, /* Succeeds if at point. */ - after_dot, /* Succeeds if after point. */ - - /* Matches any character whose syntax is specified. Followed by - a byte which contains a syntax code, e.g., Sword. */ - syntaxspec, - - /* Matches any character whose syntax is not that specified. */ - notsyntaxspec -#endif /* emacs */ -} re_opcode_t; - -/* Common operations on the compiled pattern. */ - -/* Store NUMBER in two contiguous bytes starting at DESTINATION. */ - -#define STORE_NUMBER(destination, number) \ - do { \ - (destination)[0] = (number) & 0377; \ - (destination)[1] = (number) >> 8; \ - } while (0) - -/* Same as STORE_NUMBER, except increment DESTINATION to - the byte after where the number is stored. Therefore, DESTINATION - must be an lvalue. */ - -#define STORE_NUMBER_AND_INCR(destination, number) \ - do { \ - STORE_NUMBER (destination, number); \ - (destination) += 2; \ - } while (0) - -/* Put into DESTINATION a number stored in two contiguous bytes starting - at SOURCE. */ - -#define EXTRACT_NUMBER(destination, source) \ - do { \ - (destination) = *(source) & 0377; \ - (destination) += SIGN_EXTEND_CHAR (*((source) + 1)) << 8; \ - } while (0) - -#ifdef DEBUG -static void -extract_number (dest, source) - int *dest; - unsigned char *source; -{ - int temp = SIGN_EXTEND_CHAR (*(source + 1)); - *dest = *source & 0377; - *dest += temp << 8; -} - -#ifndef EXTRACT_MACROS /* To debug the macros. */ -#undef EXTRACT_NUMBER -#define EXTRACT_NUMBER(dest, src) extract_number (&dest, src) -#endif /* not EXTRACT_MACROS */ - -#endif /* DEBUG */ - -/* Same as EXTRACT_NUMBER, except increment SOURCE to after the number. - SOURCE must be an lvalue. */ - -#define EXTRACT_NUMBER_AND_INCR(destination, source) \ - do { \ - EXTRACT_NUMBER (destination, source); \ - (source) += 2; \ - } while (0) - -#ifdef DEBUG -static void -extract_number_and_incr (destination, source) - int *destination; - unsigned char **source; -{ - extract_number (destination, *source); - *source += 2; -} - -#ifndef EXTRACT_MACROS -#undef EXTRACT_NUMBER_AND_INCR -#define EXTRACT_NUMBER_AND_INCR(dest, src) \ - extract_number_and_incr (&dest, &src) -#endif /* not EXTRACT_MACROS */ - -#endif /* DEBUG */ - -/* If DEBUG is defined, Regex prints many voluminous messages about what - it is doing (if the variable `debug' is nonzero). If linked with the - main program in `iregex.c', you can enter patterns and strings - interactively. And if linked with the main program in `main.c' and - the other test files, you can run the already-written tests. */ - -#ifdef DEBUG - -/* We use standard I/O for debugging. */ -#include - -/* It is useful to test things that ``must'' be true when debugging. */ -#include - -static int debug = 0; - -#define DEBUG_STATEMENT(e) e -#define DEBUG_PRINT1(x) if (debug) printf (x) -#define DEBUG_PRINT2(x1, x2) if (debug) printf (x1, x2) -#define DEBUG_PRINT3(x1, x2, x3) if (debug) printf (x1, x2, x3) -#define DEBUG_PRINT4(x1, x2, x3, x4) if (debug) printf (x1, x2, x3, x4) -#define DEBUG_PRINT_COMPILED_PATTERN(p, s, e) \ - if (debug) print_partial_compiled_pattern (s, e) -#define DEBUG_PRINT_DOUBLE_STRING(w, s1, sz1, s2, sz2) \ - if (debug) print_double_string (w, s1, sz1, s2, sz2) - - -extern void printchar (); - -/* Print the fastmap in human-readable form. */ - -void -print_fastmap (fastmap) - char *fastmap; -{ - unsigned was_a_range = 0; - unsigned i = 0; - - while (i < (1 << BYTEWIDTH)) - { - if (fastmap[i++]) - { - was_a_range = 0; - printchar (i - 1); - while (i < (1 << BYTEWIDTH) && fastmap[i]) - { - was_a_range = 1; - i++; - } - if (was_a_range) - { - printf ("-"); - printchar (i - 1); - } - } - } - putchar ('\n'); -} - - -/* Print a compiled pattern string in human-readable form, starting at - the START pointer into it and ending just before the pointer END. */ - -void -print_partial_compiled_pattern (start, end) - unsigned char *start; - unsigned char *end; -{ - int mcnt, mcnt2; - unsigned char *p = start; - unsigned char *pend = end; - - if (start == NULL) - { - printf ("(null)\n"); - return; - } - - /* Loop over pattern commands. */ - while (p < pend) - { - switch ((re_opcode_t) *p++) - { - case no_op: - printf ("/no_op"); - break; - - case exactn: - mcnt = *p++; - printf ("/exactn/%d", mcnt); - do - { - putchar ('/'); - printchar (*p++); - } - while (--mcnt); - break; - - case start_memory: - mcnt = *p++; - printf ("/start_memory/%d/%d", mcnt, *p++); - break; - - case stop_memory: - mcnt = *p++; - printf ("/stop_memory/%d/%d", mcnt, *p++); - break; - - case duplicate: - printf ("/duplicate/%d", *p++); - break; - - case anychar: - printf ("/anychar"); - break; - - case charset: - case charset_not: - { - register int c; - - printf ("/charset%s", - (re_opcode_t) *(p - 1) == charset_not ? "_not" : ""); - - assert (p + *p < pend); - - for (c = 0; c < *p; c++) - { - unsigned bit; - unsigned char map_byte = p[1 + c]; - - putchar ('/'); - - for (bit = 0; bit < BYTEWIDTH; bit++) - if (map_byte & (1 << bit)) - printchar (c * BYTEWIDTH + bit); - } - p += 1 + *p; - break; - } - - case begline: - printf ("/begline"); - break; - - case endline: - printf ("/endline"); - break; - - case on_failure_jump: - extract_number_and_incr (&mcnt, &p); - printf ("/on_failure_jump/0/%d", mcnt); - break; - - case on_failure_keep_string_jump: - extract_number_and_incr (&mcnt, &p); - printf ("/on_failure_keep_string_jump/0/%d", mcnt); - break; - - case dummy_failure_jump: - extract_number_and_incr (&mcnt, &p); - printf ("/dummy_failure_jump/0/%d", mcnt); - break; - - case push_dummy_failure: - printf ("/push_dummy_failure"); - break; - - case maybe_pop_jump: - extract_number_and_incr (&mcnt, &p); - printf ("/maybe_pop_jump/0/%d", mcnt); - break; - - case pop_failure_jump: - extract_number_and_incr (&mcnt, &p); - printf ("/pop_failure_jump/0/%d", mcnt); - break; - - case jump_past_alt: - extract_number_and_incr (&mcnt, &p); - printf ("/jump_past_alt/0/%d", mcnt); - break; - - case jump: - extract_number_and_incr (&mcnt, &p); - printf ("/jump/0/%d", mcnt); - break; - - case succeed_n: - extract_number_and_incr (&mcnt, &p); - extract_number_and_incr (&mcnt2, &p); - printf ("/succeed_n/0/%d/0/%d", mcnt, mcnt2); - break; - - case jump_n: - extract_number_and_incr (&mcnt, &p); - extract_number_and_incr (&mcnt2, &p); - printf ("/jump_n/0/%d/0/%d", mcnt, mcnt2); - break; - - case set_number_at: - extract_number_and_incr (&mcnt, &p); - extract_number_and_incr (&mcnt2, &p); - printf ("/set_number_at/0/%d/0/%d", mcnt, mcnt2); - break; - - case wordbound: - printf ("/wordbound"); - break; - - case notwordbound: - printf ("/notwordbound"); - break; - - case wordbeg: - printf ("/wordbeg"); - break; - - case wordend: - printf ("/wordend"); - -#ifdef emacs - case before_dot: - printf ("/before_dot"); - break; - - case at_dot: - printf ("/at_dot"); - break; - - case after_dot: - printf ("/after_dot"); - break; - - case syntaxspec: - printf ("/syntaxspec"); - mcnt = *p++; - printf ("/%d", mcnt); - break; - - case notsyntaxspec: - printf ("/notsyntaxspec"); - mcnt = *p++; - printf ("/%d", mcnt); - break; -#endif /* emacs */ - - case wordchar: - printf ("/wordchar"); - break; - - case notwordchar: - printf ("/notwordchar"); - break; - - case begbuf: - printf ("/begbuf"); - break; - - case endbuf: - printf ("/endbuf"); - break; - - default: - printf ("?%d", *(p-1)); - } - } - printf ("/\n"); -} - - -void -print_compiled_pattern (bufp) - struct re_pattern_buffer *bufp; -{ - unsigned char *buffer = bufp->buffer; - - print_partial_compiled_pattern (buffer, buffer + bufp->used); - printf ("%d bytes used/%d bytes allocated.\n", bufp->used, bufp->allocated); - - if (bufp->fastmap_accurate && bufp->fastmap) - { - printf ("fastmap: "); - print_fastmap (bufp->fastmap); - } - - printf ("re_nsub: %d\t", bufp->re_nsub); - printf ("regs_alloc: %d\t", bufp->regs_allocated); - printf ("can_be_null: %d\t", bufp->can_be_null); - printf ("newline_anchor: %d\n", bufp->newline_anchor); - printf ("no_sub: %d\t", bufp->no_sub); - printf ("not_bol: %d\t", bufp->not_bol); - printf ("not_eol: %d\t", bufp->not_eol); - printf ("syntax: %d\n", bufp->syntax); - /* Perhaps we should print the translate table? */ -} - - -void -print_double_string (where, string1, size1, string2, size2) - const char *where; - const char *string1; - const char *string2; - int size1; - int size2; -{ - unsigned this_char; - - if (where == NULL) - printf ("(null)"); - else - { - if (FIRST_STRING_P (where)) - { - for (this_char = where - string1; this_char < size1; this_char++) - printchar (string1[this_char]); - - where = string2; - } - - for (this_char = where - string2; this_char < size2; this_char++) - printchar (string2[this_char]); - } -} - -#else /* not DEBUG */ - -#undef assert -#define assert(e) - -#define DEBUG_STATEMENT(e) -#define DEBUG_PRINT1(x) -#define DEBUG_PRINT2(x1, x2) -#define DEBUG_PRINT3(x1, x2, x3) -#define DEBUG_PRINT4(x1, x2, x3, x4) -#define DEBUG_PRINT_COMPILED_PATTERN(p, s, e) -#define DEBUG_PRINT_DOUBLE_STRING(w, s1, sz1, s2, sz2) - -#endif /* not DEBUG */ - -/* Set by `re_set_syntax' to the current regexp syntax to recognize. Can - also be assigned to arbitrarily: each pattern buffer stores its own - syntax, so it can be changed between regex compilations. */ -reg_syntax_t re_syntax_options = RE_SYNTAX_EMACS; - - -/* Specify the precise syntax of regexps for compilation. This provides - for compatibility for various utilities which historically have - different, incompatible syntaxes. - - The argument SYNTAX is a bit mask comprised of the various bits - defined in regex.h. We return the old syntax. */ - -reg_syntax_t -re_set_syntax (syntax) - reg_syntax_t syntax; -{ - reg_syntax_t ret = re_syntax_options; - - re_syntax_options = syntax; - return ret; -} - -/* This table gives an error message for each of the error codes listed - in regex.h. Obviously the order here has to be same as there. */ - -static const char *re_error_msg[] = - { NULL, /* REG_NOERROR */ - "No match", /* REG_NOMATCH */ - "Invalid regular expression", /* REG_BADPAT */ - "Invalid collation character", /* REG_ECOLLATE */ - "Invalid character class name", /* REG_ECTYPE */ - "Trailing backslash", /* REG_EESCAPE */ - "Invalid back reference", /* REG_ESUBREG */ - "Unmatched [ or [^", /* REG_EBRACK */ - "Unmatched ( or \\(", /* REG_EPAREN */ - "Unmatched \\{", /* REG_EBRACE */ - "Invalid content of \\{\\}", /* REG_BADBR */ - "Invalid range end", /* REG_ERANGE */ - "Memory exhausted", /* REG_ESPACE */ - "Invalid preceding regular expression", /* REG_BADRPT */ - "Premature end of regular expression", /* REG_EEND */ - "Regular expression too big", /* REG_ESIZE */ - "Unmatched ) or \\)", /* REG_ERPAREN */ - }; - -/* Subroutine declarations and macros for regex_compile. */ - -static void store_op1 (), store_op2 (); -static void insert_op1 (), insert_op2 (); -static boolean at_begline_loc_p (), at_endline_loc_p (); -static boolean group_in_compile_stack (); -static reg_errcode_t compile_range (); - -/* Fetch the next character in the uncompiled pattern---translating it - if necessary. Also cast from a signed character in the constant - string passed to us by the user to an unsigned char that we can use - as an array index (in, e.g., `translate'). */ -#define PATFETCH(c) \ - do {if (p == pend) return REG_EEND; \ - c = (unsigned char) *p++; \ - if (translate) c = translate[c]; \ - } while (0) - -/* Fetch the next character in the uncompiled pattern, with no - translation. */ -#define PATFETCH_RAW(c) \ - do {if (p == pend) return REG_EEND; \ - c = (unsigned char) *p++; \ - } while (0) - -/* Go backwards one character in the pattern. */ -#define PATUNFETCH p-- - - -/* If `translate' is non-null, return translate[D], else just D. We - cast the subscript to translate because some data is declared as - `char *', to avoid warnings when a string constant is passed. But - when we use a character as a subscript we must make it unsigned. */ -#define TRANSLATE(d) (translate ? translate[(unsigned char) (d)] : (d)) - - -/* Macros for outputting the compiled pattern into `buffer'. */ - -/* If the buffer isn't allocated when it comes in, use this. */ -#define INIT_BUF_SIZE 32 - -/* Make sure we have at least N more bytes of space in buffer. */ -#define GET_BUFFER_SPACE(n) \ - while (b - bufp->buffer + (n) > bufp->allocated) \ - EXTEND_BUFFER () - -/* Make sure we have one more byte of buffer space and then add C to it. */ -#define BUF_PUSH(c) \ - do { \ - GET_BUFFER_SPACE (1); \ - *b++ = (unsigned char) (c); \ - } while (0) - - -/* Ensure we have two more bytes of buffer space and then append C1 and C2. */ -#define BUF_PUSH_2(c1, c2) \ - do { \ - GET_BUFFER_SPACE (2); \ - *b++ = (unsigned char) (c1); \ - *b++ = (unsigned char) (c2); \ - } while (0) - - -/* As with BUF_PUSH_2, except for three bytes. */ -#define BUF_PUSH_3(c1, c2, c3) \ - do { \ - GET_BUFFER_SPACE (3); \ - *b++ = (unsigned char) (c1); \ - *b++ = (unsigned char) (c2); \ - *b++ = (unsigned char) (c3); \ - } while (0) - - -/* Store a jump with opcode OP at LOC to location TO. We store a - relative address offset by the three bytes the jump itself occupies. */ -#define STORE_JUMP(op, loc, to) \ - store_op1 (op, loc, (to) - (loc) - 3) - -/* Likewise, for a two-argument jump. */ -#define STORE_JUMP2(op, loc, to, arg) \ - store_op2 (op, loc, (to) - (loc) - 3, arg) - -/* Like `STORE_JUMP', but for inserting. Assume `b' is the buffer end. */ -#define INSERT_JUMP(op, loc, to) \ - insert_op1 (op, loc, (to) - (loc) - 3, b) - -/* Like `STORE_JUMP2', but for inserting. Assume `b' is the buffer end. */ -#define INSERT_JUMP2(op, loc, to, arg) \ - insert_op2 (op, loc, (to) - (loc) - 3, arg, b) - - -/* This is not an arbitrary limit: the arguments which represent offsets - into the pattern are two bytes long. So if 2^16 bytes turns out to - be too small, many things would have to change. */ -#define MAX_BUF_SIZE (1L << 16) - - -/* Extend the buffer by twice its current size via realloc and - reset the pointers that pointed into the old block to point to the - correct places in the new one. If extending the buffer results in it - being larger than MAX_BUF_SIZE, then flag memory exhausted. */ -#define EXTEND_BUFFER() \ - do { \ - unsigned char *old_buffer = bufp->buffer; \ - if (bufp->allocated == MAX_BUF_SIZE) \ - return REG_ESIZE; \ - bufp->allocated <<= 1; \ - if (bufp->allocated > MAX_BUF_SIZE) \ - bufp->allocated = MAX_BUF_SIZE; \ - bufp->buffer = (unsigned char *) realloc (bufp->buffer, bufp->allocated);\ - if (bufp->buffer == NULL) \ - return REG_ESPACE; \ - /* If the buffer moved, move all the pointers into it. */ \ - if (old_buffer != bufp->buffer) \ - { \ - b = (b - old_buffer) + bufp->buffer; \ - begalt = (begalt - old_buffer) + bufp->buffer; \ - if (fixup_alt_jump) \ - fixup_alt_jump = (fixup_alt_jump - old_buffer) + bufp->buffer;\ - if (laststart) \ - laststart = (laststart - old_buffer) + bufp->buffer; \ - if (pending_exact) \ - pending_exact = (pending_exact - old_buffer) + bufp->buffer; \ - } \ - } while (0) - - -/* Since we have one byte reserved for the register number argument to - {start,stop}_memory, the maximum number of groups we can report - things about is what fits in that byte. */ -#define MAX_REGNUM 255 - -/* But patterns can have more than `MAX_REGNUM' registers. We just - ignore the excess. */ -typedef unsigned regnum_t; - - -/* Macros for the compile stack. */ - -/* Since offsets can go either forwards or backwards, this type needs to - be able to hold values from -(MAX_BUF_SIZE - 1) to MAX_BUF_SIZE - 1. */ -typedef int pattern_offset_t; - -typedef struct -{ - pattern_offset_t begalt_offset; - pattern_offset_t fixup_alt_jump; - pattern_offset_t inner_group_offset; - pattern_offset_t laststart_offset; - regnum_t regnum; -} compile_stack_elt_t; - - -typedef struct -{ - compile_stack_elt_t *stack; - unsigned size; - unsigned avail; /* Offset of next open position. */ -} compile_stack_type; - - -#define INIT_COMPILE_STACK_SIZE 32 - -#define COMPILE_STACK_EMPTY (compile_stack.avail == 0) -#define COMPILE_STACK_FULL (compile_stack.avail == compile_stack.size) - -/* The next available element. */ -#define COMPILE_STACK_TOP (compile_stack.stack[compile_stack.avail]) - - -/* Set the bit for character C in a list. */ -#define SET_LIST_BIT(c) \ - (b[((unsigned char) (c)) / BYTEWIDTH] \ - |= 1 << (((unsigned char) c) % BYTEWIDTH)) - - -/* Get the next unsigned number in the uncompiled pattern. */ -#define GET_UNSIGNED_NUMBER(num) \ - { if (p != pend) \ - { \ - PATFETCH (c); \ - while (ISDIGIT (c)) \ - { \ - if (num < 0) \ - num = 0; \ - num = num * 10 + c - '0'; \ - if (p == pend) \ - break; \ - PATFETCH (c); \ - } \ - } \ - } - -#define CHAR_CLASS_MAX_LENGTH 6 /* Namely, `xdigit'. */ - -#define IS_CHAR_CLASS(string) \ - (STREQ (string, "alpha") || STREQ (string, "upper") \ - || STREQ (string, "lower") || STREQ (string, "digit") \ - || STREQ (string, "alnum") || STREQ (string, "xdigit") \ - || STREQ (string, "space") || STREQ (string, "print") \ - || STREQ (string, "punct") || STREQ (string, "graph") \ - || STREQ (string, "cntrl") || STREQ (string, "blank")) - -/* `regex_compile' compiles PATTERN (of length SIZE) according to SYNTAX. - Returns one of error codes defined in `regex.h', or zero for success. - - Assumes the `allocated' (and perhaps `buffer') and `translate' - fields are set in BUFP on entry. - - If it succeeds, results are put in BUFP (if it returns an error, the - contents of BUFP are undefined): - `buffer' is the compiled pattern; - `syntax' is set to SYNTAX; - `used' is set to the length of the compiled pattern; - `fastmap_accurate' is zero; - `re_nsub' is the number of subexpressions in PATTERN; - `not_bol' and `not_eol' are zero; - - The `fastmap' and `newline_anchor' fields are neither - examined nor set. */ - -static reg_errcode_t -regex_compile (pattern, size, syntax, bufp) - const char *pattern; - int size; - reg_syntax_t syntax; - struct re_pattern_buffer *bufp; -{ - /* We fetch characters from PATTERN here. Even though PATTERN is - `char *' (i.e., signed), we declare these variables as unsigned, so - they can be reliably used as array indices. */ - register unsigned char c, c1; - - /* A random tempory spot in PATTERN. */ - const char *p1; - - /* Points to the end of the buffer, where we should append. */ - register unsigned char *b; - - /* Keeps track of unclosed groups. */ - compile_stack_type compile_stack; - - /* Points to the current (ending) position in the pattern. */ - const char *p = pattern; - const char *pend = pattern + size; - - /* How to translate the characters in the pattern. */ - char *translate = bufp->translate; - - /* Address of the count-byte of the most recently inserted `exactn' - command. This makes it possible to tell if a new exact-match - character can be added to that command or if the character requires - a new `exactn' command. */ - unsigned char *pending_exact = 0; - - /* Address of start of the most recently finished expression. - This tells, e.g., postfix * where to find the start of its - operand. Reset at the beginning of groups and alternatives. */ - unsigned char *laststart = 0; - - /* Address of beginning of regexp, or inside of last group. */ - unsigned char *begalt; - - /* Place in the uncompiled pattern (i.e., the {) to - which to go back if the interval is invalid. */ - const char *beg_interval; - - /* Address of the place where a forward jump should go to the end of - the containing expression. Each alternative of an `or' -- except the - last -- ends with a forward jump of this sort. */ - unsigned char *fixup_alt_jump = 0; - - /* Counts open-groups as they are encountered. Remembered for the - matching close-group on the compile stack, so the same register - number is put in the stop_memory as the start_memory. */ - regnum_t regnum = 0; - -#ifdef DEBUG - DEBUG_PRINT1 ("\nCompiling pattern: "); - if (debug) - { - unsigned debug_count; - - for (debug_count = 0; debug_count < size; debug_count++) - printchar (pattern[debug_count]); - putchar ('\n'); - } -#endif /* DEBUG */ - - /* Initialize the compile stack. */ - compile_stack.stack = TALLOC (INIT_COMPILE_STACK_SIZE, compile_stack_elt_t); - if (compile_stack.stack == NULL) - return REG_ESPACE; - - compile_stack.size = INIT_COMPILE_STACK_SIZE; - compile_stack.avail = 0; - - /* Initialize the pattern buffer. */ - bufp->syntax = syntax; - bufp->fastmap_accurate = 0; - bufp->not_bol = bufp->not_eol = 0; - - /* Set `used' to zero, so that if we return an error, the pattern - printer (for debugging) will think there's no pattern. We reset it - at the end. */ - bufp->used = 0; - - /* Always count groups, whether or not bufp->no_sub is set. */ - bufp->re_nsub = 0; - -#if !defined (emacs) && !defined (SYNTAX_TABLE) - /* Initialize the syntax table. */ - init_syntax_once (); -#endif - - if (bufp->allocated == 0) - { - if (bufp->buffer) - { /* If zero allocated, but buffer is non-null, try to realloc - enough space. This loses if buffer's address is bogus, but - that is the user's responsibility. */ - RETALLOC (bufp->buffer, INIT_BUF_SIZE, unsigned char); - } - else - { /* Caller did not allocate a buffer. Do it for them. */ - bufp->buffer = TALLOC (INIT_BUF_SIZE, unsigned char); - } - if (!bufp->buffer) return REG_ESPACE; - - bufp->allocated = INIT_BUF_SIZE; - } - - begalt = b = bufp->buffer; - - /* Loop through the uncompiled pattern until we're at the end. */ - while (p != pend) - { - PATFETCH (c); - - switch (c) - { - case '^': - { - if ( /* If at start of pattern, it's an operator. */ - p == pattern + 1 - /* If context independent, it's an operator. */ - || syntax & RE_CONTEXT_INDEP_ANCHORS - /* Otherwise, depends on what's come before. */ - || at_begline_loc_p (pattern, p, syntax)) - BUF_PUSH (begline); - else - goto normal_char; - } - break; - - - case '$': - { - if ( /* If at end of pattern, it's an operator. */ - p == pend - /* If context independent, it's an operator. */ - || syntax & RE_CONTEXT_INDEP_ANCHORS - /* Otherwise, depends on what's next. */ - || at_endline_loc_p (p, pend, syntax)) - BUF_PUSH (endline); - else - goto normal_char; - } - break; - - - case '+': - case '?': - if ((syntax & RE_BK_PLUS_QM) - || (syntax & RE_LIMITED_OPS)) - goto normal_char; - handle_plus: - case '*': - /* If there is no previous pattern... */ - if (!laststart) - { - if (syntax & RE_CONTEXT_INVALID_OPS) - return REG_BADRPT; - else if (!(syntax & RE_CONTEXT_INDEP_OPS)) - goto normal_char; - } - - { - /* Are we optimizing this jump? */ - boolean keep_string_p = false; - - /* 1 means zero (many) matches is allowed. */ - char zero_times_ok = 0, many_times_ok = 0; - - /* If there is a sequence of repetition chars, collapse it - down to just one (the right one). We can't combine - interval operators with these because of, e.g., `a{2}*', - which should only match an even number of `a's. */ - - for (;;) - { - zero_times_ok |= c != '+'; - many_times_ok |= c != '?'; - - if (p == pend) - break; - - PATFETCH (c); - - if (c == '*' - || (!(syntax & RE_BK_PLUS_QM) && (c == '+' || c == '?'))) - ; - - else if (syntax & RE_BK_PLUS_QM && c == '\\') - { - if (p == pend) return REG_EESCAPE; - - PATFETCH (c1); - if (!(c1 == '+' || c1 == '?')) - { - PATUNFETCH; - PATUNFETCH; - break; - } - - c = c1; - } - else - { - PATUNFETCH; - break; - } - - /* If we get here, we found another repeat character. */ - } - - /* Star, etc. applied to an empty pattern is equivalent - to an empty pattern. */ - if (!laststart) - break; - - /* Now we know whether or not zero matches is allowed - and also whether or not two or more matches is allowed. */ - if (many_times_ok) - { /* More than one repetition is allowed, so put in at the - end a backward relative jump from `b' to before the next - jump we're going to put in below (which jumps from - laststart to after this jump). - - But if we are at the `*' in the exact sequence `.*\n', - insert an unconditional jump backwards to the ., - instead of the beginning of the loop. This way we only - push a failure point once, instead of every time - through the loop. */ - assert (p - 1 > pattern); - - /* Allocate the space for the jump. */ - GET_BUFFER_SPACE (3); - - /* We know we are not at the first character of the pattern, - because laststart was nonzero. And we've already - incremented `p', by the way, to be the character after - the `*'. Do we have to do something analogous here - for null bytes, because of RE_DOT_NOT_NULL? */ - if (TRANSLATE (*(p - 2)) == TRANSLATE ('.') - && zero_times_ok - && p < pend && TRANSLATE (*p) == TRANSLATE ('\n') - && !(syntax & RE_DOT_NEWLINE)) - { /* We have .*\n. */ - STORE_JUMP (jump, b, laststart); - keep_string_p = true; - } - else - /* Anything else. */ - STORE_JUMP (maybe_pop_jump, b, laststart - 3); - - /* We've added more stuff to the buffer. */ - b += 3; - } - - /* On failure, jump from laststart to b + 3, which will be the - end of the buffer after this jump is inserted. */ - GET_BUFFER_SPACE (3); - INSERT_JUMP (keep_string_p ? on_failure_keep_string_jump - : on_failure_jump, - laststart, b + 3); - pending_exact = 0; - b += 3; - - if (!zero_times_ok) - { - /* At least one repetition is required, so insert a - `dummy_failure_jump' before the initial - `on_failure_jump' instruction of the loop. This - effects a skip over that instruction the first time - we hit that loop. */ - GET_BUFFER_SPACE (3); - INSERT_JUMP (dummy_failure_jump, laststart, laststart + 6); - b += 3; - } - } - break; - - - case '.': - laststart = b; - BUF_PUSH (anychar); - break; - - - case '[': - { - boolean had_char_class = false; - - if (p == pend) return REG_EBRACK; - - /* Ensure that we have enough space to push a charset: the - opcode, the length count, and the bitset; 34 bytes in all. */ - GET_BUFFER_SPACE (34); - - laststart = b; - - /* We test `*p == '^' twice, instead of using an if - statement, so we only need one BUF_PUSH. */ - BUF_PUSH (*p == '^' ? charset_not : charset); - if (*p == '^') - p++; - - /* Remember the first position in the bracket expression. */ - p1 = p; - - /* Push the number of bytes in the bitmap. */ - BUF_PUSH ((1 << BYTEWIDTH) / BYTEWIDTH); - - /* Clear the whole map. */ - bzero (b, (1 << BYTEWIDTH) / BYTEWIDTH); - - /* charset_not matches newline according to a syntax bit. */ - if ((re_opcode_t) b[-2] == charset_not - && (syntax & RE_HAT_LISTS_NOT_NEWLINE)) - SET_LIST_BIT ('\n'); - - /* Read in characters and ranges, setting map bits. */ - for (;;) - { - if (p == pend) return REG_EBRACK; - - PATFETCH (c); - - /* \ might escape characters inside [...] and [^...]. */ - if ((syntax & RE_BACKSLASH_ESCAPE_IN_LISTS) && c == '\\') - { - if (p == pend) return REG_EESCAPE; - - PATFETCH (c1); - SET_LIST_BIT (c1); - continue; - } - - /* Could be the end of the bracket expression. If it's - not (i.e., when the bracket expression is `[]' so - far), the ']' character bit gets set way below. */ - if (c == ']' && p != p1 + 1) - break; - - /* Look ahead to see if it's a range when the last thing - was a character class. */ - if (had_char_class && c == '-' && *p != ']') - return REG_ERANGE; - - /* Look ahead to see if it's a range when the last thing - was a character: if this is a hyphen not at the - beginning or the end of a list, then it's the range - operator. */ - if (c == '-' - && !(p - 2 >= pattern && p[-2] == '[') - && !(p - 3 >= pattern && p[-3] == '[' && p[-2] == '^') - && *p != ']') - { - reg_errcode_t ret - = compile_range (&p, pend, translate, syntax, b); - if (ret != REG_NOERROR) return ret; - } - - else if (p[0] == '-' && p[1] != ']') - { /* This handles ranges made up of characters only. */ - reg_errcode_t ret; - - /* Move past the `-'. */ - PATFETCH (c1); - - ret = compile_range (&p, pend, translate, syntax, b); - if (ret != REG_NOERROR) return ret; - } - - /* See if we're at the beginning of a possible character - class. */ - - else if (syntax & RE_CHAR_CLASSES && c == '[' && *p == ':') - { /* Leave room for the null. */ - char str[CHAR_CLASS_MAX_LENGTH + 1]; - - PATFETCH (c); - c1 = 0; - - /* If pattern is `[[:'. */ - if (p == pend) return REG_EBRACK; - - for (;;) - { - PATFETCH (c); - if (c == ':' || c == ']' || p == pend - || c1 == CHAR_CLASS_MAX_LENGTH) - break; - str[c1++] = c; - } - str[c1] = '\0'; - - /* If isn't a word bracketed by `[:' and:`]': - undo the ending character, the letters, and leave - the leading `:' and `[' (but set bits for them). */ - if (c == ':' && *p == ']') - { - int ch; - boolean is_alnum = STREQ (str, "alnum"); - boolean is_alpha = STREQ (str, "alpha"); - boolean is_blank = STREQ (str, "blank"); - boolean is_cntrl = STREQ (str, "cntrl"); - boolean is_digit = STREQ (str, "digit"); - boolean is_graph = STREQ (str, "graph"); - boolean is_lower = STREQ (str, "lower"); - boolean is_print = STREQ (str, "print"); - boolean is_punct = STREQ (str, "punct"); - boolean is_space = STREQ (str, "space"); - boolean is_upper = STREQ (str, "upper"); - boolean is_xdigit = STREQ (str, "xdigit"); - - if (!IS_CHAR_CLASS (str)) return REG_ECTYPE; - - /* Throw away the ] at the end of the character - class. */ - PATFETCH (c); - - if (p == pend) return REG_EBRACK; - - for (ch = 0; ch < 1 << BYTEWIDTH; ch++) - { - if ( (is_alnum && ISALNUM (ch)) - || (is_alpha && ISALPHA (ch)) - || (is_blank && ISBLANK (ch)) - || (is_cntrl && ISCNTRL (ch)) - || (is_digit && ISDIGIT (ch)) - || (is_graph && ISGRAPH (ch)) - || (is_lower && ISLOWER (ch)) - || (is_print && ISPRINT (ch)) - || (is_punct && ISPUNCT (ch)) - || (is_space && ISSPACE (ch)) - || (is_upper && ISUPPER (ch)) - || (is_xdigit && ISXDIGIT (ch))) - SET_LIST_BIT (ch); - } - had_char_class = true; - } - else - { - c1++; - while (c1--) - PATUNFETCH; - SET_LIST_BIT ('['); - SET_LIST_BIT (':'); - had_char_class = false; - } - } - else - { - had_char_class = false; - SET_LIST_BIT (c); - } - } - - /* Discard any (non)matching list bytes that are all 0 at the - end of the map. Decrease the map-length byte too. */ - while ((int) b[-1] > 0 && b[b[-1] - 1] == 0) - b[-1]--; - b += b[-1]; - } - break; - - - case '(': - if (syntax & RE_NO_BK_PARENS) - goto handle_open; - else - goto normal_char; - - - case ')': - if (syntax & RE_NO_BK_PARENS) - goto handle_close; - else - goto normal_char; - - - case '\n': - if (syntax & RE_NEWLINE_ALT) - goto handle_alt; - else - goto normal_char; - - - case '|': - if (syntax & RE_NO_BK_VBAR) - goto handle_alt; - else - goto normal_char; - - - case '{': - if (syntax & RE_INTERVALS && syntax & RE_NO_BK_BRACES) - goto handle_interval; - else - goto normal_char; - - - case '\\': - if (p == pend) return REG_EESCAPE; - - /* Do not translate the character after the \, so that we can - distinguish, e.g., \B from \b, even if we normally would - translate, e.g., B to b. */ - PATFETCH_RAW (c); - - switch (c) - { - case '(': - if (syntax & RE_NO_BK_PARENS) - goto normal_backslash; - - handle_open: - bufp->re_nsub++; - regnum++; - - if (COMPILE_STACK_FULL) - { - RETALLOC (compile_stack.stack, compile_stack.size << 1, - compile_stack_elt_t); - if (compile_stack.stack == NULL) return REG_ESPACE; - - compile_stack.size <<= 1; - } - - /* These are the values to restore when we hit end of this - group. They are all relative offsets, so that if the - whole pattern moves because of realloc, they will still - be valid. */ - COMPILE_STACK_TOP.begalt_offset = begalt - bufp->buffer; - COMPILE_STACK_TOP.fixup_alt_jump - = fixup_alt_jump ? fixup_alt_jump - bufp->buffer + 1 : 0; - COMPILE_STACK_TOP.laststart_offset = b - bufp->buffer; - COMPILE_STACK_TOP.regnum = regnum; - - /* We will eventually replace the 0 with the number of - groups inner to this one. But do not push a - start_memory for groups beyond the last one we can - represent in the compiled pattern. */ - if (regnum <= MAX_REGNUM) - { - COMPILE_STACK_TOP.inner_group_offset = b - bufp->buffer + 2; - BUF_PUSH_3 (start_memory, regnum, 0); - } - - compile_stack.avail++; - - fixup_alt_jump = 0; - laststart = 0; - begalt = b; - /* If we've reached MAX_REGNUM groups, then this open - won't actually generate any code, so we'll have to - clear pending_exact explicitly. */ - pending_exact = 0; - break; - - - case ')': - if (syntax & RE_NO_BK_PARENS) goto normal_backslash; - - if (COMPILE_STACK_EMPTY) - { - if (syntax & RE_UNMATCHED_RIGHT_PAREN_ORD) - goto normal_backslash; - else - return REG_ERPAREN; - } - - handle_close: - if (fixup_alt_jump) - { /* Push a dummy failure point at the end of the - alternative for a possible future - `pop_failure_jump' to pop. See comments at - `push_dummy_failure' in `re_match_2'. */ - BUF_PUSH (push_dummy_failure); - - /* We allocated space for this jump when we assigned - to `fixup_alt_jump', in the `handle_alt' case below. */ - STORE_JUMP (jump_past_alt, fixup_alt_jump, b - 1); - } - - /* See similar code for backslashed left paren above. */ - if (COMPILE_STACK_EMPTY) - { - if (syntax & RE_UNMATCHED_RIGHT_PAREN_ORD) - goto normal_char; - else - return REG_ERPAREN; - } - - /* Since we just checked for an empty stack above, this - ``can't happen''. */ - assert (compile_stack.avail != 0); - { - /* We don't just want to restore into `regnum', because - later groups should continue to be numbered higher, - as in `(ab)c(de)' -- the second group is #2. */ - regnum_t this_group_regnum; - - compile_stack.avail--; - begalt = bufp->buffer + COMPILE_STACK_TOP.begalt_offset; - fixup_alt_jump - = COMPILE_STACK_TOP.fixup_alt_jump - ? bufp->buffer + COMPILE_STACK_TOP.fixup_alt_jump - 1 - : 0; - laststart = bufp->buffer + COMPILE_STACK_TOP.laststart_offset; - this_group_regnum = COMPILE_STACK_TOP.regnum; - /* If we've reached MAX_REGNUM groups, then this open - won't actually generate any code, so we'll have to - clear pending_exact explicitly. */ - pending_exact = 0; - - /* We're at the end of the group, so now we know how many - groups were inside this one. */ - if (this_group_regnum <= MAX_REGNUM) - { - unsigned char *inner_group_loc - = bufp->buffer + COMPILE_STACK_TOP.inner_group_offset; - - *inner_group_loc = regnum - this_group_regnum; - BUF_PUSH_3 (stop_memory, this_group_regnum, - regnum - this_group_regnum); - } - } - break; - - - case '|': /* `\|'. */ - if (syntax & RE_LIMITED_OPS || syntax & RE_NO_BK_VBAR) - goto normal_backslash; - handle_alt: - if (syntax & RE_LIMITED_OPS) - goto normal_char; - - /* Insert before the previous alternative a jump which - jumps to this alternative if the former fails. */ - GET_BUFFER_SPACE (3); - INSERT_JUMP (on_failure_jump, begalt, b + 6); - pending_exact = 0; - b += 3; - - /* The alternative before this one has a jump after it - which gets executed if it gets matched. Adjust that - jump so it will jump to this alternative's analogous - jump (put in below, which in turn will jump to the next - (if any) alternative's such jump, etc.). The last such - jump jumps to the correct final destination. A picture: - _____ _____ - | | | | - | v | v - a | b | c - - If we are at `b', then fixup_alt_jump right now points to a - three-byte space after `a'. We'll put in the jump, set - fixup_alt_jump to right after `b', and leave behind three - bytes which we'll fill in when we get to after `c'. */ - - if (fixup_alt_jump) - STORE_JUMP (jump_past_alt, fixup_alt_jump, b); - - /* Mark and leave space for a jump after this alternative, - to be filled in later either by next alternative or - when know we're at the end of a series of alternatives. */ - fixup_alt_jump = b; - GET_BUFFER_SPACE (3); - b += 3; - - laststart = 0; - begalt = b; - break; - - - case '{': - /* If \{ is a literal. */ - if (!(syntax & RE_INTERVALS) - /* If we're at `\{' and it's not the open-interval - operator. */ - || ((syntax & RE_INTERVALS) && (syntax & RE_NO_BK_BRACES)) - || (p - 2 == pattern && p == pend)) - goto normal_backslash; - - handle_interval: - { - /* If got here, then the syntax allows intervals. */ - - /* At least (most) this many matches must be made. */ - int lower_bound = -1, upper_bound = -1; - - beg_interval = p - 1; - - if (p == pend) - { - if (syntax & RE_NO_BK_BRACES) - goto unfetch_interval; - else - return REG_EBRACE; - } - - GET_UNSIGNED_NUMBER (lower_bound); - - if (c == ',') - { - GET_UNSIGNED_NUMBER (upper_bound); - if (upper_bound < 0) upper_bound = RE_DUP_MAX; - } - else - /* Interval such as `{1}' => match exactly once. */ - upper_bound = lower_bound; - - if (lower_bound < 0 || upper_bound > RE_DUP_MAX - || lower_bound > upper_bound) - { - if (syntax & RE_NO_BK_BRACES) - goto unfetch_interval; - else - return REG_BADBR; - } - - if (!(syntax & RE_NO_BK_BRACES)) - { - if (c != '\\') return REG_EBRACE; - - PATFETCH (c); - } - - if (c != '}') - { - if (syntax & RE_NO_BK_BRACES) - goto unfetch_interval; - else - return REG_BADBR; - } - - /* We just parsed a valid interval. */ - - /* If it's invalid to have no preceding re. */ - if (!laststart) - { - if (syntax & RE_CONTEXT_INVALID_OPS) - return REG_BADRPT; - else if (syntax & RE_CONTEXT_INDEP_OPS) - laststart = b; - else - goto unfetch_interval; - } - - /* If the upper bound is zero, don't want to succeed at - all; jump from `laststart' to `b + 3', which will be - the end of the buffer after we insert the jump. */ - if (upper_bound == 0) - { - GET_BUFFER_SPACE (3); - INSERT_JUMP (jump, laststart, b + 3); - b += 3; - } - - /* Otherwise, we have a nontrivial interval. When - we're all done, the pattern will look like: - set_number_at - set_number_at - succeed_n - - jump_n - (The upper bound and `jump_n' are omitted if - `upper_bound' is 1, though.) */ - else - { /* If the upper bound is > 1, we need to insert - more at the end of the loop. */ - unsigned nbytes = 10 + (upper_bound > 1) * 10; - - GET_BUFFER_SPACE (nbytes); - - /* Initialize lower bound of the `succeed_n', even - though it will be set during matching by its - attendant `set_number_at' (inserted next), - because `re_compile_fastmap' needs to know. - Jump to the `jump_n' we might insert below. */ - INSERT_JUMP2 (succeed_n, laststart, - b + 5 + (upper_bound > 1) * 5, - lower_bound); - b += 5; - - /* Code to initialize the lower bound. Insert - before the `succeed_n'. The `5' is the last two - bytes of this `set_number_at', plus 3 bytes of - the following `succeed_n'. */ - insert_op2 (set_number_at, laststart, 5, lower_bound, b); - b += 5; - - if (upper_bound > 1) - { /* More than one repetition is allowed, so - append a backward jump to the `succeed_n' - that starts this interval. - - When we've reached this during matching, - we'll have matched the interval once, so - jump back only `upper_bound - 1' times. */ - STORE_JUMP2 (jump_n, b, laststart + 5, - upper_bound - 1); - b += 5; - - /* The location we want to set is the second - parameter of the `jump_n'; that is `b-2' as - an absolute address. `laststart' will be - the `set_number_at' we're about to insert; - `laststart+3' the number to set, the source - for the relative address. But we are - inserting into the middle of the pattern -- - so everything is getting moved up by 5. - Conclusion: (b - 2) - (laststart + 3) + 5, - i.e., b - laststart. - - We insert this at the beginning of the loop - so that if we fail during matching, we'll - reinitialize the bounds. */ - insert_op2 (set_number_at, laststart, b - laststart, - upper_bound - 1, b); - b += 5; - } - } - pending_exact = 0; - beg_interval = NULL; - } - break; - - unfetch_interval: - /* If an invalid interval, match the characters as literals. */ - assert (beg_interval); - p = beg_interval; - beg_interval = NULL; - - /* normal_char and normal_backslash need `c'. */ - PATFETCH (c); - - if (!(syntax & RE_NO_BK_BRACES)) - { - if (p > pattern && p[-1] == '\\') - goto normal_backslash; - } - goto normal_char; - -#ifdef emacs - /* There is no way to specify the before_dot and after_dot - operators. rms says this is ok. --karl */ - case '=': - BUF_PUSH (at_dot); - break; - - case 's': - laststart = b; - PATFETCH (c); - BUF_PUSH_2 (syntaxspec, syntax_spec_code[c]); - break; - - case 'S': - laststart = b; - PATFETCH (c); - BUF_PUSH_2 (notsyntaxspec, syntax_spec_code[c]); - break; -#endif /* emacs */ - - - case 'w': - laststart = b; - BUF_PUSH (wordchar); - break; - - - case 'W': - laststart = b; - BUF_PUSH (notwordchar); - break; - - - case '<': - BUF_PUSH (wordbeg); - break; - - case '>': - BUF_PUSH (wordend); - break; - - case 'b': - BUF_PUSH (wordbound); - break; - - case 'B': - BUF_PUSH (notwordbound); - break; - - case '`': - BUF_PUSH (begbuf); - break; - - case '\'': - BUF_PUSH (endbuf); - break; - - case '1': case '2': case '3': case '4': case '5': - case '6': case '7': case '8': case '9': - if (syntax & RE_NO_BK_REFS) - goto normal_char; - - c1 = c - '0'; - - if (c1 > regnum) - return REG_ESUBREG; - - /* Can't back reference to a subexpression if inside of it. */ - if (group_in_compile_stack (compile_stack, c1)) - goto normal_char; - - laststart = b; - BUF_PUSH_2 (duplicate, c1); - break; - - - case '+': - case '?': - if (syntax & RE_BK_PLUS_QM) - goto handle_plus; - else - goto normal_backslash; - - default: - normal_backslash: - /* You might think it would be useful for \ to mean - not to translate; but if we don't translate it - it will never match anything. */ - c = TRANSLATE (c); - goto normal_char; - } - break; - - - default: - /* Expects the character in `c'. */ - normal_char: - /* If no exactn currently being built. */ - if (!pending_exact - - /* If last exactn not at current position. */ - || pending_exact + *pending_exact + 1 != b - - /* We have only one byte following the exactn for the count. */ - || *pending_exact == (1 << BYTEWIDTH) - 1 - - /* If followed by a repetition operator. */ - || *p == '*' || *p == '^' - || ((syntax & RE_BK_PLUS_QM) - ? *p == '\\' && (p[1] == '+' || p[1] == '?') - : (*p == '+' || *p == '?')) - || ((syntax & RE_INTERVALS) - && ((syntax & RE_NO_BK_BRACES) - ? *p == '{' - : (p[0] == '\\' && p[1] == '{')))) - { - /* Start building a new exactn. */ - - laststart = b; - - BUF_PUSH_2 (exactn, 0); - pending_exact = b - 1; - } - - BUF_PUSH (c); - (*pending_exact)++; - break; - } /* switch (c) */ - } /* while p != pend */ - - - /* Through the pattern now. */ - - if (fixup_alt_jump) - STORE_JUMP (jump_past_alt, fixup_alt_jump, b); - - if (!COMPILE_STACK_EMPTY) - return REG_EPAREN; - - free (compile_stack.stack); - - /* We have succeeded; set the length of the buffer. */ - bufp->used = b - bufp->buffer; - -#ifdef DEBUG - if (debug) - { - DEBUG_PRINT1 ("\nCompiled pattern: "); - print_compiled_pattern (bufp); - } -#endif /* DEBUG */ - - return REG_NOERROR; -} /* regex_compile */ - -/* Subroutines for `regex_compile'. */ - -/* Store OP at LOC followed by two-byte integer parameter ARG. */ - -static void -store_op1 (op, loc, arg) - re_opcode_t op; - unsigned char *loc; - int arg; -{ - *loc = (unsigned char) op; - STORE_NUMBER (loc + 1, arg); -} - - -/* Like `store_op1', but for two two-byte parameters ARG1 and ARG2. */ - -static void -store_op2 (op, loc, arg1, arg2) - re_opcode_t op; - unsigned char *loc; - int arg1, arg2; -{ - *loc = (unsigned char) op; - STORE_NUMBER (loc + 1, arg1); - STORE_NUMBER (loc + 3, arg2); -} - - -/* Copy the bytes from LOC to END to open up three bytes of space at LOC - for OP followed by two-byte integer parameter ARG. */ - -static void -insert_op1 (op, loc, arg, end) - re_opcode_t op; - unsigned char *loc; - int arg; - unsigned char *end; -{ - register unsigned char *pfrom = end; - register unsigned char *pto = end + 3; - - while (pfrom != loc) - *--pto = *--pfrom; - - store_op1 (op, loc, arg); -} - - -/* Like `insert_op1', but for two two-byte parameters ARG1 and ARG2. */ - -static void -insert_op2 (op, loc, arg1, arg2, end) - re_opcode_t op; - unsigned char *loc; - int arg1, arg2; - unsigned char *end; -{ - register unsigned char *pfrom = end; - register unsigned char *pto = end + 5; - - while (pfrom != loc) - *--pto = *--pfrom; - - store_op2 (op, loc, arg1, arg2); -} - - -/* P points to just after a ^ in PATTERN. Return true if that ^ comes - after an alternative or a begin-subexpression. We assume there is at - least one character before the ^. */ - -static boolean -at_begline_loc_p (pattern, p, syntax) - const char *pattern, *p; - reg_syntax_t syntax; -{ - const char *prev = p - 2; - boolean prev_prev_backslash = prev > pattern && prev[-1] == '\\'; - - return - /* After a subexpression? */ - (*prev == '(' && (syntax & RE_NO_BK_PARENS || prev_prev_backslash)) - /* After an alternative? */ - || (*prev == '|' && (syntax & RE_NO_BK_VBAR || prev_prev_backslash)); -} - - -/* The dual of at_begline_loc_p. This one is for $. We assume there is - at least one character after the $, i.e., `P < PEND'. */ - -static boolean -at_endline_loc_p (p, pend, syntax) - const char *p, *pend; - int syntax; -{ - const char *next = p; - boolean next_backslash = *next == '\\'; - const char *next_next = p + 1 < pend ? p + 1 : NULL; - - return - /* Before a subexpression? */ - (syntax & RE_NO_BK_PARENS ? *next == ')' - : next_backslash && next_next && *next_next == ')') - /* Before an alternative? */ - || (syntax & RE_NO_BK_VBAR ? *next == '|' - : next_backslash && next_next && *next_next == '|'); -} - - -/* Returns true if REGNUM is in one of COMPILE_STACK's elements and - false if it's not. */ - -static boolean -group_in_compile_stack (compile_stack, regnum) - compile_stack_type compile_stack; - regnum_t regnum; -{ - int this_element; - - for (this_element = compile_stack.avail - 1; - this_element >= 0; - this_element--) - if (compile_stack.stack[this_element].regnum == regnum) - return true; - - return false; -} - - -/* Read the ending character of a range (in a bracket expression) from the - uncompiled pattern *P_PTR (which ends at PEND). We assume the - starting character is in `P[-2]'. (`P[-1]' is the character `-'.) - Then we set the translation of all bits between the starting and - ending characters (inclusive) in the compiled pattern B. - - Return an error code. - - We use these short variable names so we can use the same macros as - `regex_compile' itself. */ - -static reg_errcode_t -compile_range (p_ptr, pend, translate, syntax, b) - const char **p_ptr, *pend; - char *translate; - reg_syntax_t syntax; - unsigned char *b; -{ - unsigned this_char; - - const char *p = *p_ptr; - int range_start, range_end; - - if (p == pend) - return REG_ERANGE; - - /* Even though the pattern is a signed `char *', we need to fetch - with unsigned char *'s; if the high bit of the pattern character - is set, the range endpoints will be negative if we fetch using a - signed char *. - - We also want to fetch the endpoints without translating them; the - appropriate translation is done in the bit-setting loop below. */ - range_start = ((unsigned char *) p)[-2]; - range_end = ((unsigned char *) p)[0]; - - /* Have to increment the pointer into the pattern string, so the - caller isn't still at the ending character. */ - (*p_ptr)++; - - /* If the start is after the end, the range is empty. */ - if (range_start > range_end) - return syntax & RE_NO_EMPTY_RANGES ? REG_ERANGE : REG_NOERROR; - - /* Here we see why `this_char' has to be larger than an `unsigned - char' -- the range is inclusive, so if `range_end' == 0xff - (assuming 8-bit characters), we would otherwise go into an infinite - loop, since all characters <= 0xff. */ - for (this_char = range_start; this_char <= range_end; this_char++) - { - SET_LIST_BIT (TRANSLATE (this_char)); - } - - return REG_NOERROR; -} - -/* Failure stack declarations and macros; both re_compile_fastmap and - re_match_2 use a failure stack. These have to be macros because of - REGEX_ALLOCATE. */ - - -/* Number of failure points for which to initially allocate space - when matching. If this number is exceeded, we allocate more - space, so it is not a hard limit. */ -#ifndef INIT_FAILURE_ALLOC -#define INIT_FAILURE_ALLOC 5 -#endif - -/* Roughly the maximum number of failure points on the stack. Would be - exactly that if always used MAX_FAILURE_SPACE each time we failed. - This is a variable only so users of regex can assign to it; we never - change it ourselves. */ -int re_max_failures = 2000; - -typedef const unsigned char *fail_stack_elt_t; - -typedef struct -{ - fail_stack_elt_t *stack; - unsigned size; - unsigned avail; /* Offset of next open position. */ -} fail_stack_type; - -#define FAIL_STACK_EMPTY() (fail_stack.avail == 0) -#define FAIL_STACK_PTR_EMPTY() (fail_stack_ptr->avail == 0) -#define FAIL_STACK_FULL() (fail_stack.avail == fail_stack.size) -#define FAIL_STACK_TOP() (fail_stack.stack[fail_stack.avail]) - - -/* Initialize `fail_stack'. Do `return -2' if the alloc fails. */ - -#define INIT_FAIL_STACK() \ - do { \ - fail_stack.stack = (fail_stack_elt_t *) \ - REGEX_ALLOCATE (INIT_FAILURE_ALLOC * sizeof (fail_stack_elt_t)); \ - \ - if (fail_stack.stack == NULL) \ - return -2; \ - \ - fail_stack.size = INIT_FAILURE_ALLOC; \ - fail_stack.avail = 0; \ - } while (0) - - -/* Double the size of FAIL_STACK, up to approximately `re_max_failures' items. - - Return 1 if succeeds, and 0 if either ran out of memory - allocating space for it or it was already too large. - - REGEX_REALLOCATE requires `destination' be declared. */ - -#define DOUBLE_FAIL_STACK(fail_stack) \ - ((fail_stack).size > re_max_failures * MAX_FAILURE_ITEMS \ - ? 0 \ - : ((fail_stack).stack = (fail_stack_elt_t *) \ - REGEX_REALLOCATE ((fail_stack).stack, \ - (fail_stack).size * sizeof (fail_stack_elt_t), \ - ((fail_stack).size << 1) * sizeof (fail_stack_elt_t)), \ - \ - (fail_stack).stack == NULL \ - ? 0 \ - : ((fail_stack).size <<= 1, \ - 1))) - - -/* Push PATTERN_OP on FAIL_STACK. - - Return 1 if was able to do so and 0 if ran out of memory allocating - space to do so. */ -#define PUSH_PATTERN_OP(pattern_op, fail_stack) \ - ((FAIL_STACK_FULL () \ - && !DOUBLE_FAIL_STACK (fail_stack)) \ - ? 0 \ - : ((fail_stack).stack[(fail_stack).avail++] = pattern_op, \ - 1)) - -/* This pushes an item onto the failure stack. Must be a four-byte - value. Assumes the variable `fail_stack'. Probably should only - be called from within `PUSH_FAILURE_POINT'. */ -#define PUSH_FAILURE_ITEM(item) \ - fail_stack.stack[fail_stack.avail++] = (fail_stack_elt_t) item - -/* The complement operation. Assumes `fail_stack' is nonempty. */ -#define POP_FAILURE_ITEM() fail_stack.stack[--fail_stack.avail] - -/* Used to omit pushing failure point id's when we're not debugging. */ -#ifdef DEBUG -#define DEBUG_PUSH PUSH_FAILURE_ITEM -#define DEBUG_POP(item_addr) *(item_addr) = POP_FAILURE_ITEM () -#else -#define DEBUG_PUSH(item) -#define DEBUG_POP(item_addr) -#endif - - -/* Push the information about the state we will need - if we ever fail back to it. - - Requires variables fail_stack, regstart, regend, reg_info, and - num_regs be declared. DOUBLE_FAIL_STACK requires `destination' be - declared. - - Does `return FAILURE_CODE' if runs out of memory. */ - -#define PUSH_FAILURE_POINT(pattern_place, string_place, failure_code) \ - do { \ - char *destination; \ - /* Must be int, so when we don't save any registers, the arithmetic \ - of 0 + -1 isn't done as unsigned. */ \ - int this_reg; \ - \ - DEBUG_STATEMENT (failure_id++); \ - DEBUG_STATEMENT (nfailure_points_pushed++); \ - DEBUG_PRINT2 ("\nPUSH_FAILURE_POINT #%u:\n", failure_id); \ - DEBUG_PRINT2 (" Before push, next avail: %d\n", (fail_stack).avail);\ - DEBUG_PRINT2 (" size: %d\n", (fail_stack).size);\ - \ - DEBUG_PRINT2 (" slots needed: %d\n", NUM_FAILURE_ITEMS); \ - DEBUG_PRINT2 (" available: %d\n", REMAINING_AVAIL_SLOTS); \ - \ - /* Ensure we have enough space allocated for what we will push. */ \ - while (REMAINING_AVAIL_SLOTS < NUM_FAILURE_ITEMS) \ - { \ - if (!DOUBLE_FAIL_STACK (fail_stack)) \ - return failure_code; \ - \ - DEBUG_PRINT2 ("\n Doubled stack; size now: %d\n", \ - (fail_stack).size); \ - DEBUG_PRINT2 (" slots available: %d\n", REMAINING_AVAIL_SLOTS);\ - } \ - \ - /* Push the info, starting with the registers. */ \ - DEBUG_PRINT1 ("\n"); \ - \ - for (this_reg = lowest_active_reg; this_reg <= highest_active_reg; \ - this_reg++) \ - { \ - DEBUG_PRINT2 (" Pushing reg: %d\n", this_reg); \ - DEBUG_STATEMENT (num_regs_pushed++); \ - \ - DEBUG_PRINT2 (" start: 0x%x\n", regstart[this_reg]); \ - PUSH_FAILURE_ITEM (regstart[this_reg]); \ - \ - DEBUG_PRINT2 (" end: 0x%x\n", regend[this_reg]); \ - PUSH_FAILURE_ITEM (regend[this_reg]); \ - \ - DEBUG_PRINT2 (" info: 0x%x\n ", reg_info[this_reg]); \ - DEBUG_PRINT2 (" match_null=%d", \ - REG_MATCH_NULL_STRING_P (reg_info[this_reg])); \ - DEBUG_PRINT2 (" active=%d", IS_ACTIVE (reg_info[this_reg])); \ - DEBUG_PRINT2 (" matched_something=%d", \ - MATCHED_SOMETHING (reg_info[this_reg])); \ - DEBUG_PRINT2 (" ever_matched=%d", \ - EVER_MATCHED_SOMETHING (reg_info[this_reg])); \ - DEBUG_PRINT1 ("\n"); \ - PUSH_FAILURE_ITEM (reg_info[this_reg].word); \ - } \ - \ - DEBUG_PRINT2 (" Pushing low active reg: %d\n", lowest_active_reg);\ - PUSH_FAILURE_ITEM (lowest_active_reg); \ - \ - DEBUG_PRINT2 (" Pushing high active reg: %d\n", highest_active_reg);\ - PUSH_FAILURE_ITEM (highest_active_reg); \ - \ - DEBUG_PRINT2 (" Pushing pattern 0x%x: ", pattern_place); \ - DEBUG_PRINT_COMPILED_PATTERN (bufp, pattern_place, pend); \ - PUSH_FAILURE_ITEM (pattern_place); \ - \ - DEBUG_PRINT2 (" Pushing string 0x%x: `", string_place); \ - DEBUG_PRINT_DOUBLE_STRING (string_place, string1, size1, string2, \ - size2); \ - DEBUG_PRINT1 ("'\n"); \ - PUSH_FAILURE_ITEM (string_place); \ - \ - DEBUG_PRINT2 (" Pushing failure id: %u\n", failure_id); \ - DEBUG_PUSH (failure_id); \ - } while (0) - -/* This is the number of items that are pushed and popped on the stack - for each register. */ -#define NUM_REG_ITEMS 3 - -/* Individual items aside from the registers. */ -#ifdef DEBUG -#define NUM_NONREG_ITEMS 5 /* Includes failure point id. */ -#else -#define NUM_NONREG_ITEMS 4 -#endif - -/* We push at most this many items on the stack. */ -#define MAX_FAILURE_ITEMS ((num_regs - 1) * NUM_REG_ITEMS + NUM_NONREG_ITEMS) - -/* We actually push this many items. */ -#define NUM_FAILURE_ITEMS \ - ((highest_active_reg - lowest_active_reg + 1) * NUM_REG_ITEMS \ - + NUM_NONREG_ITEMS) - -/* How many items can still be added to the stack without overflowing it. */ -#define REMAINING_AVAIL_SLOTS ((fail_stack).size - (fail_stack).avail) - - -/* Pops what PUSH_FAIL_STACK pushes. - - We restore into the parameters, all of which should be lvalues: - STR -- the saved data position. - PAT -- the saved pattern position. - LOW_REG, HIGH_REG -- the highest and lowest active registers. - REGSTART, REGEND -- arrays of string positions. - REG_INFO -- array of information about each subexpression. - - Also assumes the variables `fail_stack' and (if debugging), `bufp', - `pend', `string1', `size1', `string2', and `size2'. */ - -#define POP_FAILURE_POINT(str, pat, low_reg, high_reg, regstart, regend, reg_info)\ -{ \ - DEBUG_STATEMENT (fail_stack_elt_t failure_id;) \ - int this_reg; \ - const unsigned char *string_temp; \ - \ - assert (!FAIL_STACK_EMPTY ()); \ - \ - /* Remove failure points and point to how many regs pushed. */ \ - DEBUG_PRINT1 ("POP_FAILURE_POINT:\n"); \ - DEBUG_PRINT2 (" Before pop, next avail: %d\n", fail_stack.avail); \ - DEBUG_PRINT2 (" size: %d\n", fail_stack.size); \ - \ - assert (fail_stack.avail >= NUM_NONREG_ITEMS); \ - \ - DEBUG_POP (&failure_id); \ - DEBUG_PRINT2 (" Popping failure id: %u\n", failure_id); \ - \ - /* If the saved string location is NULL, it came from an \ - on_failure_keep_string_jump opcode, and we want to throw away the \ - saved NULL, thus retaining our current position in the string. */ \ - string_temp = POP_FAILURE_ITEM (); \ - if (string_temp != NULL) \ - str = (const char *) string_temp; \ - \ - DEBUG_PRINT2 (" Popping string 0x%x: `", str); \ - DEBUG_PRINT_DOUBLE_STRING (str, string1, size1, string2, size2); \ - DEBUG_PRINT1 ("'\n"); \ - \ - pat = (unsigned char *) POP_FAILURE_ITEM (); \ - DEBUG_PRINT2 (" Popping pattern 0x%x: ", pat); \ - DEBUG_PRINT_COMPILED_PATTERN (bufp, pat, pend); \ - \ - /* Restore register info. */ \ - high_reg = (unsigned) POP_FAILURE_ITEM (); \ - DEBUG_PRINT2 (" Popping high active reg: %d\n", high_reg); \ - \ - low_reg = (unsigned) POP_FAILURE_ITEM (); \ - DEBUG_PRINT2 (" Popping low active reg: %d\n", low_reg); \ - \ - for (this_reg = high_reg; this_reg >= low_reg; this_reg--) \ - { \ - DEBUG_PRINT2 (" Popping reg: %d\n", this_reg); \ - \ - reg_info[this_reg].word = POP_FAILURE_ITEM (); \ - DEBUG_PRINT2 (" info: 0x%x\n", reg_info[this_reg]); \ - \ - regend[this_reg] = (const char *) POP_FAILURE_ITEM (); \ - DEBUG_PRINT2 (" end: 0x%x\n", regend[this_reg]); \ - \ - regstart[this_reg] = (const char *) POP_FAILURE_ITEM (); \ - DEBUG_PRINT2 (" start: 0x%x\n", regstart[this_reg]); \ - } \ - \ - DEBUG_STATEMENT (nfailure_points_popped++); \ -} /* POP_FAILURE_POINT */ - -/* re_compile_fastmap computes a ``fastmap'' for the compiled pattern in - BUFP. A fastmap records which of the (1 << BYTEWIDTH) possible - characters can start a string that matches the pattern. This fastmap - is used by re_search to skip quickly over impossible starting points. - - The caller must supply the address of a (1 << BYTEWIDTH)-byte data - area as BUFP->fastmap. - - We set the `fastmap', `fastmap_accurate', and `can_be_null' fields in - the pattern buffer. - - Returns 0 if we succeed, -2 if an internal error. */ - -int -re_compile_fastmap (bufp) - struct re_pattern_buffer *bufp; -{ - int j, k; - fail_stack_type fail_stack; -#ifndef REGEX_MALLOC - char *destination; -#endif - /* We don't push any register information onto the failure stack. */ - unsigned num_regs = 0; - - register char *fastmap = bufp->fastmap; - unsigned char *pattern = bufp->buffer; - unsigned long size = bufp->used; - const unsigned char *p = pattern; - register unsigned char *pend = pattern + size; - - /* Assume that each path through the pattern can be null until - proven otherwise. We set this false at the bottom of switch - statement, to which we get only if a particular path doesn't - match the empty string. */ - boolean path_can_be_null = true; - - /* We aren't doing a `succeed_n' to begin with. */ - boolean succeed_n_p = false; - - assert (fastmap != NULL && p != NULL); - - INIT_FAIL_STACK (); - bzero (fastmap, 1 << BYTEWIDTH); /* Assume nothing's valid. */ - bufp->fastmap_accurate = 1; /* It will be when we're done. */ - bufp->can_be_null = 0; - - while (p != pend || !FAIL_STACK_EMPTY ()) - { - if (p == pend) - { - bufp->can_be_null |= path_can_be_null; - - /* Reset for next path. */ - path_can_be_null = true; - - p = fail_stack.stack[--fail_stack.avail]; - } - - /* We should never be about to go beyond the end of the pattern. */ - assert (p < pend); - -#ifdef SWITCH_ENUM_BUG - switch ((int) ((re_opcode_t) *p++)) -#else - switch ((re_opcode_t) *p++) -#endif - { - - /* I guess the idea here is to simply not bother with a fastmap - if a backreference is used, since it's too hard to figure out - the fastmap for the corresponding group. Setting - `can_be_null' stops `re_search_2' from using the fastmap, so - that is all we do. */ - case duplicate: - bufp->can_be_null = 1; - return 0; - - - /* Following are the cases which match a character. These end - with `break'. */ - - case exactn: - fastmap[p[1]] = 1; - break; - - - case charset: - for (j = *p++ * BYTEWIDTH - 1; j >= 0; j--) - if (p[j / BYTEWIDTH] & (1 << (j % BYTEWIDTH))) - fastmap[j] = 1; - break; - - - case charset_not: - /* Chars beyond end of map must be allowed. */ - for (j = *p * BYTEWIDTH; j < (1 << BYTEWIDTH); j++) - fastmap[j] = 1; - - for (j = *p++ * BYTEWIDTH - 1; j >= 0; j--) - if (!(p[j / BYTEWIDTH] & (1 << (j % BYTEWIDTH)))) - fastmap[j] = 1; - break; - - - case wordchar: - for (j = 0; j < (1 << BYTEWIDTH); j++) - if (SYNTAX (j) == Sword) - fastmap[j] = 1; - break; - - - case notwordchar: - for (j = 0; j < (1 << BYTEWIDTH); j++) - if (SYNTAX (j) != Sword) - fastmap[j] = 1; - break; - - - case anychar: - /* `.' matches anything ... */ - for (j = 0; j < (1 << BYTEWIDTH); j++) - fastmap[j] = 1; - - /* ... except perhaps newline. */ - if (!(bufp->syntax & RE_DOT_NEWLINE)) - fastmap['\n'] = 0; - - /* Return if we have already set `can_be_null'; if we have, - then the fastmap is irrelevant. Something's wrong here. */ - else if (bufp->can_be_null) - return 0; - - /* Otherwise, have to check alternative paths. */ - break; - - -#ifdef emacs - case syntaxspec: - k = *p++; - for (j = 0; j < (1 << BYTEWIDTH); j++) - if (SYNTAX (j) == (enum syntaxcode) k) - fastmap[j] = 1; - break; - - - case notsyntaxspec: - k = *p++; - for (j = 0; j < (1 << BYTEWIDTH); j++) - if (SYNTAX (j) != (enum syntaxcode) k) - fastmap[j] = 1; - break; - - - /* All cases after this match the empty string. These end with - `continue'. */ - - - case before_dot: - case at_dot: - case after_dot: - continue; -#endif /* not emacs */ - - - case no_op: - case begline: - case endline: - case begbuf: - case endbuf: - case wordbound: - case notwordbound: - case wordbeg: - case wordend: - case push_dummy_failure: - continue; - - - case jump_n: - case pop_failure_jump: - case maybe_pop_jump: - case jump: - case jump_past_alt: - case dummy_failure_jump: - EXTRACT_NUMBER_AND_INCR (j, p); - p += j; - if (j > 0) - continue; - - /* Jump backward implies we just went through the body of a - loop and matched nothing. Opcode jumped to should be - `on_failure_jump' or `succeed_n'. Just treat it like an - ordinary jump. For a * loop, it has pushed its failure - point already; if so, discard that as redundant. */ - if ((re_opcode_t) *p != on_failure_jump - && (re_opcode_t) *p != succeed_n) - continue; - - p++; - EXTRACT_NUMBER_AND_INCR (j, p); - p += j; - - /* If what's on the stack is where we are now, pop it. */ - if (!FAIL_STACK_EMPTY () - && fail_stack.stack[fail_stack.avail - 1] == p) - fail_stack.avail--; - - continue; - - - case on_failure_jump: - case on_failure_keep_string_jump: - handle_on_failure_jump: - EXTRACT_NUMBER_AND_INCR (j, p); - - /* For some patterns, e.g., `(a?)?', `p+j' here points to the - end of the pattern. We don't want to push such a point, - since when we restore it above, entering the switch will - increment `p' past the end of the pattern. We don't need - to push such a point since we obviously won't find any more - fastmap entries beyond `pend'. Such a pattern can match - the null string, though. */ - if (p + j < pend) - { - if (!PUSH_PATTERN_OP (p + j, fail_stack)) - return -2; - } - else - bufp->can_be_null = 1; - - if (succeed_n_p) - { - EXTRACT_NUMBER_AND_INCR (k, p); /* Skip the n. */ - succeed_n_p = false; - } - - continue; - - - case succeed_n: - /* Get to the number of times to succeed. */ - p += 2; - - /* Increment p past the n for when k != 0. */ - EXTRACT_NUMBER_AND_INCR (k, p); - if (k == 0) - { - p -= 4; - succeed_n_p = true; /* Spaghetti code alert. */ - goto handle_on_failure_jump; - } - continue; - - - case set_number_at: - p += 4; - continue; - - - case start_memory: - case stop_memory: - p += 2; - continue; - - - default: - abort (); /* We have listed all the cases. */ - } /* switch *p++ */ - - /* Getting here means we have found the possible starting - characters for one path of the pattern -- and that the empty - string does not match. We need not follow this path further. - Instead, look at the next alternative (remembered on the - stack), or quit if no more. The test at the top of the loop - does these things. */ - path_can_be_null = false; - p = pend; - } /* while p */ - - /* Set `can_be_null' for the last path (also the first path, if the - pattern is empty). */ - bufp->can_be_null |= path_can_be_null; - return 0; -} /* re_compile_fastmap */ - -/* Set REGS to hold NUM_REGS registers, storing them in STARTS and - ENDS. Subsequent matches using PATTERN_BUFFER and REGS will use - this memory for recording register information. STARTS and ENDS - must be allocated using the malloc library routine, and must each - be at least NUM_REGS * sizeof (regoff_t) bytes long. - - If NUM_REGS == 0, then subsequent matches should allocate their own - register data. - - Unless this function is called, the first search or match using - PATTERN_BUFFER will allocate its own register data, without - freeing the old data. */ - -void -re_set_registers (bufp, regs, num_regs, starts, ends) - struct re_pattern_buffer *bufp; - struct re_registers *regs; - unsigned num_regs; - regoff_t *starts, *ends; -{ - if (num_regs) - { - bufp->regs_allocated = REGS_REALLOCATE; - regs->num_regs = num_regs; - regs->start = starts; - regs->end = ends; - } - else - { - bufp->regs_allocated = REGS_UNALLOCATED; - regs->num_regs = 0; - regs->start = regs->end = (regoff_t) 0; - } -} - -/* Searching routines. */ - -/* Like re_search_2, below, but only one string is specified, and - doesn't let you say where to stop matching. */ - -int -re_search (bufp, string, size, startpos, range, regs) - struct re_pattern_buffer *bufp; - const char *string; - int size, startpos, range; - struct re_registers *regs; -{ - return re_search_2 (bufp, NULL, 0, string, size, startpos, range, - regs, size); -} - - -/* Using the compiled pattern in BUFP->buffer, first tries to match the - virtual concatenation of STRING1 and STRING2, starting first at index - STARTPOS, then at STARTPOS + 1, and so on. - - STRING1 and STRING2 have length SIZE1 and SIZE2, respectively. - - RANGE is how far to scan while trying to match. RANGE = 0 means try - only at STARTPOS; in general, the last start tried is STARTPOS + - RANGE. - - In REGS, return the indices of the virtual concatenation of STRING1 - and STRING2 that matched the entire BUFP->buffer and its contained - subexpressions. - - Do not consider matching one past the index STOP in the virtual - concatenation of STRING1 and STRING2. - - We return either the position in the strings at which the match was - found, -1 if no match, or -2 if error (such as failure - stack overflow). */ - -int -re_search_2 (bufp, string1, size1, string2, size2, startpos, range, regs, stop) - struct re_pattern_buffer *bufp; - const char *string1, *string2; - int size1, size2; - int startpos; - int range; - struct re_registers *regs; - int stop; -{ - int val; - register char *fastmap = bufp->fastmap; - register char *translate = bufp->translate; - int total_size = size1 + size2; - int endpos = startpos + range; - - /* Check for out-of-range STARTPOS. */ - if (startpos < 0 || startpos > total_size) - return -1; - - /* Fix up RANGE if it might eventually take us outside - the virtual concatenation of STRING1 and STRING2. */ - if (endpos < -1) - range = -1 - startpos; - else if (endpos > total_size) - range = total_size - startpos; - - /* If the search isn't to be a backwards one, don't waste time in a - search for a pattern that must be anchored. */ - if (bufp->used > 0 && (re_opcode_t) bufp->buffer[0] == begbuf && range > 0) - { - if (startpos > 0) - return -1; - else - range = 1; - } - - /* Update the fastmap now if not correct already. */ - if (fastmap && !bufp->fastmap_accurate) - if (re_compile_fastmap (bufp) == -2) - return -2; - - /* Loop through the string, looking for a place to start matching. */ - for (;;) - { - /* If a fastmap is supplied, skip quickly over characters that - cannot be the start of a match. If the pattern can match the - null string, however, we don't need to skip characters; we want - the first null string. */ - if (fastmap && startpos < total_size && !bufp->can_be_null) - { - if (range > 0) /* Searching forwards. */ - { - register const char *d; - register int lim = 0; - int irange = range; - - if (startpos < size1 && startpos + range >= size1) - lim = range - (size1 - startpos); - - d = (startpos >= size1 ? string2 - size1 : string1) + startpos; - - /* Written out as an if-else to avoid testing `translate' - inside the loop. */ - if (translate) - while (range > lim - && !fastmap[(unsigned char) - translate[(unsigned char) *d++]]) - range--; - else - while (range > lim && !fastmap[(unsigned char) *d++]) - range--; - - startpos += irange - range; - } - else /* Searching backwards. */ - { - register char c = (size1 == 0 || startpos >= size1 - ? string2[startpos - size1] - : string1[startpos]); - - if (!fastmap[(unsigned char) TRANSLATE (c)]) - goto advance; - } - } - - /* If can't match the null string, and that's all we have left, fail. */ - if (range >= 0 && startpos == total_size && fastmap - && !bufp->can_be_null) - return -1; - - val = re_match_2 (bufp, string1, size1, string2, size2, - startpos, regs, stop); - if (val >= 0) - return startpos; - - if (val == -2) - return -2; - - advance: - if (!range) - break; - else if (range > 0) - { - range--; - startpos++; - } - else - { - range++; - startpos--; - } - } - return -1; -} /* re_search_2 */ - -/* Declarations and macros for re_match_2. */ - -static int bcmp_translate (); -static boolean alt_match_null_string_p (), - common_op_match_null_string_p (), - group_match_null_string_p (); - -/* Structure for per-register (a.k.a. per-group) information. - This must not be longer than one word, because we push this value - onto the failure stack. Other register information, such as the - starting and ending positions (which are addresses), and the list of - inner groups (which is a bits list) are maintained in separate - variables. - - We are making a (strictly speaking) nonportable assumption here: that - the compiler will pack our bit fields into something that fits into - the type of `word', i.e., is something that fits into one item on the - failure stack. */ -typedef union -{ - fail_stack_elt_t word; - struct - { - /* This field is one if this group can match the empty string, - zero if not. If not yet determined, `MATCH_NULL_UNSET_VALUE'. */ -#define MATCH_NULL_UNSET_VALUE 3 - unsigned match_null_string_p : 2; - unsigned is_active : 1; - unsigned matched_something : 1; - unsigned ever_matched_something : 1; - } bits; -} register_info_type; - -#define REG_MATCH_NULL_STRING_P(R) ((R).bits.match_null_string_p) -#define IS_ACTIVE(R) ((R).bits.is_active) -#define MATCHED_SOMETHING(R) ((R).bits.matched_something) -#define EVER_MATCHED_SOMETHING(R) ((R).bits.ever_matched_something) - - -/* Call this when have matched a real character; it sets `matched' flags - for the subexpressions which we are currently inside. Also records - that those subexprs have matched. */ -#define SET_REGS_MATCHED() \ - do \ - { \ - unsigned r; \ - for (r = lowest_active_reg; r <= highest_active_reg; r++) \ - { \ - MATCHED_SOMETHING (reg_info[r]) \ - = EVER_MATCHED_SOMETHING (reg_info[r]) \ - = 1; \ - } \ - } \ - while (0) - - -/* This converts PTR, a pointer into one of the search strings `string1' - and `string2' into an offset from the beginning of that string. */ -#define POINTER_TO_OFFSET(ptr) \ - (FIRST_STRING_P (ptr) ? (ptr) - string1 : (ptr) - string2 + size1) - -/* Registers are set to a sentinel when they haven't yet matched. */ -#define REG_UNSET_VALUE ((char *) -1) -#define REG_UNSET(e) ((e) == REG_UNSET_VALUE) - - -/* Macros for dealing with the split strings in re_match_2. */ - -#define MATCHING_IN_FIRST_STRING (dend == end_match_1) - -/* Call before fetching a character with *d. This switches over to - string2 if necessary. */ -#define PREFETCH() \ - while (d == dend) \ - { \ - /* End of string2 => fail. */ \ - if (dend == end_match_2) \ - goto fail; \ - /* End of string1 => advance to string2. */ \ - d = string2; \ - dend = end_match_2; \ - } - - -/* Test if at very beginning or at very end of the virtual concatenation - of `string1' and `string2'. If only one string, it's `string2'. */ -#define AT_STRINGS_BEG(d) ((d) == (size1 ? string1 : string2) || !size2) -#define AT_STRINGS_END(d) ((d) == end2) - - -/* Test if D points to a character which is word-constituent. We have - two special cases to check for: if past the end of string1, look at - the first character in string2; and if before the beginning of - string2, look at the last character in string1. */ -#define WORDCHAR_P(d) \ - (SYNTAX ((d) == end1 ? *string2 \ - : (d) == string2 - 1 ? *(end1 - 1) : *(d)) \ - == Sword) - -/* Test if the character before D and the one at D differ with respect - to being word-constituent. */ -#define AT_WORD_BOUNDARY(d) \ - (AT_STRINGS_BEG (d) || AT_STRINGS_END (d) \ - || WORDCHAR_P (d - 1) != WORDCHAR_P (d)) - - -/* Free everything we malloc. */ -#ifdef REGEX_MALLOC -#define FREE_VAR(var) if (var) free (var); var = NULL -#define FREE_VARIABLES() \ - do { \ - FREE_VAR (fail_stack.stack); \ - FREE_VAR (regstart); \ - FREE_VAR (regend); \ - FREE_VAR (old_regstart); \ - FREE_VAR (old_regend); \ - FREE_VAR (best_regstart); \ - FREE_VAR (best_regend); \ - FREE_VAR (reg_info); \ - FREE_VAR (reg_dummy); \ - FREE_VAR (reg_info_dummy); \ - } while (0) -#else /* not REGEX_MALLOC */ -/* Some MIPS systems (at least) want this to free alloca'd storage. */ -#define FREE_VARIABLES() alloca (0) -#endif /* not REGEX_MALLOC */ - - -/* These values must meet several constraints. They must not be valid - register values; since we have a limit of 255 registers (because - we use only one byte in the pattern for the register number), we can - use numbers larger than 255. They must differ by 1, because of - NUM_FAILURE_ITEMS above. And the value for the lowest register must - be larger than the value for the highest register, so we do not try - to actually save any registers when none are active. */ -#define NO_HIGHEST_ACTIVE_REG (1 << BYTEWIDTH) -#define NO_LOWEST_ACTIVE_REG (NO_HIGHEST_ACTIVE_REG + 1) - -/* Matching routines. */ - -#ifndef emacs /* Emacs never uses this. */ -/* re_match is like re_match_2 except it takes only a single string. */ - -int -re_match (bufp, string, size, pos, regs) - struct re_pattern_buffer *bufp; - const char *string; - int size, pos; - struct re_registers *regs; - { - return re_match_2 (bufp, NULL, 0, string, size, pos, regs, size); -} -#endif /* not emacs */ - - -/* re_match_2 matches the compiled pattern in BUFP against the - the (virtual) concatenation of STRING1 and STRING2 (of length SIZE1 - and SIZE2, respectively). We start matching at POS, and stop - matching at STOP. - - If REGS is non-null and the `no_sub' field of BUFP is nonzero, we - store offsets for the substring each group matched in REGS. See the - documentation for exactly how many groups we fill. - - We return -1 if no match, -2 if an internal error (such as the - failure stack overflowing). Otherwise, we return the length of the - matched substring. */ - -int -re_match_2 (bufp, string1, size1, string2, size2, pos, regs, stop) - struct re_pattern_buffer *bufp; - const char *string1, *string2; - int size1, size2; - int pos; - struct re_registers *regs; - int stop; -{ - /* General temporaries. */ - int mcnt; - unsigned char *p1; - - /* Just past the end of the corresponding string. */ - const char *end1, *end2; - - /* Pointers into string1 and string2, just past the last characters in - each to consider matching. */ - const char *end_match_1, *end_match_2; - - /* Where we are in the data, and the end of the current string. */ - const char *d, *dend; - - /* Where we are in the pattern, and the end of the pattern. */ - unsigned char *p = bufp->buffer; - register unsigned char *pend = p + bufp->used; - - /* We use this to map every character in the string. */ - char *translate = bufp->translate; - - /* Failure point stack. Each place that can handle a failure further - down the line pushes a failure point on this stack. It consists of - restart, regend, and reg_info for all registers corresponding to - the subexpressions we're currently inside, plus the number of such - registers, and, finally, two char *'s. The first char * is where - to resume scanning the pattern; the second one is where to resume - scanning the strings. If the latter is zero, the failure point is - a ``dummy''; if a failure happens and the failure point is a dummy, - it gets discarded and the next next one is tried. */ - fail_stack_type fail_stack; -#ifdef DEBUG - static unsigned failure_id = 0; - unsigned nfailure_points_pushed = 0, nfailure_points_popped = 0; -#endif - - /* We fill all the registers internally, independent of what we - return, for use in backreferences. The number here includes - an element for register zero. */ - unsigned num_regs = bufp->re_nsub + 1; - - /* The currently active registers. */ - unsigned lowest_active_reg = NO_LOWEST_ACTIVE_REG; - unsigned highest_active_reg = NO_HIGHEST_ACTIVE_REG; - - /* Information on the contents of registers. These are pointers into - the input strings; they record just what was matched (on this - attempt) by a subexpression part of the pattern, that is, the - regnum-th regstart pointer points to where in the pattern we began - matching and the regnum-th regend points to right after where we - stopped matching the regnum-th subexpression. (The zeroth register - keeps track of what the whole pattern matches.) */ - const char **regstart = NULL, **regend = NULL; - - /* If a group that's operated upon by a repetition operator fails to - match anything, then the register for its start will need to be - restored because it will have been set to wherever in the string we - are when we last see its open-group operator. Similarly for a - register's end. */ - const char **old_regstart = NULL, **old_regend = NULL; - - /* The is_active field of reg_info helps us keep track of which (possibly - nested) subexpressions we are currently in. The matched_something - field of reg_info[reg_num] helps us tell whether or not we have - matched any of the pattern so far this time through the reg_num-th - subexpression. These two fields get reset each time through any - loop their register is in. */ - register_info_type *reg_info = NULL; - - /* The following record the register info as found in the above - variables when we find a match better than any we've seen before. - This happens as we backtrack through the failure points, which in - turn happens only if we have not yet matched the entire string. */ - unsigned best_regs_set = false; - const char **best_regstart = NULL, **best_regend = NULL; - - /* Logically, this is `best_regend[0]'. But we don't want to have to - allocate space for that if we're not allocating space for anything - else (see below). Also, we never need info about register 0 for - any of the other register vectors, and it seems rather a kludge to - treat `best_regend' differently than the rest. So we keep track of - the end of the best match so far in a separate variable. We - initialize this to NULL so that when we backtrack the first time - and need to test it, it's not garbage. */ - const char *match_end = NULL; - - /* Used when we pop values we don't care about. */ - const char **reg_dummy = NULL; - register_info_type *reg_info_dummy = NULL; - -#ifdef DEBUG - /* Counts the total number of registers pushed. */ - unsigned num_regs_pushed = 0; -#endif - - DEBUG_PRINT1 ("\n\nEntering re_match_2.\n"); - - INIT_FAIL_STACK (); - - /* Do not bother to initialize all the register variables if there are - no groups in the pattern, as it takes a fair amount of time. If - there are groups, we include space for register 0 (the whole - pattern), even though we never use it, since it simplifies the - array indexing. We should fix this. */ - if (bufp->re_nsub) - { - regstart = REGEX_TALLOC (num_regs, const char *); - regend = REGEX_TALLOC (num_regs, const char *); - old_regstart = REGEX_TALLOC (num_regs, const char *); - old_regend = REGEX_TALLOC (num_regs, const char *); - best_regstart = REGEX_TALLOC (num_regs, const char *); - best_regend = REGEX_TALLOC (num_regs, const char *); - reg_info = REGEX_TALLOC (num_regs, register_info_type); - reg_dummy = REGEX_TALLOC (num_regs, const char *); - reg_info_dummy = REGEX_TALLOC (num_regs, register_info_type); - - if (!(regstart && regend && old_regstart && old_regend && reg_info - && best_regstart && best_regend && reg_dummy && reg_info_dummy)) - { - FREE_VARIABLES (); - return -2; - } - } -#ifdef REGEX_MALLOC - else - { - /* We must initialize all our variables to NULL, so that - `FREE_VARIABLES' doesn't try to free them. */ - regstart = regend = old_regstart = old_regend = best_regstart - = best_regend = reg_dummy = NULL; - reg_info = reg_info_dummy = (register_info_type *) NULL; - } -#endif /* REGEX_MALLOC */ - - /* The starting position is bogus. */ - if (pos < 0 || pos > size1 + size2) - { - FREE_VARIABLES (); - return -1; - } - - /* Initialize subexpression text positions to -1 to mark ones that no - start_memory/stop_memory has been seen for. Also initialize the - register information struct. */ - for (mcnt = 1; mcnt < num_regs; mcnt++) - { - regstart[mcnt] = regend[mcnt] - = old_regstart[mcnt] = old_regend[mcnt] = REG_UNSET_VALUE; - - REG_MATCH_NULL_STRING_P (reg_info[mcnt]) = MATCH_NULL_UNSET_VALUE; - IS_ACTIVE (reg_info[mcnt]) = 0; - MATCHED_SOMETHING (reg_info[mcnt]) = 0; - EVER_MATCHED_SOMETHING (reg_info[mcnt]) = 0; - } - - /* We move `string1' into `string2' if the latter's empty -- but not if - `string1' is null. */ - if (size2 == 0 && string1 != NULL) - { - string2 = string1; - size2 = size1; - string1 = 0; - size1 = 0; - } - end1 = string1 + size1; - end2 = string2 + size2; - - /* Compute where to stop matching, within the two strings. */ - if (stop <= size1) - { - end_match_1 = string1 + stop; - end_match_2 = string2; - } - else - { - end_match_1 = end1; - end_match_2 = string2 + stop - size1; - } - - /* `p' scans through the pattern as `d' scans through the data. - `dend' is the end of the input string that `d' points within. `d' - is advanced into the following input string whenever necessary, but - this happens before fetching; therefore, at the beginning of the - loop, `d' can be pointing at the end of a string, but it cannot - equal `string2'. */ - if (size1 > 0 && pos <= size1) - { - d = string1 + pos; - dend = end_match_1; - } - else - { - d = string2 + pos - size1; - dend = end_match_2; - } - - DEBUG_PRINT1 ("The compiled pattern is: "); - DEBUG_PRINT_COMPILED_PATTERN (bufp, p, pend); - DEBUG_PRINT1 ("The string to match is: `"); - DEBUG_PRINT_DOUBLE_STRING (d, string1, size1, string2, size2); - DEBUG_PRINT1 ("'\n"); - - /* This loops over pattern commands. It exits by returning from the - function if the match is complete, or it drops through if the match - fails at this starting point in the input data. */ - for (;;) - { - DEBUG_PRINT2 ("\n0x%x: ", p); - - if (p == pend) - { /* End of pattern means we might have succeeded. */ - DEBUG_PRINT1 ("end of pattern ... "); - - /* If we haven't matched the entire string, and we want the - longest match, try backtracking. */ - if (d != end_match_2) - { - DEBUG_PRINT1 ("backtracking.\n"); - - if (!FAIL_STACK_EMPTY ()) - { /* More failure points to try. */ - boolean same_str_p = (FIRST_STRING_P (match_end) - == MATCHING_IN_FIRST_STRING); - - /* If exceeds best match so far, save it. */ - if (!best_regs_set - || (same_str_p && d > match_end) - || (!same_str_p && !MATCHING_IN_FIRST_STRING)) - { - best_regs_set = true; - match_end = d; - - DEBUG_PRINT1 ("\nSAVING match as best so far.\n"); - - for (mcnt = 1; mcnt < num_regs; mcnt++) - { - best_regstart[mcnt] = regstart[mcnt]; - best_regend[mcnt] = regend[mcnt]; - } - } - goto fail; - } - - /* If no failure points, don't restore garbage. */ - else if (best_regs_set) - { - restore_best_regs: - /* Restore best match. It may happen that `dend == - end_match_1' while the restored d is in string2. - For example, the pattern `x.*y.*z' against the - strings `x-' and `y-z-', if the two strings are - not consecutive in memory. */ - DEBUG_PRINT1 ("Restoring best registers.\n"); - - d = match_end; - dend = ((d >= string1 && d <= end1) - ? end_match_1 : end_match_2); - - for (mcnt = 1; mcnt < num_regs; mcnt++) - { - regstart[mcnt] = best_regstart[mcnt]; - regend[mcnt] = best_regend[mcnt]; - } - } - } /* d != end_match_2 */ - - DEBUG_PRINT1 ("Accepting match.\n"); - - /* If caller wants register contents data back, do it. */ - if (regs && !bufp->no_sub) - { - /* Have the register data arrays been allocated? */ - if (bufp->regs_allocated == REGS_UNALLOCATED) - { /* No. So allocate them with malloc. We need one - extra element beyond `num_regs' for the `-1' marker - GNU code uses. */ - regs->num_regs = MAX (RE_NREGS, num_regs + 1); - regs->start = TALLOC (regs->num_regs, regoff_t); - regs->end = TALLOC (regs->num_regs, regoff_t); - if (regs->start == NULL || regs->end == NULL) - return -2; - bufp->regs_allocated = REGS_REALLOCATE; - } - else if (bufp->regs_allocated == REGS_REALLOCATE) - { /* Yes. If we need more elements than were already - allocated, reallocate them. If we need fewer, just - leave it alone. */ - if (regs->num_regs < num_regs + 1) - { - regs->num_regs = num_regs + 1; - RETALLOC (regs->start, regs->num_regs, regoff_t); - RETALLOC (regs->end, regs->num_regs, regoff_t); - if (regs->start == NULL || regs->end == NULL) - return -2; - } - } - else - assert (bufp->regs_allocated == REGS_FIXED); - - /* Convert the pointer data in `regstart' and `regend' to - indices. Register zero has to be set differently, - since we haven't kept track of any info for it. */ - if (regs->num_regs > 0) - { - regs->start[0] = pos; - regs->end[0] = (MATCHING_IN_FIRST_STRING ? d - string1 - : d - string2 + size1); - } - - /* Go through the first `min (num_regs, regs->num_regs)' - registers, since that is all we initialized. */ - for (mcnt = 1; mcnt < MIN (num_regs, regs->num_regs); mcnt++) - { - if (REG_UNSET (regstart[mcnt]) || REG_UNSET (regend[mcnt])) - regs->start[mcnt] = regs->end[mcnt] = -1; - else - { - regs->start[mcnt] = POINTER_TO_OFFSET (regstart[mcnt]); - regs->end[mcnt] = POINTER_TO_OFFSET (regend[mcnt]); - } - } - - /* If the regs structure we return has more elements than - were in the pattern, set the extra elements to -1. If - we (re)allocated the registers, this is the case, - because we always allocate enough to have at least one - -1 at the end. */ - for (mcnt = num_regs; mcnt < regs->num_regs; mcnt++) - regs->start[mcnt] = regs->end[mcnt] = -1; - } /* regs && !bufp->no_sub */ - - FREE_VARIABLES (); - DEBUG_PRINT4 ("%u failure points pushed, %u popped (%u remain).\n", - nfailure_points_pushed, nfailure_points_popped, - nfailure_points_pushed - nfailure_points_popped); - DEBUG_PRINT2 ("%u registers pushed.\n", num_regs_pushed); - - mcnt = d - pos - (MATCHING_IN_FIRST_STRING - ? string1 - : string2 - size1); - - DEBUG_PRINT2 ("Returning %d from re_match_2.\n", mcnt); - - return mcnt; - } - - /* Otherwise match next pattern command. */ -#ifdef SWITCH_ENUM_BUG - switch ((int) ((re_opcode_t) *p++)) -#else - switch ((re_opcode_t) *p++) -#endif - { - /* Ignore these. Used to ignore the n of succeed_n's which - currently have n == 0. */ - case no_op: - DEBUG_PRINT1 ("EXECUTING no_op.\n"); - break; - - - /* Match the next n pattern characters exactly. The following - byte in the pattern defines n, and the n bytes after that - are the characters to match. */ - case exactn: - mcnt = *p++; - DEBUG_PRINT2 ("EXECUTING exactn %d.\n", mcnt); - - /* This is written out as an if-else so we don't waste time - testing `translate' inside the loop. */ - if (translate) - { - do - { - PREFETCH (); - if (translate[(unsigned char) *d++] != (char) *p++) - goto fail; - } - while (--mcnt); - } - else - { - do - { - PREFETCH (); - if (*d++ != (char) *p++) goto fail; - } - while (--mcnt); - } - SET_REGS_MATCHED (); - break; - - - /* Match any character except possibly a newline or a null. */ - case anychar: - DEBUG_PRINT1 ("EXECUTING anychar.\n"); - - PREFETCH (); - - if ((!(bufp->syntax & RE_DOT_NEWLINE) && TRANSLATE (*d) == '\n') - || (bufp->syntax & RE_DOT_NOT_NULL && TRANSLATE (*d) == '\000')) - goto fail; - - SET_REGS_MATCHED (); - DEBUG_PRINT2 (" Matched `%d'.\n", *d); - d++; - break; - - - case charset: - case charset_not: - { - register unsigned char c; - boolean not = (re_opcode_t) *(p - 1) == charset_not; - - DEBUG_PRINT2 ("EXECUTING charset%s.\n", not ? "_not" : ""); - - PREFETCH (); - c = TRANSLATE (*d); /* The character to match. */ - - /* Cast to `unsigned' instead of `unsigned char' in case the - bit list is a full 32 bytes long. */ - if (c < (unsigned) (*p * BYTEWIDTH) - && p[1 + c / BYTEWIDTH] & (1 << (c % BYTEWIDTH))) - not = !not; - - p += 1 + *p; - - if (!not) goto fail; - - SET_REGS_MATCHED (); - d++; - break; - } - - - /* The beginning of a group is represented by start_memory. - The arguments are the register number in the next byte, and the - number of groups inner to this one in the next. The text - matched within the group is recorded (in the internal - registers data structure) under the register number. */ - case start_memory: - DEBUG_PRINT3 ("EXECUTING start_memory %d (%d):\n", *p, p[1]); - - /* Find out if this group can match the empty string. */ - p1 = p; /* To send to group_match_null_string_p. */ - - if (REG_MATCH_NULL_STRING_P (reg_info[*p]) == MATCH_NULL_UNSET_VALUE) - REG_MATCH_NULL_STRING_P (reg_info[*p]) - = group_match_null_string_p (&p1, pend, reg_info); - - /* Save the position in the string where we were the last time - we were at this open-group operator in case the group is - operated upon by a repetition operator, e.g., with `(a*)*b' - against `ab'; then we want to ignore where we are now in - the string in case this attempt to match fails. */ - old_regstart[*p] = REG_MATCH_NULL_STRING_P (reg_info[*p]) - ? REG_UNSET (regstart[*p]) ? d : regstart[*p] - : regstart[*p]; - DEBUG_PRINT2 (" old_regstart: %d\n", - POINTER_TO_OFFSET (old_regstart[*p])); - - regstart[*p] = d; - DEBUG_PRINT2 (" regstart: %d\n", POINTER_TO_OFFSET (regstart[*p])); - - IS_ACTIVE (reg_info[*p]) = 1; - MATCHED_SOMETHING (reg_info[*p]) = 0; - - /* This is the new highest active register. */ - highest_active_reg = *p; - - /* If nothing was active before, this is the new lowest active - register. */ - if (lowest_active_reg == NO_LOWEST_ACTIVE_REG) - lowest_active_reg = *p; - - /* Move past the register number and inner group count. */ - p += 2; - break; - - - /* The stop_memory opcode represents the end of a group. Its - arguments are the same as start_memory's: the register - number, and the number of inner groups. */ - case stop_memory: - DEBUG_PRINT3 ("EXECUTING stop_memory %d (%d):\n", *p, p[1]); - - /* We need to save the string position the last time we were at - this close-group operator in case the group is operated - upon by a repetition operator, e.g., with `((a*)*(b*)*)*' - against `aba'; then we want to ignore where we are now in - the string in case this attempt to match fails. */ - old_regend[*p] = REG_MATCH_NULL_STRING_P (reg_info[*p]) - ? REG_UNSET (regend[*p]) ? d : regend[*p] - : regend[*p]; - DEBUG_PRINT2 (" old_regend: %d\n", - POINTER_TO_OFFSET (old_regend[*p])); - - regend[*p] = d; - DEBUG_PRINT2 (" regend: %d\n", POINTER_TO_OFFSET (regend[*p])); - - /* This register isn't active anymore. */ - IS_ACTIVE (reg_info[*p]) = 0; - - /* If this was the only register active, nothing is active - anymore. */ - if (lowest_active_reg == highest_active_reg) - { - lowest_active_reg = NO_LOWEST_ACTIVE_REG; - highest_active_reg = NO_HIGHEST_ACTIVE_REG; - } - else - { /* We must scan for the new highest active register, since - it isn't necessarily one less than now: consider - (a(b)c(d(e)f)g). When group 3 ends, after the f), the - new highest active register is 1. */ - unsigned char r = *p - 1; - while (r > 0 && !IS_ACTIVE (reg_info[r])) - r--; - - /* If we end up at register zero, that means that we saved - the registers as the result of an `on_failure_jump', not - a `start_memory', and we jumped to past the innermost - `stop_memory'. For example, in ((.)*) we save - registers 1 and 2 as a result of the *, but when we pop - back to the second ), we are at the stop_memory 1. - Thus, nothing is active. */ - if (r == 0) - { - lowest_active_reg = NO_LOWEST_ACTIVE_REG; - highest_active_reg = NO_HIGHEST_ACTIVE_REG; - } - else - highest_active_reg = r; - } - - /* If just failed to match something this time around with a - group that's operated on by a repetition operator, try to - force exit from the ``loop'', and restore the register - information for this group that we had before trying this - last match. */ - if ((!MATCHED_SOMETHING (reg_info[*p]) - || (re_opcode_t) p[-3] == start_memory) - && (p + 2) < pend) - { - boolean is_a_jump_n = false; - - p1 = p + 2; - mcnt = 0; - switch ((re_opcode_t) *p1++) - { - case jump_n: - is_a_jump_n = true; - case pop_failure_jump: - case maybe_pop_jump: - case jump: - case dummy_failure_jump: - EXTRACT_NUMBER_AND_INCR (mcnt, p1); - if (is_a_jump_n) - p1 += 2; - break; - - default: - /* do nothing */ ; - } - p1 += mcnt; - - /* If the next operation is a jump backwards in the pattern - to an on_failure_jump right before the start_memory - corresponding to this stop_memory, exit from the loop - by forcing a failure after pushing on the stack the - on_failure_jump's jump in the pattern, and d. */ - if (mcnt < 0 && (re_opcode_t) *p1 == on_failure_jump - && (re_opcode_t) p1[3] == start_memory && p1[4] == *p) - { - /* If this group ever matched anything, then restore - what its registers were before trying this last - failed match, e.g., with `(a*)*b' against `ab' for - regstart[1], and, e.g., with `((a*)*(b*)*)*' - against `aba' for regend[3]. - - Also restore the registers for inner groups for, - e.g., `((a*)(b*))*' against `aba' (register 3 would - otherwise get trashed). */ - - if (EVER_MATCHED_SOMETHING (reg_info[*p])) - { - unsigned r; - - EVER_MATCHED_SOMETHING (reg_info[*p]) = 0; - - /* Restore this and inner groups' (if any) registers. */ - for (r = *p; r < *p + *(p + 1); r++) - { - regstart[r] = old_regstart[r]; - - /* xx why this test? */ - if ((int) old_regend[r] >= (int) regstart[r]) - regend[r] = old_regend[r]; - } - } - p1++; - EXTRACT_NUMBER_AND_INCR (mcnt, p1); - PUSH_FAILURE_POINT (p1 + mcnt, d, -2); - - goto fail; - } - } - - /* Move past the register number and the inner group count. */ - p += 2; - break; - - - /* \ has been turned into a `duplicate' command which is - followed by the numeric value of as the register number. */ - case duplicate: - { - register const char *d2, *dend2; - int regno = *p++; /* Get which register to match against. */ - DEBUG_PRINT2 ("EXECUTING duplicate %d.\n", regno); - - /* Can't back reference a group which we've never matched. */ - if (REG_UNSET (regstart[regno]) || REG_UNSET (regend[regno])) - goto fail; - - /* Where in input to try to start matching. */ - d2 = regstart[regno]; - - /* Where to stop matching; if both the place to start and - the place to stop matching are in the same string, then - set to the place to stop, otherwise, for now have to use - the end of the first string. */ - - dend2 = ((FIRST_STRING_P (regstart[regno]) - == FIRST_STRING_P (regend[regno])) - ? regend[regno] : end_match_1); - for (;;) - { - /* If necessary, advance to next segment in register - contents. */ - while (d2 == dend2) - { - if (dend2 == end_match_2) break; - if (dend2 == regend[regno]) break; - - /* End of string1 => advance to string2. */ - d2 = string2; - dend2 = regend[regno]; - } - /* At end of register contents => success */ - if (d2 == dend2) break; - - /* If necessary, advance to next segment in data. */ - PREFETCH (); - - /* How many characters left in this segment to match. */ - mcnt = dend - d; - - /* Want how many consecutive characters we can match in - one shot, so, if necessary, adjust the count. */ - if (mcnt > dend2 - d2) - mcnt = dend2 - d2; - - /* Compare that many; failure if mismatch, else move - past them. */ - if (translate - ? bcmp_translate (d, d2, mcnt, translate) - : bcmp (d, d2, mcnt)) - goto fail; - d += mcnt, d2 += mcnt; - } - } - break; - - - /* begline matches the empty string at the beginning of the string - (unless `not_bol' is set in `bufp'), and, if - `newline_anchor' is set, after newlines. */ - case begline: - DEBUG_PRINT1 ("EXECUTING begline.\n"); - - if (AT_STRINGS_BEG (d)) - { - if (!bufp->not_bol) break; - } - else if (d[-1] == '\n' && bufp->newline_anchor) - { - break; - } - /* In all other cases, we fail. */ - goto fail; - - - /* endline is the dual of begline. */ - case endline: - DEBUG_PRINT1 ("EXECUTING endline.\n"); - - if (AT_STRINGS_END (d)) - { - if (!bufp->not_eol) break; - } - - /* We have to ``prefetch'' the next character. */ - else if ((d == end1 ? *string2 : *d) == '\n' - && bufp->newline_anchor) - { - break; - } - goto fail; - - - /* Match at the very beginning of the data. */ - case begbuf: - DEBUG_PRINT1 ("EXECUTING begbuf.\n"); - if (AT_STRINGS_BEG (d)) - break; - goto fail; - - - /* Match at the very end of the data. */ - case endbuf: - DEBUG_PRINT1 ("EXECUTING endbuf.\n"); - if (AT_STRINGS_END (d)) - break; - goto fail; - - - /* on_failure_keep_string_jump is used to optimize `.*\n'. It - pushes NULL as the value for the string on the stack. Then - `pop_failure_point' will keep the current value for the - string, instead of restoring it. To see why, consider - matching `foo\nbar' against `.*\n'. The .* matches the foo; - then the . fails against the \n. But the next thing we want - to do is match the \n against the \n; if we restored the - string value, we would be back at the foo. - - Because this is used only in specific cases, we don't need to - check all the things that `on_failure_jump' does, to make - sure the right things get saved on the stack. Hence we don't - share its code. The only reason to push anything on the - stack at all is that otherwise we would have to change - `anychar's code to do something besides goto fail in this - case; that seems worse than this. */ - case on_failure_keep_string_jump: - DEBUG_PRINT1 ("EXECUTING on_failure_keep_string_jump"); - - EXTRACT_NUMBER_AND_INCR (mcnt, p); - DEBUG_PRINT3 (" %d (to 0x%x):\n", mcnt, p + mcnt); - - PUSH_FAILURE_POINT (p + mcnt, NULL, -2); - break; - - - /* Uses of on_failure_jump: - - Each alternative starts with an on_failure_jump that points - to the beginning of the next alternative. Each alternative - except the last ends with a jump that in effect jumps past - the rest of the alternatives. (They really jump to the - ending jump of the following alternative, because tensioning - these jumps is a hassle.) - - Repeats start with an on_failure_jump that points past both - the repetition text and either the following jump or - pop_failure_jump back to this on_failure_jump. */ - case on_failure_jump: - on_failure: - DEBUG_PRINT1 ("EXECUTING on_failure_jump"); - - EXTRACT_NUMBER_AND_INCR (mcnt, p); - DEBUG_PRINT3 (" %d (to 0x%x)", mcnt, p + mcnt); - - /* If this on_failure_jump comes right before a group (i.e., - the original * applied to a group), save the information - for that group and all inner ones, so that if we fail back - to this point, the group's information will be correct. - For example, in \(a*\)*\1, we need the preceding group, - and in \(\(a*\)b*\)\2, we need the inner group. */ - - /* We can't use `p' to check ahead because we push - a failure point to `p + mcnt' after we do this. */ - p1 = p; - - /* We need to skip no_op's before we look for the - start_memory in case this on_failure_jump is happening as - the result of a completed succeed_n, as in \(a\)\{1,3\}b\1 - against aba. */ - while (p1 < pend && (re_opcode_t) *p1 == no_op) - p1++; - - if (p1 < pend && (re_opcode_t) *p1 == start_memory) - { - /* We have a new highest active register now. This will - get reset at the start_memory we are about to get to, - but we will have saved all the registers relevant to - this repetition op, as described above. */ - highest_active_reg = *(p1 + 1) + *(p1 + 2); - if (lowest_active_reg == NO_LOWEST_ACTIVE_REG) - lowest_active_reg = *(p1 + 1); - } - - DEBUG_PRINT1 (":\n"); - PUSH_FAILURE_POINT (p + mcnt, d, -2); - break; - - - /* A smart repeat ends with `maybe_pop_jump'. - We change it to either `pop_failure_jump' or `jump'. */ - case maybe_pop_jump: - EXTRACT_NUMBER_AND_INCR (mcnt, p); - DEBUG_PRINT2 ("EXECUTING maybe_pop_jump %d.\n", mcnt); - { - register unsigned char *p2 = p; - - /* Compare the beginning of the repeat with what in the - pattern follows its end. If we can establish that there - is nothing that they would both match, i.e., that we - would have to backtrack because of (as in, e.g., `a*a') - then we can change to pop_failure_jump, because we'll - never have to backtrack. - - This is not true in the case of alternatives: in - `(a|ab)*' we do need to backtrack to the `ab' alternative - (e.g., if the string was `ab'). But instead of trying to - detect that here, the alternative has put on a dummy - failure point which is what we will end up popping. */ - - /* Skip over open/close-group commands. */ - while (p2 + 2 < pend - && ((re_opcode_t) *p2 == stop_memory - || (re_opcode_t) *p2 == start_memory)) - p2 += 3; /* Skip over args, too. */ - - /* If we're at the end of the pattern, we can change. */ - if (p2 == pend) - { - /* Consider what happens when matching ":\(.*\)" - against ":/". I don't really understand this code - yet. */ - p[-3] = (unsigned char) pop_failure_jump; - DEBUG_PRINT1 - (" End of pattern: change to `pop_failure_jump'.\n"); - } - - else if ((re_opcode_t) *p2 == exactn - || (bufp->newline_anchor && (re_opcode_t) *p2 == endline)) - { - register unsigned char c - = *p2 == (unsigned char) endline ? '\n' : p2[2]; - p1 = p + mcnt; - - /* p1[0] ... p1[2] are the `on_failure_jump' corresponding - to the `maybe_finalize_jump' of this case. Examine what - follows. */ - if ((re_opcode_t) p1[3] == exactn && p1[5] != c) - { - p[-3] = (unsigned char) pop_failure_jump; - DEBUG_PRINT3 (" %c != %c => pop_failure_jump.\n", - c, p1[5]); - } - - else if ((re_opcode_t) p1[3] == charset - || (re_opcode_t) p1[3] == charset_not) - { - int not = (re_opcode_t) p1[3] == charset_not; - - if (c < (unsigned char) (p1[4] * BYTEWIDTH) - && p1[5 + c / BYTEWIDTH] & (1 << (c % BYTEWIDTH))) - not = !not; - - /* `not' is equal to 1 if c would match, which means - that we can't change to pop_failure_jump. */ - if (!not) - { - p[-3] = (unsigned char) pop_failure_jump; - DEBUG_PRINT1 (" No match => pop_failure_jump.\n"); - } - } - } - } - p -= 2; /* Point at relative address again. */ - if ((re_opcode_t) p[-1] != pop_failure_jump) - { - p[-1] = (unsigned char) jump; - DEBUG_PRINT1 (" Match => jump.\n"); - goto unconditional_jump; - } - /* Note fall through. */ - - - /* The end of a simple repeat has a pop_failure_jump back to - its matching on_failure_jump, where the latter will push a - failure point. The pop_failure_jump takes off failure - points put on by this pop_failure_jump's matching - on_failure_jump; we got through the pattern to here from the - matching on_failure_jump, so didn't fail. */ - case pop_failure_jump: - { - /* We need to pass separate storage for the lowest and - highest registers, even though we don't care about the - actual values. Otherwise, we will restore only one - register from the stack, since lowest will == highest in - `pop_failure_point'. */ - unsigned dummy_low_reg, dummy_high_reg; - unsigned char *pdummy; - const char *sdummy; - - DEBUG_PRINT1 ("EXECUTING pop_failure_jump.\n"); - POP_FAILURE_POINT (sdummy, pdummy, - dummy_low_reg, dummy_high_reg, - reg_dummy, reg_dummy, reg_info_dummy); - } - /* Note fall through. */ - - - /* Unconditionally jump (without popping any failure points). */ - case jump: - unconditional_jump: - EXTRACT_NUMBER_AND_INCR (mcnt, p); /* Get the amount to jump. */ - DEBUG_PRINT2 ("EXECUTING jump %d ", mcnt); - p += mcnt; /* Do the jump. */ - DEBUG_PRINT2 ("(to 0x%x).\n", p); - break; - - - /* We need this opcode so we can detect where alternatives end - in `group_match_null_string_p' et al. */ - case jump_past_alt: - DEBUG_PRINT1 ("EXECUTING jump_past_alt.\n"); - goto unconditional_jump; - - - /* Normally, the on_failure_jump pushes a failure point, which - then gets popped at pop_failure_jump. We will end up at - pop_failure_jump, also, and with a pattern of, say, `a+', we - are skipping over the on_failure_jump, so we have to push - something meaningless for pop_failure_jump to pop. */ - case dummy_failure_jump: - DEBUG_PRINT1 ("EXECUTING dummy_failure_jump.\n"); - /* It doesn't matter what we push for the string here. What - the code at `fail' tests is the value for the pattern. */ - PUSH_FAILURE_POINT (0, 0, -2); - goto unconditional_jump; - - - /* At the end of an alternative, we need to push a dummy failure - point in case we are followed by a `pop_failure_jump', because - we don't want the failure point for the alternative to be - popped. For example, matching `(a|ab)*' against `aab' - requires that we match the `ab' alternative. */ - case push_dummy_failure: - DEBUG_PRINT1 ("EXECUTING push_dummy_failure.\n"); - /* See comments just above at `dummy_failure_jump' about the - two zeroes. */ - PUSH_FAILURE_POINT (0, 0, -2); - break; - - /* Have to succeed matching what follows at least n times. - After that, handle like `on_failure_jump'. */ - case succeed_n: - EXTRACT_NUMBER (mcnt, p + 2); - DEBUG_PRINT2 ("EXECUTING succeed_n %d.\n", mcnt); - - assert (mcnt >= 0); - /* Originally, this is how many times we HAVE to succeed. */ - if (mcnt > 0) - { - mcnt--; - p += 2; - STORE_NUMBER_AND_INCR (p, mcnt); - DEBUG_PRINT3 (" Setting 0x%x to %d.\n", p, mcnt); - } - else if (mcnt == 0) - { - DEBUG_PRINT2 (" Setting two bytes from 0x%x to no_op.\n", p+2); - p[2] = (unsigned char) no_op; - p[3] = (unsigned char) no_op; - goto on_failure; - } - break; - - case jump_n: - EXTRACT_NUMBER (mcnt, p + 2); - DEBUG_PRINT2 ("EXECUTING jump_n %d.\n", mcnt); - - /* Originally, this is how many times we CAN jump. */ - if (mcnt) - { - mcnt--; - STORE_NUMBER (p + 2, mcnt); - goto unconditional_jump; - } - /* If don't have to jump any more, skip over the rest of command. */ - else - p += 4; - break; - - case set_number_at: - { - DEBUG_PRINT1 ("EXECUTING set_number_at.\n"); - - EXTRACT_NUMBER_AND_INCR (mcnt, p); - p1 = p + mcnt; - EXTRACT_NUMBER_AND_INCR (mcnt, p); - DEBUG_PRINT3 (" Setting 0x%x to %d.\n", p1, mcnt); - STORE_NUMBER (p1, mcnt); - break; - } - - case wordbound: - DEBUG_PRINT1 ("EXECUTING wordbound.\n"); - if (AT_WORD_BOUNDARY (d)) - break; - goto fail; - - case notwordbound: - DEBUG_PRINT1 ("EXECUTING notwordbound.\n"); - if (AT_WORD_BOUNDARY (d)) - goto fail; - break; - - case wordbeg: - DEBUG_PRINT1 ("EXECUTING wordbeg.\n"); - if (WORDCHAR_P (d) && (AT_STRINGS_BEG (d) || !WORDCHAR_P (d - 1))) - break; - goto fail; - - case wordend: - DEBUG_PRINT1 ("EXECUTING wordend.\n"); - if (!AT_STRINGS_BEG (d) && WORDCHAR_P (d - 1) - && (!WORDCHAR_P (d) || AT_STRINGS_END (d))) - break; - goto fail; - -#ifdef emacs -#ifdef emacs19 - case before_dot: - DEBUG_PRINT1 ("EXECUTING before_dot.\n"); - if (PTR_CHAR_POS ((unsigned char *) d) >= point) - goto fail; - break; - - case at_dot: - DEBUG_PRINT1 ("EXECUTING at_dot.\n"); - if (PTR_CHAR_POS ((unsigned char *) d) != point) - goto fail; - break; - - case after_dot: - DEBUG_PRINT1 ("EXECUTING after_dot.\n"); - if (PTR_CHAR_POS ((unsigned char *) d) <= point) - goto fail; - break; -#else /* not emacs19 */ - case at_dot: - DEBUG_PRINT1 ("EXECUTING at_dot.\n"); - if (PTR_CHAR_POS ((unsigned char *) d) + 1 != point) - goto fail; - break; -#endif /* not emacs19 */ - - case syntaxspec: - DEBUG_PRINT2 ("EXECUTING syntaxspec %d.\n", mcnt); - mcnt = *p++; - goto matchsyntax; - - case wordchar: - DEBUG_PRINT1 ("EXECUTING Emacs wordchar.\n"); - mcnt = (int) Sword; - matchsyntax: - PREFETCH (); - if (SYNTAX (*d++) != (enum syntaxcode) mcnt) - goto fail; - SET_REGS_MATCHED (); - break; - - case notsyntaxspec: - DEBUG_PRINT2 ("EXECUTING notsyntaxspec %d.\n", mcnt); - mcnt = *p++; - goto matchnotsyntax; - - case notwordchar: - DEBUG_PRINT1 ("EXECUTING Emacs notwordchar.\n"); - mcnt = (int) Sword; - matchnotsyntax: - PREFETCH (); - if (SYNTAX (*d++) == (enum syntaxcode) mcnt) - goto fail; - SET_REGS_MATCHED (); - break; - -#else /* not emacs */ - case wordchar: - DEBUG_PRINT1 ("EXECUTING non-Emacs wordchar.\n"); - PREFETCH (); - if (!WORDCHAR_P (d)) - goto fail; - SET_REGS_MATCHED (); - d++; - break; - - case notwordchar: - DEBUG_PRINT1 ("EXECUTING non-Emacs notwordchar.\n"); - PREFETCH (); - if (WORDCHAR_P (d)) - goto fail; - SET_REGS_MATCHED (); - d++; - break; -#endif /* not emacs */ - - default: - abort (); - } - continue; /* Successfully executed one pattern command; keep going. */ - - - /* We goto here if a matching operation fails. */ - fail: - if (!FAIL_STACK_EMPTY ()) - { /* A restart point is known. Restore to that state. */ - DEBUG_PRINT1 ("\nFAIL:\n"); - POP_FAILURE_POINT (d, p, - lowest_active_reg, highest_active_reg, - regstart, regend, reg_info); - - /* If this failure point is a dummy, try the next one. */ - if (!p) - goto fail; - - /* If we failed to the end of the pattern, don't examine *p. */ - assert (p <= pend); - if (p < pend) - { - boolean is_a_jump_n = false; - - /* If failed to a backwards jump that's part of a repetition - loop, need to pop this failure point and use the next one. */ - switch ((re_opcode_t) *p) - { - case jump_n: - is_a_jump_n = true; - case maybe_pop_jump: - case pop_failure_jump: - case jump: - p1 = p + 1; - EXTRACT_NUMBER_AND_INCR (mcnt, p1); - p1 += mcnt; - - if ((is_a_jump_n && (re_opcode_t) *p1 == succeed_n) - || (!is_a_jump_n - && (re_opcode_t) *p1 == on_failure_jump)) - goto fail; - break; - default: - /* do nothing */ ; - } - } - - if (d >= string1 && d <= end1) - dend = end_match_1; - } - else - break; /* Matching at this starting point really fails. */ - } /* for (;;) */ - - if (best_regs_set) - goto restore_best_regs; - - FREE_VARIABLES (); - - return -1; /* Failure to match. */ -} /* re_match_2 */ - -/* Subroutine definitions for re_match_2. */ - - -/* We are passed P pointing to a register number after a start_memory. - - Return true if the pattern up to the corresponding stop_memory can - match the empty string, and false otherwise. - - If we find the matching stop_memory, sets P to point to one past its number. - Otherwise, sets P to an undefined byte less than or equal to END. - - We don't handle duplicates properly (yet). */ - -static boolean -group_match_null_string_p (p, end, reg_info) - unsigned char **p, *end; - register_info_type *reg_info; -{ - int mcnt; - /* Point to after the args to the start_memory. */ - unsigned char *p1 = *p + 2; - - while (p1 < end) - { - /* Skip over opcodes that can match nothing, and return true or - false, as appropriate, when we get to one that can't, or to the - matching stop_memory. */ - - switch ((re_opcode_t) *p1) - { - /* Could be either a loop or a series of alternatives. */ - case on_failure_jump: - p1++; - EXTRACT_NUMBER_AND_INCR (mcnt, p1); - - /* If the next operation is not a jump backwards in the - pattern. */ - - if (mcnt >= 0) - { - /* Go through the on_failure_jumps of the alternatives, - seeing if any of the alternatives cannot match nothing. - The last alternative starts with only a jump, - whereas the rest start with on_failure_jump and end - with a jump, e.g., here is the pattern for `a|b|c': - - /on_failure_jump/0/6/exactn/1/a/jump_past_alt/0/6 - /on_failure_jump/0/6/exactn/1/b/jump_past_alt/0/3 - /exactn/1/c - - So, we have to first go through the first (n-1) - alternatives and then deal with the last one separately. */ - - - /* Deal with the first (n-1) alternatives, which start - with an on_failure_jump (see above) that jumps to right - past a jump_past_alt. */ - - while ((re_opcode_t) p1[mcnt-3] == jump_past_alt) - { - /* `mcnt' holds how many bytes long the alternative - is, including the ending `jump_past_alt' and - its number. */ - - if (!alt_match_null_string_p (p1, p1 + mcnt - 3, - reg_info)) - return false; - - /* Move to right after this alternative, including the - jump_past_alt. */ - p1 += mcnt; - - /* Break if it's the beginning of an n-th alternative - that doesn't begin with an on_failure_jump. */ - if ((re_opcode_t) *p1 != on_failure_jump) - break; - - /* Still have to check that it's not an n-th - alternative that starts with an on_failure_jump. */ - p1++; - EXTRACT_NUMBER_AND_INCR (mcnt, p1); - if ((re_opcode_t) p1[mcnt-3] != jump_past_alt) - { - /* Get to the beginning of the n-th alternative. */ - p1 -= 3; - break; - } - } - - /* Deal with the last alternative: go back and get number - of the `jump_past_alt' just before it. `mcnt' contains - the length of the alternative. */ - EXTRACT_NUMBER (mcnt, p1 - 2); - - if (!alt_match_null_string_p (p1, p1 + mcnt, reg_info)) - return false; - - p1 += mcnt; /* Get past the n-th alternative. */ - } /* if mcnt > 0 */ - break; - - - case stop_memory: - assert (p1[1] == **p); - *p = p1 + 2; - return true; - - - default: - if (!common_op_match_null_string_p (&p1, end, reg_info)) - return false; - } - } /* while p1 < end */ - - return false; -} /* group_match_null_string_p */ - - -/* Similar to group_match_null_string_p, but doesn't deal with alternatives: - It expects P to be the first byte of a single alternative and END one - byte past the last. The alternative can contain groups. */ - -static boolean -alt_match_null_string_p (p, end, reg_info) - unsigned char *p, *end; - register_info_type *reg_info; -{ - int mcnt; - unsigned char *p1 = p; - - while (p1 < end) - { - /* Skip over opcodes that can match nothing, and break when we get - to one that can't. */ - - switch ((re_opcode_t) *p1) - { - /* It's a loop. */ - case on_failure_jump: - p1++; - EXTRACT_NUMBER_AND_INCR (mcnt, p1); - p1 += mcnt; - break; - - default: - if (!common_op_match_null_string_p (&p1, end, reg_info)) - return false; - } - } /* while p1 < end */ - - return true; -} /* alt_match_null_string_p */ - - -/* Deals with the ops common to group_match_null_string_p and - alt_match_null_string_p. - - Sets P to one after the op and its arguments, if any. */ - -static boolean -common_op_match_null_string_p (p, end, reg_info) - unsigned char **p, *end; - register_info_type *reg_info; -{ - int mcnt; - boolean ret; - int reg_no; - unsigned char *p1 = *p; - - switch ((re_opcode_t) *p1++) - { - case no_op: - case begline: - case endline: - case begbuf: - case endbuf: - case wordbeg: - case wordend: - case wordbound: - case notwordbound: -#ifdef emacs - case before_dot: - case at_dot: - case after_dot: -#endif - break; - - case start_memory: - reg_no = *p1; - assert (reg_no > 0 && reg_no <= MAX_REGNUM); - ret = group_match_null_string_p (&p1, end, reg_info); - - /* Have to set this here in case we're checking a group which - contains a group and a back reference to it. */ - - if (REG_MATCH_NULL_STRING_P (reg_info[reg_no]) == MATCH_NULL_UNSET_VALUE) - REG_MATCH_NULL_STRING_P (reg_info[reg_no]) = ret; - - if (!ret) - return false; - break; - - /* If this is an optimized succeed_n for zero times, make the jump. */ - case jump: - EXTRACT_NUMBER_AND_INCR (mcnt, p1); - if (mcnt >= 0) - p1 += mcnt; - else - return false; - break; - - case succeed_n: - /* Get to the number of times to succeed. */ - p1 += 2; - EXTRACT_NUMBER_AND_INCR (mcnt, p1); - - if (mcnt == 0) - { - p1 -= 4; - EXTRACT_NUMBER_AND_INCR (mcnt, p1); - p1 += mcnt; - } - else - return false; - break; - - case duplicate: - if (!REG_MATCH_NULL_STRING_P (reg_info[*p1])) - return false; - break; - - case set_number_at: - p1 += 4; - - default: - /* All other opcodes mean we cannot match the empty string. */ - return false; - } - - *p = p1; - return true; -} /* common_op_match_null_string_p */ - - -/* Return zero if TRANSLATE[S1] and TRANSLATE[S2] are identical for LEN - bytes; nonzero otherwise. */ - -static int -bcmp_translate (s1, s2, len, translate) - unsigned char *s1, *s2; - register int len; - char *translate; -{ - register unsigned char *p1 = s1, *p2 = s2; - while (len) - { - if (translate[*p1++] != translate[*p2++]) return 1; - len--; - } - return 0; -} - -/* Entry points for GNU code. */ - -/* re_compile_pattern is the GNU regular expression compiler: it - compiles PATTERN (of length SIZE) and puts the result in BUFP. - Returns 0 if the pattern was valid, otherwise an error string. - - Assumes the `allocated' (and perhaps `buffer') and `translate' fields - are set in BUFP on entry. - - We call regex_compile to do the actual compilation. */ - -const char * -re_compile_pattern (pattern, length, bufp) - const char *pattern; - int length; - struct re_pattern_buffer *bufp; -{ - reg_errcode_t ret; - - /* GNU code is written to assume at least RE_NREGS registers will be set - (and at least one extra will be -1). */ - bufp->regs_allocated = REGS_UNALLOCATED; - - /* And GNU code determines whether or not to get register information - by passing null for the REGS argument to re_match, etc., not by - setting no_sub. */ - bufp->no_sub = 0; - - /* Match anchors at newline. */ - bufp->newline_anchor = 1; - - ret = regex_compile (pattern, length, re_syntax_options, bufp); - - return re_error_msg[(int) ret]; -} - -/* Entry points compatible with 4.2 BSD regex library. We don't define - them if this is an Emacs or POSIX compilation. */ - -#if !defined (emacs) && !defined (_POSIX_SOURCE) - -/* BSD has one and only one pattern buffer. */ -static struct re_pattern_buffer re_comp_buf; - -char * -re_comp (s) - const char *s; -{ - reg_errcode_t ret; - - if (!s) - { - if (!re_comp_buf.buffer) - return "No previous regular expression"; - return 0; - } - - if (!re_comp_buf.buffer) - { - re_comp_buf.buffer = (unsigned char *) malloc (200); - if (re_comp_buf.buffer == NULL) - return "Memory exhausted"; - re_comp_buf.allocated = 200; - - re_comp_buf.fastmap = (char *) malloc (1 << BYTEWIDTH); - if (re_comp_buf.fastmap == NULL) - return "Memory exhausted"; - } - - /* Since `re_exec' always passes NULL for the `regs' argument, we - don't need to initialize the pattern buffer fields which affect it. */ - - /* Match anchors at newlines. */ - re_comp_buf.newline_anchor = 1; - - ret = regex_compile (s, strlen (s), re_syntax_options, &re_comp_buf); - - /* Yes, we're discarding `const' here. */ - return (char *) re_error_msg[(int) ret]; -} - - -int -re_exec (s) - const char *s; -{ - const int len = strlen (s); - return - 0 <= re_search (&re_comp_buf, s, len, 0, len, (struct re_registers *) 0); -} -#endif /* not emacs and not _POSIX_SOURCE */ - -/* POSIX.2 functions. Don't define these for Emacs. */ - -#ifndef emacs - -/* regcomp takes a regular expression as a string and compiles it. - - PREG is a regex_t *. We do not expect any fields to be initialized, - since POSIX says we shouldn't. Thus, we set - - `buffer' to the compiled pattern; - `used' to the length of the compiled pattern; - `syntax' to RE_SYNTAX_POSIX_EXTENDED if the - REG_EXTENDED bit in CFLAGS is set; otherwise, to - RE_SYNTAX_POSIX_BASIC; - `newline_anchor' to REG_NEWLINE being set in CFLAGS; - `fastmap' and `fastmap_accurate' to zero; - `re_nsub' to the number of subexpressions in PATTERN. - - PATTERN is the address of the pattern string. - - CFLAGS is a series of bits which affect compilation. - - If REG_EXTENDED is set, we use POSIX extended syntax; otherwise, we - use POSIX basic syntax. - - If REG_NEWLINE is set, then . and [^...] don't match newline. - Also, regexec will try a match beginning after every newline. - - If REG_ICASE is set, then we considers upper- and lowercase - versions of letters to be equivalent when matching. - - If REG_NOSUB is set, then when PREG is passed to regexec, that - routine will report only success or failure, and nothing about the - registers. - - It returns 0 if it succeeds, nonzero if it doesn't. (See regex.h for - the return codes and their meanings.) */ - -int -regcomp (preg, pattern, cflags) - regex_t *preg; - const char *pattern; - int cflags; -{ - reg_errcode_t ret; - unsigned syntax - = (cflags & REG_EXTENDED) ? - RE_SYNTAX_POSIX_EXTENDED : RE_SYNTAX_POSIX_BASIC; - - /* regex_compile will allocate the space for the compiled pattern. */ - preg->buffer = 0; - preg->allocated = 0; - - /* Don't bother to use a fastmap when searching. This simplifies the - REG_NEWLINE case: if we used a fastmap, we'd have to put all the - characters after newlines into the fastmap. This way, we just try - every character. */ - preg->fastmap = 0; - - if (cflags & REG_ICASE) - { - unsigned i; - - preg->translate = (char *) malloc (CHAR_SET_SIZE); - if (preg->translate == NULL) - return (int) REG_ESPACE; - - /* Map uppercase characters to corresponding lowercase ones. */ - for (i = 0; i < CHAR_SET_SIZE; i++) - preg->translate[i] = ISUPPER (i) ? tolower (i) : i; - } - else - preg->translate = NULL; - - /* If REG_NEWLINE is set, newlines are treated differently. */ - if (cflags & REG_NEWLINE) - { /* REG_NEWLINE implies neither . nor [^...] match newline. */ - syntax &= ~RE_DOT_NEWLINE; - syntax |= RE_HAT_LISTS_NOT_NEWLINE; - /* It also changes the matching behavior. */ - preg->newline_anchor = 1; - } - else - preg->newline_anchor = 0; - - preg->no_sub = !!(cflags & REG_NOSUB); - - /* POSIX says a null character in the pattern terminates it, so we - can use strlen here in compiling the pattern. */ - ret = regex_compile (pattern, strlen (pattern), syntax, preg); - - /* POSIX doesn't distinguish between an unmatched open-group and an - unmatched close-group: both are REG_EPAREN. */ - if (ret == REG_ERPAREN) ret = REG_EPAREN; - - return (int) ret; -} - - -/* regexec searches for a given pattern, specified by PREG, in the - string STRING. - - If NMATCH is zero or REG_NOSUB was set in the cflags argument to - `regcomp', we ignore PMATCH. Otherwise, we assume PMATCH has at - least NMATCH elements, and we set them to the offsets of the - corresponding matched substrings. - - EFLAGS specifies `execution flags' which affect matching: if - REG_NOTBOL is set, then ^ does not match at the beginning of the - string; if REG_NOTEOL is set, then $ does not match at the end. - - We return 0 if we find a match and REG_NOMATCH if not. */ - -int -regexec (preg, string, nmatch, pmatch, eflags) - const regex_t *preg; - const char *string; - size_t nmatch; - regmatch_t pmatch[]; - int eflags; -{ - int ret; - struct re_registers regs; - regex_t private_preg; - int len = strlen (string); - boolean want_reg_info = !preg->no_sub && nmatch > 0; - - private_preg = *preg; - - private_preg.not_bol = !!(eflags & REG_NOTBOL); - private_preg.not_eol = !!(eflags & REG_NOTEOL); - - /* The user has told us exactly how many registers to return - information about, via `nmatch'. We have to pass that on to the - matching routines. */ - private_preg.regs_allocated = REGS_FIXED; - - if (want_reg_info) - { - regs.num_regs = nmatch; - regs.start = TALLOC (nmatch, regoff_t); - regs.end = TALLOC (nmatch, regoff_t); - if (regs.start == NULL || regs.end == NULL) - return (int) REG_NOMATCH; - } - - /* Perform the searching operation. */ - ret = re_search (&private_preg, string, len, - /* start: */ 0, /* range: */ len, - want_reg_info ? ®s : (struct re_registers *) 0); - - /* Copy the register information to the POSIX structure. */ - if (want_reg_info) - { - if (ret >= 0) - { - unsigned r; - - for (r = 0; r < nmatch; r++) - { - pmatch[r].rm_so = regs.start[r]; - pmatch[r].rm_eo = regs.end[r]; - } - } - - /* If we needed the temporary register info, free the space now. */ - free (regs.start); - free (regs.end); - } - - /* We want zero return to mean success, unlike `re_search'. */ - return ret >= 0 ? (int) REG_NOERROR : (int) REG_NOMATCH; -} - - -/* Returns a message corresponding to an error code, ERRCODE, returned - from either regcomp or regexec. We don't use PREG here. */ - -size_t -regerror (errcode, preg, errbuf, errbuf_size) - int errcode; - const regex_t *preg; - char *errbuf; - size_t errbuf_size; -{ - const char *msg; - size_t msg_size; - - if (errcode < 0 - || errcode >= (sizeof (re_error_msg) / sizeof (re_error_msg[0]))) - /* Only error codes returned by the rest of the code should be passed - to this routine. If we are given anything else, or if other regex - code generates an invalid error code, then the program has a bug. - Dump core so we can fix it. */ - abort (); - - msg = re_error_msg[errcode]; - - /* POSIX doesn't require that we do anything in this case, but why - not be nice. */ - if (! msg) - msg = "Success"; - - msg_size = strlen (msg) + 1; /* Includes the null. */ - - if (errbuf_size != 0) - { - if (msg_size > errbuf_size) - { - strncpy (errbuf, msg, errbuf_size - 1); - errbuf[errbuf_size - 1] = 0; - } - else - strcpy (errbuf, msg); - } - - return msg_size; -} - - -/* Free dynamically allocated space used by PREG. */ - -void -regfree (preg) - regex_t *preg; -{ - if (preg->buffer != NULL) - free (preg->buffer); - preg->buffer = NULL; - - preg->allocated = 0; - preg->used = 0; - - if (preg->fastmap != NULL) - free (preg->fastmap); - preg->fastmap = NULL; - preg->fastmap_accurate = 0; - - if (preg->translate != NULL) - free (preg->translate); - preg->translate = NULL; -} - -#endif /* not emacs */ - -/* -Local variables: -make-backup-files: t -version-control: t -trim-versions-without-asking: nil -End: -*/ diff --git a/lib/contrib/regex.h b/lib/contrib/regex.h deleted file mode 100644 index 408dd210..00000000 --- a/lib/contrib/regex.h +++ /dev/null @@ -1,490 +0,0 @@ -/* Definitions for data structures and routines for the regular - expression library, version 0.12. - - Copyright (C) 1985, 1989, 1990, 1991, 1992, 1993 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 2, 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, write to the Free Software - Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. */ - -#ifndef __REGEXP_LIBRARY_H__ -#define __REGEXP_LIBRARY_H__ - -/* POSIX says that must be included (by the caller) before - . */ - -#ifdef VMS -/* VMS doesn't have `size_t' in , even though POSIX says it - should be there. */ -#include -#endif - - -/* The following bits are used to determine the regexp syntax we - recognize. The set/not-set meanings are chosen so that Emacs syntax - remains the value 0. The bits are given in alphabetical order, and - the definitions shifted by one from the previous bit; thus, when we - add or remove a bit, only one other definition need change. */ -typedef unsigned reg_syntax_t; - -/* If this bit is not set, then \ inside a bracket expression is literal. - If set, then such a \ quotes the following character. */ -#define RE_BACKSLASH_ESCAPE_IN_LISTS (1) - -/* If this bit is not set, then + and ? are operators, and \+ and \? are - literals. - If set, then \+ and \? are operators and + and ? are literals. */ -#define RE_BK_PLUS_QM (RE_BACKSLASH_ESCAPE_IN_LISTS << 1) - -/* If this bit is set, then character classes are supported. They are: - [:alpha:], [:upper:], [:lower:], [:digit:], [:alnum:], [:xdigit:], - [:space:], [:print:], [:punct:], [:graph:], and [:cntrl:]. - If not set, then character classes are not supported. */ -#define RE_CHAR_CLASSES (RE_BK_PLUS_QM << 1) - -/* If this bit is set, then ^ and $ are always anchors (outside bracket - expressions, of course). - If this bit is not set, then it depends: - ^ is an anchor if it is at the beginning of a regular - expression or after an open-group or an alternation operator; - $ is an anchor if it is at the end of a regular expression, or - before a close-group or an alternation operator. - - This bit could be (re)combined with RE_CONTEXT_INDEP_OPS, because - POSIX draft 11.2 says that * etc. in leading positions is undefined. - We already implemented a previous draft which made those constructs - invalid, though, so we haven't changed the code back. */ -#define RE_CONTEXT_INDEP_ANCHORS (RE_CHAR_CLASSES << 1) - -/* If this bit is set, then special characters are always special - regardless of where they are in the pattern. - If this bit is not set, then special characters are special only in - some contexts; otherwise they are ordinary. Specifically, - * + ? and intervals are only special when not after the beginning, - open-group, or alternation operator. */ -#define RE_CONTEXT_INDEP_OPS (RE_CONTEXT_INDEP_ANCHORS << 1) - -/* If this bit is set, then *, +, ?, and { cannot be first in an re or - immediately after an alternation or begin-group operator. */ -#define RE_CONTEXT_INVALID_OPS (RE_CONTEXT_INDEP_OPS << 1) - -/* If this bit is set, then . matches newline. - If not set, then it doesn't. */ -#define RE_DOT_NEWLINE (RE_CONTEXT_INVALID_OPS << 1) - -/* If this bit is set, then . doesn't match NUL. - If not set, then it does. */ -#define RE_DOT_NOT_NULL (RE_DOT_NEWLINE << 1) - -/* If this bit is set, nonmatching lists [^...] do not match newline. - If not set, they do. */ -#define RE_HAT_LISTS_NOT_NEWLINE (RE_DOT_NOT_NULL << 1) - -/* If this bit is set, either \{...\} or {...} defines an - interval, depending on RE_NO_BK_BRACES. - If not set, \{, \}, {, and } are literals. */ -#define RE_INTERVALS (RE_HAT_LISTS_NOT_NEWLINE << 1) - -/* If this bit is set, +, ? and | aren't recognized as operators. - If not set, they are. */ -#define RE_LIMITED_OPS (RE_INTERVALS << 1) - -/* If this bit is set, newline is an alternation operator. - If not set, newline is literal. */ -#define RE_NEWLINE_ALT (RE_LIMITED_OPS << 1) - -/* If this bit is set, then `{...}' defines an interval, and \{ and \} - are literals. - If not set, then `\{...\}' defines an interval. */ -#define RE_NO_BK_BRACES (RE_NEWLINE_ALT << 1) - -/* If this bit is set, (...) defines a group, and \( and \) are literals. - If not set, \(...\) defines a group, and ( and ) are literals. */ -#define RE_NO_BK_PARENS (RE_NO_BK_BRACES << 1) - -/* If this bit is set, then \ matches . - If not set, then \ is a back-reference. */ -#define RE_NO_BK_REFS (RE_NO_BK_PARENS << 1) - -/* If this bit is set, then | is an alternation operator, and \| is literal. - If not set, then \| is an alternation operator, and | is literal. */ -#define RE_NO_BK_VBAR (RE_NO_BK_REFS << 1) - -/* If this bit is set, then an ending range point collating higher - than the starting range point, as in [z-a], is invalid. - If not set, then when ending range point collates higher than the - starting range point, the range is ignored. */ -#define RE_NO_EMPTY_RANGES (RE_NO_BK_VBAR << 1) - -/* If this bit is set, then an unmatched ) is ordinary. - If not set, then an unmatched ) is invalid. */ -#define RE_UNMATCHED_RIGHT_PAREN_ORD (RE_NO_EMPTY_RANGES << 1) - -/* This global variable defines the particular regexp syntax to use (for - some interfaces). When a regexp is compiled, the syntax used is - stored in the pattern buffer, so changing this does not affect - already-compiled regexps. */ -extern reg_syntax_t re_syntax_options; - -/* Define combinations of the above bits for the standard possibilities. - (The [[[ comments delimit what gets put into the Texinfo file, so - don't delete them!) */ -/* [[[begin syntaxes]]] */ -#define RE_SYNTAX_EMACS 0 - -#define RE_SYNTAX_AWK \ - (RE_BACKSLASH_ESCAPE_IN_LISTS | RE_DOT_NOT_NULL \ - | RE_NO_BK_PARENS | RE_NO_BK_REFS \ - | RE_NO_BK_VBAR | RE_NO_EMPTY_RANGES \ - | RE_UNMATCHED_RIGHT_PAREN_ORD) - -#define RE_SYNTAX_POSIX_AWK \ - (RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS) - -#define RE_SYNTAX_GREP \ - (RE_BK_PLUS_QM | RE_CHAR_CLASSES \ - | RE_HAT_LISTS_NOT_NEWLINE | RE_INTERVALS \ - | RE_NEWLINE_ALT) - -#define RE_SYNTAX_EGREP \ - (RE_CHAR_CLASSES | RE_CONTEXT_INDEP_ANCHORS \ - | RE_CONTEXT_INDEP_OPS | RE_HAT_LISTS_NOT_NEWLINE \ - | RE_NEWLINE_ALT | RE_NO_BK_PARENS \ - | RE_NO_BK_VBAR) - -#define RE_SYNTAX_POSIX_EGREP \ - (RE_SYNTAX_EGREP | RE_INTERVALS | RE_NO_BK_BRACES) - -/* P1003.2/D11.2, section 4.20.7.1, lines 5078ff. */ -#define RE_SYNTAX_ED RE_SYNTAX_POSIX_BASIC - -#define RE_SYNTAX_SED RE_SYNTAX_POSIX_BASIC - -/* Syntax bits common to both basic and extended POSIX regex syntax. */ -#define _RE_SYNTAX_POSIX_COMMON \ - (RE_CHAR_CLASSES | RE_DOT_NEWLINE | RE_DOT_NOT_NULL \ - | RE_INTERVALS | RE_NO_EMPTY_RANGES) - -#define RE_SYNTAX_POSIX_BASIC \ - (_RE_SYNTAX_POSIX_COMMON | RE_BK_PLUS_QM) - -/* Differs from ..._POSIX_BASIC only in that RE_BK_PLUS_QM becomes - RE_LIMITED_OPS, i.e., \? \+ \| are not recognized. Actually, this - isn't minimal, since other operators, such as \`, aren't disabled. */ -#define RE_SYNTAX_POSIX_MINIMAL_BASIC \ - (_RE_SYNTAX_POSIX_COMMON | RE_LIMITED_OPS) - -#define RE_SYNTAX_POSIX_EXTENDED \ - (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \ - | RE_CONTEXT_INDEP_OPS | RE_NO_BK_BRACES \ - | RE_NO_BK_PARENS | RE_NO_BK_VBAR \ - | RE_UNMATCHED_RIGHT_PAREN_ORD) - -/* Differs from ..._POSIX_EXTENDED in that RE_CONTEXT_INVALID_OPS - replaces RE_CONTEXT_INDEP_OPS and RE_NO_BK_REFS is added. */ -#define RE_SYNTAX_POSIX_MINIMAL_EXTENDED \ - (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \ - | RE_CONTEXT_INVALID_OPS | RE_NO_BK_BRACES \ - | RE_NO_BK_PARENS | RE_NO_BK_REFS \ - | RE_NO_BK_VBAR | RE_UNMATCHED_RIGHT_PAREN_ORD) -/* [[[end syntaxes]]] */ - -/* Maximum number of duplicates an interval can allow. Some systems - (erroneously) define this in other header files, but we want our - value, so remove any previous define. */ -#ifdef RE_DUP_MAX -#undef RE_DUP_MAX -#endif -#define RE_DUP_MAX ((1 << 15) - 1) - - -/* POSIX `cflags' bits (i.e., information for `regcomp'). */ - -/* If this bit is set, then use extended regular expression syntax. - If not set, then use basic regular expression syntax. */ -#define REG_EXTENDED 1 - -/* If this bit is set, then ignore case when matching. - If not set, then case is significant. */ -#define REG_ICASE (REG_EXTENDED << 1) - -/* If this bit is set, then anchors do not match at newline - characters in the string. - If not set, then anchors do match at newlines. */ -#define REG_NEWLINE (REG_ICASE << 1) - -/* If this bit is set, then report only success or fail in regexec. - If not set, then returns differ between not matching and errors. */ -#define REG_NOSUB (REG_NEWLINE << 1) - - -/* POSIX `eflags' bits (i.e., information for regexec). */ - -/* If this bit is set, then the beginning-of-line operator doesn't match - the beginning of the string (presumably because it's not the - beginning of a line). - If not set, then the beginning-of-line operator does match the - beginning of the string. */ -#define REG_NOTBOL 1 - -/* Like REG_NOTBOL, except for the end-of-line. */ -#define REG_NOTEOL (1 << 1) - - -/* If any error codes are removed, changed, or added, update the - `re_error_msg' table in regex.c. */ -typedef enum -{ - REG_NOERROR = 0, /* Success. */ - REG_NOMATCH, /* Didn't find a match (for regexec). */ - - /* POSIX regcomp return error codes. (In the order listed in the - standard.) */ - REG_BADPAT, /* Invalid pattern. */ - REG_ECOLLATE, /* Not implemented. */ - REG_ECTYPE, /* Invalid character class name. */ - REG_EESCAPE, /* Trailing backslash. */ - REG_ESUBREG, /* Invalid back reference. */ - REG_EBRACK, /* Unmatched left bracket. */ - REG_EPAREN, /* Parenthesis imbalance. */ - REG_EBRACE, /* Unmatched \{. */ - REG_BADBR, /* Invalid contents of \{\}. */ - REG_ERANGE, /* Invalid range end. */ - REG_ESPACE, /* Ran out of memory. */ - REG_BADRPT, /* No preceding re for repetition op. */ - - /* Error codes we've added. */ - REG_EEND, /* Premature end. */ - REG_ESIZE, /* Compiled pattern bigger than 2^16 bytes. */ - REG_ERPAREN /* Unmatched ) or \); not returned from regcomp. */ -} reg_errcode_t; - -/* This data structure represents a compiled pattern. Before calling - the pattern compiler, the fields `buffer', `allocated', `fastmap', - `translate', and `no_sub' can be set. After the pattern has been - compiled, the `re_nsub' field is available. All other fields are - private to the regex routines. */ - -struct re_pattern_buffer -{ -/* [[[begin pattern_buffer]]] */ - /* Space that holds the compiled pattern. It is declared as - `unsigned char *' because its elements are - sometimes used as array indexes. */ - unsigned char *buffer; - - /* Number of bytes to which `buffer' points. */ - unsigned long allocated; - - /* Number of bytes actually used in `buffer'. */ - unsigned long used; - - /* Syntax setting with which the pattern was compiled. */ - reg_syntax_t syntax; - - /* Pointer to a fastmap, if any, otherwise zero. re_search uses - the fastmap, if there is one, to skip over impossible - starting points for matches. */ - char *fastmap; - - /* Either a translate table to apply to all characters before - comparing them, or zero for no translation. The translation - is applied to a pattern when it is compiled and to a string - when it is matched. */ - char *translate; - - /* Number of subexpressions found by the compiler. */ - size_t re_nsub; - - /* Zero if this pattern cannot match the empty string, one else. - Well, in truth it's used only in `re_search_2', to see - whether or not we should use the fastmap, so we don't set - this absolutely perfectly; see `re_compile_fastmap' (the - `duplicate' case). */ - unsigned can_be_null : 1; - - /* If REGS_UNALLOCATED, allocate space in the `regs' structure - for `max (RE_NREGS, re_nsub + 1)' groups. - If REGS_REALLOCATE, reallocate space if necessary. - If REGS_FIXED, use what's there. */ -#define REGS_UNALLOCATED 0 -#define REGS_REALLOCATE 1 -#define REGS_FIXED 2 - unsigned regs_allocated : 2; - - /* Set to zero when `regex_compile' compiles a pattern; set to one - by `re_compile_fastmap' if it updates the fastmap. */ - unsigned fastmap_accurate : 1; - - /* If set, `re_match_2' does not return information about - subexpressions. */ - unsigned no_sub : 1; - - /* If set, a beginning-of-line anchor doesn't match at the - beginning of the string. */ - unsigned not_bol : 1; - - /* Similarly for an end-of-line anchor. */ - unsigned not_eol : 1; - - /* If true, an anchor at a newline matches. */ - unsigned newline_anchor : 1; - -/* [[[end pattern_buffer]]] */ -}; - -typedef struct re_pattern_buffer regex_t; - - -/* search.c (search_buffer) in Emacs needs this one opcode value. It is - defined both in `regex.c' and here. */ -#define RE_EXACTN_VALUE 1 - -/* Type for byte offsets within the string. POSIX mandates this. */ -typedef int regoff_t; - - -/* This is the structure we store register match data in. See - regex.texinfo for a full description of what registers match. */ -struct re_registers -{ - unsigned num_regs; - regoff_t *start; - regoff_t *end; -}; - - -/* If `regs_allocated' is REGS_UNALLOCATED in the pattern buffer, - `re_match_2' returns information about at least this many registers - the first time a `regs' structure is passed. */ -#ifndef RE_NREGS -#define RE_NREGS 30 -#endif - - -/* POSIX specification for registers. Aside from the different names than - `re_registers', POSIX uses an array of structures, instead of a - structure of arrays. */ -typedef struct -{ - regoff_t rm_so; /* Byte offset from string's start to substring's start. */ - regoff_t rm_eo; /* Byte offset from string's start to substring's end. */ -} regmatch_t; - -/* Declarations for routines. */ - -/* To avoid duplicating every routine declaration -- once with a - prototype (if we are ANSI), and once without (if we aren't) -- we - use the following macro to declare argument types. This - unfortunately clutters up the declarations a bit, but I think it's - worth it. */ - -#if __STDC__ - -#define _RE_ARGS(args) args - -#else /* not __STDC__ */ - -#define _RE_ARGS(args) () - -#endif /* not __STDC__ */ - -/* Sets the current default syntax to SYNTAX, and return the old syntax. - You can also simply assign to the `re_syntax_options' variable. */ -extern reg_syntax_t re_set_syntax _RE_ARGS ((reg_syntax_t syntax)); - -/* Compile the regular expression PATTERN, with length LENGTH - and syntax given by the global `re_syntax_options', into the buffer - BUFFER. Return NULL if successful, and an error string if not. */ -extern const char *re_compile_pattern - _RE_ARGS ((const char *pattern, int length, - struct re_pattern_buffer *buffer)); - - -/* Compile a fastmap for the compiled pattern in BUFFER; used to - accelerate searches. Return 0 if successful and -2 if was an - internal error. */ -extern int re_compile_fastmap _RE_ARGS ((struct re_pattern_buffer *buffer)); - - -/* Search in the string STRING (with length LENGTH) for the pattern - compiled into BUFFER. Start searching at position START, for RANGE - characters. Return the starting position of the match, -1 for no - match, or -2 for an internal error. Also return register - information in REGS (if REGS and BUFFER->no_sub are nonzero). */ -extern int re_search - _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string, - int length, int start, int range, struct re_registers *regs)); - - -/* Like `re_search', but search in the concatenation of STRING1 and - STRING2. Also, stop searching at index START + STOP. */ -extern int re_search_2 - _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string1, - int length1, const char *string2, int length2, - int start, int range, struct re_registers *regs, int stop)); - - -/* Like `re_search', but return how many characters in STRING the regexp - in BUFFER matched, starting at position START. */ -extern int re_match - _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string, - int length, int start, struct re_registers *regs)); - - -/* Relates to `re_match' as `re_search_2' relates to `re_search'. */ -extern int re_match_2 - _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string1, - int length1, const char *string2, int length2, - int start, struct re_registers *regs, int stop)); - - -/* Set REGS to hold NUM_REGS registers, storing them in STARTS and - ENDS. Subsequent matches using BUFFER and REGS will use this memory - for recording register information. STARTS and ENDS must be - allocated with malloc, and must each be at least `NUM_REGS * sizeof - (regoff_t)' bytes long. - - If NUM_REGS == 0, then subsequent matches should allocate their own - register data. - - Unless this function is called, the first search or match using - PATTERN_BUFFER will allocate its own register data, without - freeing the old data. */ -extern void re_set_registers - _RE_ARGS ((struct re_pattern_buffer *buffer, struct re_registers *regs, - unsigned num_regs, regoff_t *starts, regoff_t *ends)); - -/* 4.2 bsd compatibility. */ -extern char *re_comp _RE_ARGS ((const char *)); -extern int re_exec _RE_ARGS ((const char *)); - -/* POSIX compatibility. */ -extern int regcomp _RE_ARGS ((regex_t *preg, const char *pattern, int cflags)); -extern int regexec - _RE_ARGS ((const regex_t *preg, const char *string, size_t nmatch, - regmatch_t pmatch[], int eflags)); -extern size_t regerror - _RE_ARGS ((int errcode, const regex_t *preg, char *errbuf, - size_t errbuf_size)); -extern void regfree _RE_ARGS ((regex_t *preg)); - -#endif /* not __REGEXP_LIBRARY_H__ */ - -/* -Local variables: -make-backup-files: t -version-control: t -trim-versions-without-asking: nil -End: -*/ diff --git a/lib/contrib/regexpr.c b/lib/contrib/regexpr.c new file mode 100644 index 00000000..ca5e7413 --- /dev/null +++ b/lib/contrib/regexpr.c @@ -0,0 +1,1722 @@ +/* + +regexpr.c + +Author: Tatu Ylonen + +Copyright (c) 1991 Tatu Ylonen, Espoo, Finland + +Permission to use, copy, modify, distribute, and sell this software +and its documentation is hereby granted without fee, provided that the +above copyright notice appears in all source code copies, the name of +Tatu Ylonen is not used to advertise products containing this software +or a derivation thereof, and all modified versions are clearly marked +as such. + +This software is provided "as is" without express or implied warranty. + +Created: Thu Sep 26 17:14:05 1991 ylo +Last modified: Sun Mar 29 16:47:31 1992 ylo + +This code draws many ideas from the regular expression packages by +Henry Spencer of the University of Toronto and Richard Stallman of the +Free Software Foundation. + +Emacs-specific code and syntax table code is almost directly borrowed +from GNU regexp. + +$Id$ + +*/ + +#include +#include +#include +#include +#include +#include +#include +#include + +#include +#include +//#include "silcincludes.h" +#include "regexpr.h" + +#define MACRO_BEGIN do { +#define MACRO_END } while (0) + +enum regexp_compiled_ops /* opcodes for compiled regexp */ +{ + Cend, /* end of pattern reached */ + Cbol, /* beginning of line */ + Ceol, /* end of line */ + Cset, /* character set. Followed by 32 bytes of set. */ + Cexact, /* followed by a byte to match */ + Canychar, /* matches any character except newline */ + Cstart_memory, /* set register start addr (followed by reg number) */ + Cend_memory, /* set register end addr (followed by reg number) */ + Cmatch_memory, /* match a duplicate of reg contents (regnum follows)*/ + Cjump, /* followed by two bytes (lsb,msb) of displacement. */ + Cstar_jump, /* will change to jump/update_failure_jump at runtime */ + Cfailure_jump, /* jump to addr on failure */ + Cupdate_failure_jump, /* update topmost failure point and jump */ + Cdummy_failure_jump, /* push a dummy failure point and jump */ + Cbegbuf, /* match at beginning of buffer */ + Cendbuf, /* match at end of buffer */ + Cwordbeg, /* match at beginning of word */ + Cwordend, /* match at end of word */ + Cwordbound, /* match if at word boundary */ + Cnotwordbound, /* match if not at word boundary */ +#ifdef emacs + Cemacs_at_dot, /* emacs only: matches at dot */ +#endif /* emacs */ + Csyntaxspec, /* matches syntax code (1 byte follows) */ + Cnotsyntaxspec /* matches if syntax code does not match (1 byte foll)*/ +}; + +enum regexp_syntax_op /* syntax codes for plain and quoted characters */ +{ + Rend, /* special code for end of regexp */ + Rnormal, /* normal character */ + Ranychar, /* any character except newline */ + Rquote, /* the quote character */ + Rbol, /* match beginning of line */ + Reol, /* match end of line */ + Roptional, /* match preceding expression optionally */ + Rstar, /* match preceding expr zero or more times */ + Rplus, /* match preceding expr one or more times */ + Ror, /* match either of alternatives */ + Ropenpar, /* opening parenthesis */ + Rclosepar, /* closing parenthesis */ + Rmemory, /* match memory register */ + Rextended_memory, /* \vnn to match registers 10-99 */ + Ropenset, /* open set. Internal syntax hard-coded below. */ + /* the following are gnu extensions to "normal" regexp syntax */ + Rbegbuf, /* beginning of buffer */ + Rendbuf, /* end of buffer */ + Rwordchar, /* word character */ + Rnotwordchar, /* not word character */ + Rwordbeg, /* beginning of word */ + Rwordend, /* end of word */ + Rwordbound, /* word bound */ + Rnotwordbound, /* not word bound */ +#ifdef emacs + Remacs_at_dot, /* emacs: at dot */ + Remacs_syntaxspec, /* syntaxspec */ + Remacs_notsyntaxspec, /* notsyntaxspec */ +#endif /* emacs */ + Rnum_ops +}; + +static int re_compile_initialized = 0; +static int regexp_syntax = 0; +static unsigned char regexp_plain_ops[256]; +static unsigned char regexp_quoted_ops[256]; +static unsigned char regexp_precedences[Rnum_ops]; +static int regexp_context_indep_ops; +static int regexp_ansi_sequences; + +#define NUM_LEVELS 5 /* number of precedence levels in use */ +#define MAX_NESTING 100 /* max nesting level of operators */ + +#ifdef emacs + +/* This code is for emacs compatibility only. */ + +#include "config.h" +#include "lisp.h" +#include "buffer.h" +#include "syntax.h" + +/* emacs defines NULL in some strange way? */ +#undef NULL +#define NULL 0 + +#else /* emacs */ + +#define SYNTAX(ch) re_syntax_table[(unsigned char)(ch)] +#define Sword 1 + +#ifdef SYNTAX_TABLE +char *re_syntax_table; +#else +static char re_syntax_table[256]; +#endif /* SYNTAX_TABLE */ + +#endif /* emacs */ + +static void re_compile_initialize() +{ + int a; + +#if !defined(emacs) && !defined(SYNTAX_TABLE) + static int syntax_table_inited = 0; + + if (!syntax_table_inited) + { + syntax_table_inited = 1; + memset(re_syntax_table, 0, 256); + for (a = 'a'; a <= 'z'; a++) + re_syntax_table[a] = Sword; + for (a = 'A'; a <= 'Z'; a++) + re_syntax_table[a] = Sword; + for (a = '0'; a <= '9'; a++) + re_syntax_table[a] = Sword; + } +#endif /* !emacs && !SYNTAX_TABLE */ + re_compile_initialized = 1; + for (a = 0; a < 256; a++) + { + regexp_plain_ops[a] = Rnormal; + regexp_quoted_ops[a] = Rnormal; + } + for (a = '0'; a <= '9'; a++) + regexp_quoted_ops[a] = Rmemory; + regexp_plain_ops['\134'] = Rquote; + if (regexp_syntax & RE_NO_BK_PARENS) + { + regexp_plain_ops['('] = Ropenpar; + regexp_plain_ops[')'] = Rclosepar; + } + else + { + regexp_quoted_ops['('] = Ropenpar; + regexp_quoted_ops[')'] = Rclosepar; + } + if (regexp_syntax & RE_NO_BK_VBAR) + regexp_plain_ops['\174'] = Ror; + else + regexp_quoted_ops['\174'] = Ror; + regexp_plain_ops['*'] = Rstar; + if (regexp_syntax & RE_BK_PLUS_QM) + { + regexp_quoted_ops['+'] = Rplus; + regexp_quoted_ops['?'] = Roptional; + } + else + { + regexp_plain_ops['+'] = Rplus; + regexp_plain_ops['?'] = Roptional; + } + if (regexp_syntax & RE_NEWLINE_OR) + regexp_plain_ops['\n'] = Ror; + regexp_plain_ops['\133'] = Ropenset; + regexp_plain_ops['\136'] = Rbol; + regexp_plain_ops['$'] = Reol; + regexp_plain_ops['.'] = Ranychar; + if (!(regexp_syntax & RE_NO_GNU_EXTENSIONS)) + { +#ifdef emacs + regexp_quoted_ops['='] = Remacs_at_dot; + regexp_quoted_ops['s'] = Remacs_syntaxspec; + regexp_quoted_ops['S'] = Remacs_notsyntaxspec; +#endif /* emacs */ + regexp_quoted_ops['w'] = Rwordchar; + regexp_quoted_ops['W'] = Rnotwordchar; + regexp_quoted_ops['<'] = Rwordbeg; + regexp_quoted_ops['>'] = Rwordend; + regexp_quoted_ops['b'] = Rwordbound; + regexp_quoted_ops['B'] = Rnotwordbound; + regexp_quoted_ops['`'] = Rbegbuf; + regexp_quoted_ops['\''] = Rendbuf; + } + if (regexp_syntax & RE_ANSI_HEX) + regexp_quoted_ops['v'] = Rextended_memory; + for (a = 0; a < Rnum_ops; a++) + regexp_precedences[a] = 4; + if (regexp_syntax & RE_TIGHT_VBAR) + { + regexp_precedences[Ror] = 3; + regexp_precedences[Rbol] = 2; + regexp_precedences[Reol] = 2; + } + else + { + regexp_precedences[Ror] = 2; + regexp_precedences[Rbol] = 3; + regexp_precedences[Reol] = 3; + } + regexp_precedences[Rclosepar] = 1; + regexp_precedences[Rend] = 0; + regexp_context_indep_ops = (regexp_syntax & RE_CONTEXT_INDEP_OPS) != 0; + regexp_ansi_sequences = (regexp_syntax & RE_ANSI_HEX) != 0; +} + +int re_set_syntax(syntax) +int syntax; +{ + int ret; + + ret = regexp_syntax; + regexp_syntax = syntax; + re_compile_initialize(); + return ret; +} + +static int hex_char_to_decimal(ch) +int ch; +{ + if (ch >= '0' && ch <= '9') + return ch - '0'; + if (ch >= 'a' && ch <= 'f') + return ch - 'a' + 10; + if (ch >= 'A' && ch <= 'F') + return ch - 'A' + 10; + return 16; +} + +char *re_compile_pattern(regex, size, bufp) +char *regex; +int size; +regexp_t bufp; +{ + int a, pos, op, current_level, level, opcode; + int pattern_offset = 0, alloc; + int starts[NUM_LEVELS * MAX_NESTING], starts_base; + int future_jumps[MAX_NESTING], num_jumps; + unsigned char ch = 0; + char *pattern, *translate; + int next_register, paren_depth, num_open_registers, open_registers[RE_NREGS]; + int beginning_context; + +#define NEXTCHAR(var) \ + MACRO_BEGIN \ + if (pos >= size) \ + goto ends_prematurely; \ + (var) = regex[pos]; \ + pos++; \ + MACRO_END + +#define ALLOC(amount) \ + MACRO_BEGIN \ + if (pattern_offset+(amount) > alloc) \ + { \ + alloc += 256 + (amount); \ + pattern = realloc(pattern, alloc); \ + if (!pattern) \ + goto out_of_memory; \ + } \ + MACRO_END + +#define STORE(ch) pattern[pattern_offset++] = (ch) + +#define CURRENT_LEVEL_START (starts[starts_base + current_level]) + +#define SET_LEVEL_START starts[starts_base + current_level] = pattern_offset + +#define PUSH_LEVEL_STARTS if (starts_base < (MAX_NESTING-1)*NUM_LEVELS) \ + starts_base += NUM_LEVELS; \ + else \ + goto too_complex + +#define POP_LEVEL_STARTS starts_base -= NUM_LEVELS + +#define PUT_ADDR(offset,addr) \ + MACRO_BEGIN \ + int disp = (addr) - (offset) - 2; \ + pattern[(offset)] = disp & 0xff; \ + pattern[(offset)+1] = (disp>>8) & 0xff; \ + MACRO_END + +#define INSERT_JUMP(pos,type,addr) \ + MACRO_BEGIN \ + int a, p = (pos), t = (type), ad = (addr); \ + for (a = pattern_offset - 1; a >= p; a--) \ + pattern[a + 3] = pattern[a]; \ + pattern[p] = t; \ + PUT_ADDR(p+1,ad); \ + pattern_offset += 3; \ + MACRO_END + +#define SETBIT(buf,offset,bit) (buf)[(offset)+(bit)/8] |= (1<<((bit) & 7)) + +#define SET_FIELDS \ + MACRO_BEGIN \ + bufp->allocated = alloc; \ + bufp->buffer = pattern; \ + bufp->used = pattern_offset; \ + MACRO_END + +#define GETHEX(var) \ + MACRO_BEGIN \ + char gethex_ch, gethex_value; \ + NEXTCHAR(gethex_ch); \ + gethex_value = hex_char_to_decimal(gethex_ch); \ + if (gethex_value == 16) \ + goto hex_error; \ + NEXTCHAR(gethex_ch); \ + gethex_ch = hex_char_to_decimal(gethex_ch); \ + if (gethex_ch == 16) \ + goto hex_error; \ + (var) = gethex_value * 16 + gethex_ch; \ + MACRO_END + +#define ANSI_TRANSLATE(ch) \ + MACRO_BEGIN \ + switch (ch) \ + { \ + case 'a': \ + case 'A': \ + ch = 7; /* audible bell */ \ + break; \ + case 'b': \ + case 'B': \ + ch = 8; /* backspace */ \ + break; \ + case 'f': \ + case 'F': \ + ch = 12; /* form feed */ \ + break; \ + case 'n': \ + case 'N': \ + ch = 10; /* line feed */ \ + break; \ + case 'r': \ + case 'R': \ + ch = 13; /* carriage return */ \ + break; \ + case 't': \ + case 'T': \ + ch = 9; /* tab */ \ + break; \ + case 'v': \ + case 'V': \ + ch = 11; /* vertical tab */ \ + break; \ + case 'x': /* hex code */ \ + case 'X': \ + GETHEX(ch); \ + break; \ + default: \ + /* other characters passed through */ \ + if (translate) \ + ch = translate[(unsigned char)ch]; \ + break; \ + } \ + MACRO_END + + if (!re_compile_initialized) + re_compile_initialize(); + bufp->used = 0; + bufp->fastmap_accurate = 0; + bufp->uses_registers = 0; + translate = bufp->translate; + pattern = bufp->buffer; + alloc = bufp->allocated; + if (alloc == 0 || pattern == NULL) + { + alloc = 256; + pattern = malloc(alloc); + if (!pattern) + goto out_of_memory; + } + pattern_offset = 0; + starts_base = 0; + num_jumps = 0; + current_level = 0; + SET_LEVEL_START; + num_open_registers = 0; + next_register = 1; + paren_depth = 0; + beginning_context = 1; + op = -1; + /* we use Rend dummy to ensure that pending jumps are updated (due to + low priority of Rend) before exiting the loop. */ + pos = 0; + while (op != Rend) + { + if (pos >= size) + op = Rend; + else + { + NEXTCHAR(ch); + if (translate) + ch = translate[(unsigned char)ch]; + op = regexp_plain_ops[(unsigned char)ch]; + if (op == Rquote) + { + NEXTCHAR(ch); + op = regexp_quoted_ops[(unsigned char)ch]; + if (op == Rnormal && regexp_ansi_sequences) + ANSI_TRANSLATE(ch); + } + } + level = regexp_precedences[op]; + /* printf("ch='%c' op=%d level=%d current_level=%d curlevstart=%d\n", + ch, op, level, current_level, CURRENT_LEVEL_START); */ + if (level > current_level) + { + for (current_level++; current_level < level; current_level++) + SET_LEVEL_START; + SET_LEVEL_START; + } + else + if (level < current_level) + { + current_level = level; + for (;num_jumps > 0 && + future_jumps[num_jumps-1] >= CURRENT_LEVEL_START; + num_jumps--) + PUT_ADDR(future_jumps[num_jumps-1], pattern_offset); + } + switch (op) + { + case Rend: + break; + case Rnormal: + normal_char: + opcode = Cexact; + store_opcode_and_arg: /* opcode & ch must be set */ + SET_LEVEL_START; + ALLOC(2); + STORE(opcode); + STORE(ch); + break; + case Ranychar: + opcode = Canychar; + store_opcode: + SET_LEVEL_START; + ALLOC(1); + STORE(opcode); + break; + case Rquote: + abort(); + /*NOTREACHED*/ + case Rbol: + if (!beginning_context) { + if (regexp_context_indep_ops) + goto op_error; + else + goto normal_char; + } + opcode = Cbol; + goto store_opcode; + case Reol: + if (!((pos >= size) || + ((regexp_syntax & RE_NO_BK_VBAR) ? + (regex[pos] == '\174') : + (pos+1 < size && regex[pos] == '\134' && + regex[pos+1] == '\174')) || + ((regexp_syntax & RE_NO_BK_PARENS)? + (regex[pos] == ')'): + (pos+1 < size && regex[pos] == '\134' && + regex[pos+1] == ')')))) { + if (regexp_context_indep_ops) + goto op_error; + else + goto normal_char; + } + opcode = Ceol; + goto store_opcode; + break; + case Roptional: + if (beginning_context) { + if (regexp_context_indep_ops) + goto op_error; + else + goto normal_char; + } + if (CURRENT_LEVEL_START == pattern_offset) + break; /* ignore empty patterns for ? */ + ALLOC(3); + INSERT_JUMP(CURRENT_LEVEL_START, Cfailure_jump, + pattern_offset + 3); + break; + case Rstar: + case Rplus: + if (beginning_context) { + if (regexp_context_indep_ops) + goto op_error; + else + goto normal_char; + } + if (CURRENT_LEVEL_START == pattern_offset) + break; /* ignore empty patterns for + and * */ + ALLOC(9); + INSERT_JUMP(CURRENT_LEVEL_START, Cfailure_jump, + pattern_offset + 6); + INSERT_JUMP(pattern_offset, Cstar_jump, CURRENT_LEVEL_START); + if (op == Rplus) /* jump over initial failure_jump */ + INSERT_JUMP(CURRENT_LEVEL_START, Cdummy_failure_jump, + CURRENT_LEVEL_START + 6); + break; + case Ror: + ALLOC(6); + INSERT_JUMP(CURRENT_LEVEL_START, Cfailure_jump, + pattern_offset + 6); + if (num_jumps >= MAX_NESTING) + goto too_complex; + STORE(Cjump); + future_jumps[num_jumps++] = pattern_offset; + STORE(0); + STORE(0); + SET_LEVEL_START; + break; + case Ropenpar: + SET_LEVEL_START; + if (next_register < RE_NREGS) + { + bufp->uses_registers = 1; + ALLOC(2); + STORE(Cstart_memory); + STORE(next_register); + open_registers[num_open_registers++] = next_register; + next_register++; + } + paren_depth++; + PUSH_LEVEL_STARTS; + current_level = 0; + SET_LEVEL_START; + break; + case Rclosepar: + if (paren_depth <= 0) + goto parenthesis_error; + POP_LEVEL_STARTS; + current_level = regexp_precedences[Ropenpar]; + paren_depth--; + if (paren_depth < num_open_registers) + { + bufp->uses_registers = 1; + ALLOC(2); + STORE(Cend_memory); + num_open_registers--; + STORE(open_registers[num_open_registers]); + } + break; + case Rmemory: + if (ch == '0') + goto bad_match_register; + assert(ch >= '0' && ch <= '9'); + bufp->uses_registers = 1; + opcode = Cmatch_memory; + ch -= '0'; + goto store_opcode_and_arg; + case Rextended_memory: + NEXTCHAR(ch); + if (ch < '0' || ch > '9') + goto bad_match_register; + NEXTCHAR(a); + if (a < '0' || a > '9') + goto bad_match_register; + ch = 10 * (a - '0') + ch - '0'; + if (ch <= 0 || ch >= RE_NREGS) + goto bad_match_register; + bufp->uses_registers = 1; + opcode = Cmatch_memory; + goto store_opcode_and_arg; + case Ropenset: + { + int complement,prev,offset,range,firstchar; + + SET_LEVEL_START; + ALLOC(1+256/8); + STORE(Cset); + offset = pattern_offset; + for (a = 0; a < 256/8; a++) + STORE(0); + NEXTCHAR(ch); + if (translate) + ch = translate[(unsigned char)ch]; + if (ch == '\136') + { + complement = 1; + NEXTCHAR(ch); + if (translate) + ch = translate[(unsigned char)ch]; + } + else + complement = 0; + prev = -1; + range = 0; + firstchar = 1; + while (ch != '\135' || firstchar) + { + firstchar = 0; + if (regexp_ansi_sequences && ch == '\134') + { + NEXTCHAR(ch); + ANSI_TRANSLATE(ch); + } + if (range) + { + for (a = prev; a <= ch; a++) + SETBIT(pattern, offset, a); + prev = -1; + range = 0; + } + else + if (prev != -1 && ch == '-') + range = 1; + else + { + SETBIT(pattern, offset, ch); + prev = ch; + } + NEXTCHAR(ch); + if (translate) + ch = translate[(unsigned char)ch]; + } + if (range) + SETBIT(pattern, offset, '-'); + if (complement) + { + for (a = 0; a < 256/8; a++) + pattern[offset+a] ^= 0xff; + } + break; + } + case Rbegbuf: + opcode = Cbegbuf; + goto store_opcode; + case Rendbuf: + opcode = Cendbuf; + goto store_opcode; + case Rwordchar: + opcode = Csyntaxspec; + ch = Sword; + goto store_opcode_and_arg; + case Rnotwordchar: + opcode = Cnotsyntaxspec; + ch = Sword; + goto store_opcode_and_arg; + case Rwordbeg: + opcode = Cwordbeg; + goto store_opcode; + case Rwordend: + opcode = Cwordend; + goto store_opcode; + case Rwordbound: + opcode = Cwordbound; + goto store_opcode; + case Rnotwordbound: + opcode = Cnotwordbound; + goto store_opcode; +#ifdef emacs + case Remacs_at_dot: + opcode = Cemacs_at_dot; + goto store_opcode; + case Remacs_syntaxspec: + NEXTCHAR(ch); + if (translate) + ch = translate[(unsigned char)ch]; + opcode = Csyntaxspec; + ch = syntax_spec_code[(unsigned char)ch]; + goto store_opcode_and_arg; + case Remacs_notsyntaxspec: + NEXTCHAR(ch); + if (translate) + ch = translate[(unsigned char)ch]; + opcode = Cnotsyntaxspec; + ch = syntax_spec_code[(unsigned char)ch]; + goto store_opcode_and_arg; +#endif /* emacs */ + default: + abort(); + } + beginning_context = (op == Ropenpar || op == Ror); + } + if (starts_base != 0) + goto parenthesis_error; + assert(num_jumps == 0); + ALLOC(1); + STORE(Cend); + SET_FIELDS; + return NULL; + + op_error: + SET_FIELDS; + return "Badly placed special character"; + + bad_match_register: + SET_FIELDS; + return "Bad match register number"; + + hex_error: + SET_FIELDS; + return "Bad hexadecimal number"; + + parenthesis_error: + SET_FIELDS; + return "Badly placed parenthesis"; + + out_of_memory: + SET_FIELDS; + return "Out of memory"; + + ends_prematurely: + SET_FIELDS; + return "Regular expression ends prematurely"; + + too_complex: + SET_FIELDS; + return "Regular expression too complex"; +} +#undef CHARAT +#undef NEXTCHAR +#undef GETHEX +#undef ALLOC +#undef STORE +#undef CURRENT_LEVEL_START +#undef SET_LEVEL_START +#undef PUSH_LEVEL_STARTS +#undef POP_LEVEL_STARTS +#undef PUT_ADDR +#undef INSERT_JUMP +#undef SETBIT +#undef SET_FIELDS + +static void re_compile_fastmap_aux(code, pos, visited, can_be_null, fastmap) +char *code, *visited, *can_be_null, *fastmap; +int pos; +{ + int a, b, syntaxcode; + + if (visited[pos]) + return; /* we have already been here */ + visited[pos] = 1; + for (;;) + switch (code[pos++]) + { + case Cend: + *can_be_null = 1; + return; + case Cbol: + case Cbegbuf: + case Cendbuf: + case Cwordbeg: + case Cwordend: + case Cwordbound: + case Cnotwordbound: +#ifdef emacs + case Cemacs_at_dot: +#endif /* emacs */ + break; + case Csyntaxspec: + syntaxcode = code[pos++]; + for (a = 0; a < 256; a++) + if (SYNTAX(a) == syntaxcode) + fastmap[a] = 1; + return; + case Cnotsyntaxspec: + syntaxcode = code[pos++]; + for (a = 0; a < 256; a++) + if (SYNTAX(a) != syntaxcode) + fastmap[a] = 1; + return; + case Ceol: + fastmap['\n'] = 1; + if (*can_be_null == 0) + *can_be_null = 2; /* can match null, but only at end of buffer*/ + return; + case Cset: + for (a = 0; a < 256/8; a++) + if (code[pos + a] != 0) + for (b = 0; b < 8; b++) + if (code[pos + a] & (1 << b)) + fastmap[(a << 3) + b] = 1; + pos += 256/8; + return; + case Cexact: + fastmap[(unsigned char)code[pos]] = 1; + return; + case Canychar: + for (a = 0; a < 256; a++) + if (a != '\n') + fastmap[a] = 1; + return; + case Cstart_memory: + case Cend_memory: + pos++; + break; + case Cmatch_memory: + for (a = 0; a < 256; a++) + fastmap[a] = 1; + *can_be_null = 1; + return; + case Cjump: + case Cdummy_failure_jump: + case Cupdate_failure_jump: + case Cstar_jump: + a = (unsigned char)code[pos++]; + a |= (unsigned char)code[pos++] << 8; + pos += (int)(short)a; + if (visited[pos]) + { + /* argh... the regexp contains empty loops. This is not + good, as this may cause a failure stack overflow when + matching. Oh well. */ + /* this path leads nowhere; pursue other paths. */ + return; + } + visited[pos] = 1; + break; + case Cfailure_jump: + a = (unsigned char)code[pos++]; + a |= (unsigned char)code[pos++] << 8; + a = pos + (int)(short)a; + re_compile_fastmap_aux(code, a, visited, can_be_null, fastmap); + break; + default: + abort(); /* probably some opcode is missing from this switch */ + /*NOTREACHED*/ + } +} + +static int re_do_compile_fastmap(buffer, used, pos, can_be_null, fastmap) +char *buffer, *fastmap, *can_be_null; +int used, pos; +{ + char small_visited[512], *visited; + + if (used <= sizeof(small_visited)) + visited = small_visited; + else + { + visited = malloc(used); + if (!visited) + return 0; + } + *can_be_null = 0; + memset(fastmap, 0, 256); + memset(visited, 0, used); + re_compile_fastmap_aux(buffer, pos, visited, can_be_null, fastmap); + if (visited != small_visited) + free(visited); + return 1; +} + +void re_compile_fastmap(bufp) +regexp_t bufp; +{ + if (!bufp->fastmap || bufp->fastmap_accurate) + return; + assert(bufp->used > 0); + if (!re_do_compile_fastmap(bufp->buffer, bufp->used, 0, &bufp->can_be_null, + bufp->fastmap)) + return; + if (bufp->buffer[0] == Cbol) + bufp->anchor = 1; /* begline */ + else + if (bufp->buffer[0] == Cbegbuf) + bufp->anchor = 2; /* begbuf */ + else + bufp->anchor = 0; /* none */ + bufp->fastmap_accurate = 1; +} + +#define INITIAL_FAILURES 128 /* initial # failure points to allocate */ +#define MAX_FAILURES 4100 /* max # of failure points before failing */ + +int re_match_2(bufp, string1, size1, string2, size2, pos, regs, mstop) +regexp_t bufp; +char *string1, *string2; +int size1, size2, pos, mstop; +regexp_registers_t regs; +{ + struct failure_point { char *text, *partend, *code; } + *failure_stack_start, *failure_sp, *failure_stack_end, + initial_failure_stack[INITIAL_FAILURES]; + char *code, *translate, *text, *textend, *partend, *part_2_end; + char *regstart_text[RE_NREGS], *regstart_partend[RE_NREGS]; + char *regend_text[RE_NREGS], *regend_partend[RE_NREGS]; + int a, b, ch, reg, regch, match_end; + char *regtext, *regpartend, *regtextend; + +#define PREFETCH \ + MACRO_BEGIN \ + if (text == partend) \ + { \ + if (text == textend) \ + goto fail; \ + text = string2; \ + partend = part_2_end; \ + } \ + MACRO_END + +#define NEXTCHAR(var) \ + MACRO_BEGIN \ + PREFETCH; \ + (var) = (unsigned char)*text++; \ + if (translate) \ + (var) = (unsigned char)translate[(var)]; \ + MACRO_END + + assert(pos >= 0 && size1 >= 0 && size2 >= 0 && mstop >= 0); + assert(mstop <= size1 + size2); + assert(pos <= mstop); + + if (pos <= size1) + { + text = string1 + pos; + if (mstop <= size1) + { + partend = string1 + mstop; + textend = partend; + } + else + { + partend = string1 + size1; + textend = string2 + mstop - size1; + } + part_2_end = string2 + mstop - size1; + } + else + { + text = string2 + pos - size1; + partend = string2 + mstop - size1; + textend = partend; + part_2_end = partend; + } + + if (bufp->uses_registers && regs != NULL) + for (a = 0; a < RE_NREGS; a++) + regend_text[a] = NULL; + + code = bufp->buffer; + translate = bufp->translate; + failure_stack_start = failure_sp = initial_failure_stack; + failure_stack_end = initial_failure_stack + INITIAL_FAILURES; + +#if 0 + /* re_search_2 has already done this, and otherwise we get little benefit + from this. So I'll leave this out. */ + if (bufp->fastmap_accurate && !bufp->can_be_null && + text != textend && + !bufp->fastmap[translate ? + (unsigned char)translate[(unsigned char)*text] : + (unsigned char)*text]) + return -1; /* it can't possibly match */ +#endif + + continue_matching: + for (;;) + { + switch (*code++) + { + case Cend: + if (partend != part_2_end) + match_end = text - string1; + else + match_end = text - string2 + size1; + if (regs) + { + regs->start[0] = pos; + regs->end[0] = match_end; + if (!bufp->uses_registers) + { + for (a = 1; a < RE_NREGS; a++) + { + regs->start[a] = -1; + regs->end[a] = -1; + } + } + else + { + for (a = 1; a < RE_NREGS; a++) + { + if (regend_text[a] == NULL) + { + regs->start[a] = -1; + regs->end[a] = -1; + continue; + } + if (regstart_partend[a] != part_2_end) + regs->start[a] = regstart_text[a] - string1; + else + regs->start[a] = regstart_text[a] - string2 + size1; + if (regend_partend[a] != part_2_end) + regs->end[a] = regend_text[a] - string1; + else + regs->end[a] = regend_text[a] - string2 + size1; + } + } + } + if (failure_stack_start != initial_failure_stack) + free((char *)failure_stack_start); + return match_end - pos; + case Cbol: + if (text == string1 || text[-1] == '\n') /* text[-1] always valid */ + break; + goto fail; + case Ceol: + if (text == string2 + size2 || + (text == string1 + size1 ? + (size2 == 0 || *string2 == '\n') : + *text == '\n')) + break; + goto fail; + case Cset: + NEXTCHAR(ch); + if (code[ch/8] & (1<<(ch & 7))) + { + code += 256/8; + break; + } + goto fail; + case Cexact: + NEXTCHAR(ch); + if (ch != (unsigned char)*code++) + goto fail; + break; + case Canychar: + NEXTCHAR(ch); + if (ch == '\n') + goto fail; + break; + case Cstart_memory: + reg = *code++; + regstart_text[reg] = text; + regstart_partend[reg] = partend; + break; + case Cend_memory: + reg = *code++; + regend_text[reg] = text; + regend_partend[reg] = partend; + break; + case Cmatch_memory: + reg = *code++; + if (regend_text[reg] == NULL) + goto fail; /* or should we just match nothing? */ + regtext = regstart_text[reg]; + regtextend = regend_text[reg]; + if (regstart_partend[reg] == regend_partend[reg]) + regpartend = regtextend; + else + regpartend = string1 + size1; + + for (;regtext != regtextend;) + { + NEXTCHAR(ch); + if (regtext == regpartend) + regtext = string2; + regch = (unsigned char)*regtext++; + if (translate) + regch = (unsigned char)translate[regch]; + if (regch != ch) + goto fail; + } + break; + case Cstar_jump: + /* star is coded as: + 1: failure_jump 2 + ... code for operand of star + star_jump 1 + 2: ... code after star + We change the star_jump to update_failure_jump if we can determine + that it is safe to do so; otherwise we change it to an ordinary + jump. + plus is coded as + jump 2 + 1: failure_jump 3 + 2: ... code for operand of plus + star_jump 1 + 3: ... code after plus + For star_jump considerations this is processed identically + to star. */ + a = (unsigned char)*code++; + a |= (unsigned char)*code++ << 8; + a = (int)(short)a; + { + char map[256], can_be_null; + char *p1, *p2; + + p1 = code + a + 3; /* skip the failure_jump */ + assert(p1[-3] == Cfailure_jump); + p2 = code; + /* p1 points inside loop, p2 points to after loop */ + if (!re_do_compile_fastmap(bufp->buffer, bufp->used, + p2 - bufp->buffer, &can_be_null, map)) + goto make_normal_jump; + /* If we might introduce a new update point inside the loop, + we can't optimize because then update_jump would update a + wrong failure point. Thus we have to be quite careful here. */ + loop_p1: + /* loop until we find something that consumes a character */ + switch (*p1++) + { + case Cbol: + case Ceol: + case Cbegbuf: + case Cendbuf: + case Cwordbeg: + case Cwordend: + case Cwordbound: + case Cnotwordbound: +#ifdef emacs + case Cemacs_at_dot: +#endif /* emacs */ + goto loop_p1; + case Cstart_memory: + case Cend_memory: + p1++; + goto loop_p1; + case Cexact: + ch = (unsigned char)*p1++; + if (map[ch]) + goto make_normal_jump; + break; + case Canychar: + for (b = 0; b < 256; b++) + if (b != '\n' && map[b]) + goto make_normal_jump; + break; + case Cset: + for (b = 0; b < 256; b++) + if ((p1[b >> 3] & (1 << (b & 7))) && map[b]) + goto make_normal_jump; + p1 += 256/8; + break; + default: + goto make_normal_jump; + } + /* now we know that we can't backtrack. */ + while (p1 != p2 - 3) + { + switch (*p1++) + { + case Cend: + abort(); /* we certainly shouldn't get this inside loop */ + /*NOTREACHED*/ + case Cbol: + case Ceol: + case Canychar: + case Cbegbuf: + case Cendbuf: + case Cwordbeg: + case Cwordend: + case Cwordbound: + case Cnotwordbound: +#ifdef emacs + case Cemacs_at_dot: +#endif /* emacs */ + break; + case Cset: + p1 += 256/8; + break; + case Cexact: + case Cstart_memory: + case Cend_memory: + case Cmatch_memory: + case Csyntaxspec: + case Cnotsyntaxspec: + p1++; + break; + case Cjump: + case Cstar_jump: + case Cfailure_jump: + case Cupdate_failure_jump: + case Cdummy_failure_jump: + goto make_normal_jump; + default: + printf("regexpr.c: processing star_jump: unknown op %d\n", p1[-1]); + break; + } + } + goto make_update_jump; + } + make_normal_jump: + /* printf("changing to normal jump\n"); */ + code -= 3; + *code = Cjump; + break; + make_update_jump: + /* printf("changing to update jump\n"); */ + code -= 2; + a += 3; /* jump to after the Cfailure_jump */ + code[-1] = Cupdate_failure_jump; + code[0] = a & 0xff; + code[1] = a >> 8; + /* fall to next case */ + case Cupdate_failure_jump: + failure_sp[-1].text = text; + failure_sp[-1].partend = partend; + /* fall to next case */ + case Cjump: + a = (unsigned char)*code++; + a |= (unsigned char)*code++ << 8; + code += (int)(short)a; + break; + case Cdummy_failure_jump: + case Cfailure_jump: + if (failure_sp == failure_stack_end) + { + if (failure_stack_start != initial_failure_stack) + goto error; + failure_stack_start = (struct failure_point *) + malloc(MAX_FAILURES * sizeof(*failure_stack_start)); + failure_stack_end = failure_stack_start + MAX_FAILURES; + memcpy((char *)failure_stack_start, (char *)initial_failure_stack, + INITIAL_FAILURES * sizeof(*failure_stack_start)); + failure_sp = failure_stack_start + INITIAL_FAILURES; + } + a = (unsigned char)*code++; + a |= (unsigned char)*code++ << 8; + a = (int)(short)a; + if (code[-3] == Cdummy_failure_jump) + { /* this is only used in plus */ + assert(*code == Cfailure_jump); + b = (unsigned char)code[1]; + b |= (unsigned char)code[2] << 8; + failure_sp->code = code + (int)(short)b + 3; + failure_sp->text = NULL; + code += a; + } + else + { + failure_sp->code = code + a; + failure_sp->text = text; + failure_sp->partend = partend; + } + failure_sp++; + break; + case Cbegbuf: + if (text == string1) + break; + goto fail; + case Cendbuf: + if (size2 == 0 ? text == string1 + size1 : text == string2 + size2) + break; + goto fail; + case Cwordbeg: + if (text == string2 + size2) + goto fail; + if (size2 == 0 && text == string1 + size1) + goto fail; + if (SYNTAX(text == string1 + size1 ? *string1 : *text) != Sword) + goto fail; + if (text == string1) + break; + if (SYNTAX(text[-1]) != Sword) + break; + goto fail; + case Cwordend: + if (text == string1) + goto fail; + if (SYNTAX(text[-1]) != Sword) + goto fail; + if (text == string2 + size2) + break; + if (size2 == 0 && text == string1 + size1) + break; + if (SYNTAX(*text) == Sword) + goto fail; + break; + case Cwordbound: + /* Note: as in gnu regexp, this also matches at the beginning + and end of buffer. */ + if (text == string1 || text == string2 + size2 || + (size2 == 0 && text == string1 + size1)) + break; + if ((SYNTAX(text[-1]) == Sword) ^ + (SYNTAX(text == string1 + size1 ? *string2 : *text) == Sword)) + break; + goto fail; + case Cnotwordbound: + /* Note: as in gnu regexp, this never matches at the beginning + and end of buffer. */ + if (text == string1 || text == string2 + size2 || + (size2 == 0 && text == string1 + size1)) + goto fail; + if (!((SYNTAX(text[-1]) == Sword) ^ + (SYNTAX(text == string1 + size1 ? *string2 : *text) == Sword))) + goto fail; + break; + case Csyntaxspec: + NEXTCHAR(ch); + if (SYNTAX(ch) != (unsigned char)*code++) + goto fail; + break; + case Cnotsyntaxspec: + NEXTCHAR(ch); + if (SYNTAX(ch) != (unsigned char)*code++) + break; + goto fail; +#ifdef emacs + case Cemacs_at_dot: + if (PTR_CHAR_POS((unsigned char *)text) + 1 != point) + goto fail; + break; +#endif /* emacs */ + default: + abort(); + /*NOTREACHED*/ + } + } + abort(); + /*NOTREACHED*/ + + fail: + if (failure_sp != failure_stack_start) + { + failure_sp--; + text = failure_sp->text; + if (text == NULL) + goto fail; + partend = failure_sp->partend; + code = failure_sp->code; + goto continue_matching; + } + if (failure_stack_start != initial_failure_stack) + free((char *)failure_stack_start); + return -1; + + error: + if (failure_stack_start != initial_failure_stack) + free((char *)failure_stack_start); + return -2; +} + +#undef PREFETCH +#undef NEXTCHAR +#undef PUSH_FAILURE + +int re_match(bufp, string, size, pos, regs) +regexp_t bufp; +char *string; +int size, pos; +regexp_registers_t regs; +{ + return re_match_2(bufp, string, size, (char *)NULL, 0, pos, regs, size); +} + +int re_search_2(bufp, string1, size1, string2, size2, pos, range, regs, + mstop) +regexp_t bufp; +char *string1, *string2; +int size1, size2, pos, range, mstop; +regexp_registers_t regs; +{ + char *fastmap, *translate, *text, *partstart, *partend; + int dir, ret; + char anchor; + + assert(size1 >= 0 && size2 >= 0 && pos >= 0 && mstop >= 0); + assert(pos + range >= 0 && pos + range <= size1 + size2); + assert(pos <= mstop); + + fastmap = bufp->fastmap; + translate = bufp->translate; + if (fastmap && !bufp->fastmap_accurate) + re_compile_fastmap(bufp); + anchor = bufp->anchor; + if (bufp->can_be_null == 1) /* can_be_null == 2: can match null at eob */ + fastmap = NULL; + if (range < 0) + { + dir = -1; + range = -range; + } + else + dir = 1; + if (anchor == 2) { + if (pos != 0) + return -1; + else + range = 0; + } + for (; range >= 0; range--, pos += dir) + { + if (fastmap) + { + if (dir == 1) + { /* searching forwards */ + if (pos < size1) + { + text = string1 + pos; + if (pos + range > size1) + partend = string1 + size1; + else + partend = string1 + pos + range; + } + else + { + text = string2 + pos - size1; + partend = string2 + pos + range - size1; + } + partstart = text; + if (translate) + while (text != partend && + !fastmap[(unsigned char) + translate[(unsigned char)*text]]) + text++; + else + while (text != partend && !fastmap[(unsigned char)*text]) + text++; + pos += text - partstart; + range -= text - partstart; + if (pos == size1 + size2 && bufp->can_be_null == 0) + return -1; + } + else + { /* searching backwards */ + if (pos <= size1) + { + text = string1 + pos; + partstart = string1 + pos - range; + } + else + { + text = string2 + pos - size1; + if (range < pos - size1) + partstart = string2 + pos - size1 - range; + else + partstart = string2; + } + partend = text; + if (translate) + while (text != partstart && + !fastmap[(unsigned char) + translate[(unsigned char)*text]]) + text--; + else + while (text != partstart && + !fastmap[(unsigned char)*text]) + text--; + pos -= partend - text; + range -= partend - text; + } + } + if (anchor == 1) + { /* anchored to begline */ + if (pos > 0 && + (pos <= size1 ? string1[pos - 1] : + string2[pos - size1 - 1]) != '\n') + continue; + } + assert(pos >= 0 && pos <= size1 + size2); + ret = re_match_2(bufp, string1, size1, string2, size2, pos, regs, mstop); + if (ret >= 0) + return pos; + if (ret == -2) + return -2; + } + return -1; +} + +int re_search(bufp, string, size, startpos, range, regs) +regexp_t bufp; +char *string; +int size, startpos, range; +regexp_registers_t regs; +{ + return re_search_2(bufp, string, size, (char *)NULL, 0, + startpos, range, regs, size); +} + +static struct re_pattern_buffer re_comp_buf; + +char *re_comp(s) +char *s; +{ + if (s == NULL) + { + if (!re_comp_buf.buffer) + return "Out of memory"; + return NULL; + } + if (!re_comp_buf.buffer) + { + /* the buffer will be allocated automatically */ + re_comp_buf.fastmap = malloc(256); + re_comp_buf.translate = NULL; + } + return re_compile_pattern(s, strlen(s), &re_comp_buf); +} + +int re_exec(s) +char *s; +{ + int len = strlen(s); + + return re_search(&re_comp_buf, s, len, 0, len, (regexp_registers_t)NULL) >= 0; +} + +/* POSIX Compatibility */ + +int regcomp(regex_t *preg, const char *regex, int cflags) +{ + int syntax = 0; + memset(preg, 0, sizeof(*preg)); + if (cflags & REG_EXTENDED) + syntax |= (RE_CONTEXT_INDEP_OPS | RE_NO_BK_PARENS | RE_NO_BK_VBAR); + re_set_syntax(syntax); + if (re_compile_pattern((char *)regex, strlen(regex), preg) == NULL) + return 0; + return -1; +} + +int regexec(const regex_t *preg, const char *string, size_t nmatch, + regmatch_t pmatch[], int eflags) +{ + int len = strlen(string); + int ret; + + ret = re_search((regex_t *)preg, (char *)string, len, 0, len, (regexp_registers_t)NULL); + if (ret >= 0) + return 0; + + return ret; +} + +size_t regerror(int errcode, const regex_t *preg, char *errbuf, + size_t errbuf_size) +{ + return -1; +} + +void regfree(regex_t *preg) +{ + free(preg->buffer); +} + +#ifdef TEST_REGEXP + +int main() +{ + char buf[500]; + char *cp; + struct re_pattern_buffer exp; + struct re_registers regs; + int a,pos; + char fastmap[256]; + + exp.allocated = 0; + exp.buffer = 0; + exp.translate = NULL; + exp.fastmap = fastmap; + + /* re_set_syntax(RE_NO_BK_PARENS|RE_NO_BK_VBAR|RE_ANSI_HEX); */ + + while (1) + { + printf("Enter regexp:\n"); + gets(buf); + cp=re_compile_pattern(buf, strlen(buf), &exp); + if (cp) + { + printf("Error: %s\n", cp); + continue; + } + re_compile_fastmap(&exp); + printf("dump:\n"); + for (pos = 0; pos < exp.used;) + { + printf("%d: ", pos); + switch (exp.buffer[pos++]) + { + case Cend: + strcpy(buf, "end"); + break; + case Cbol: + strcpy(buf, "bol"); + break; + case Ceol: + strcpy(buf, "eol"); + break; + case Cset: + strcpy(buf, "set "); + for (a = 0; a < 256/8; a++) + sprintf(buf+strlen(buf)," %02x", + (unsigned char)exp.buffer[pos++]); + break; + case Cexact: + sprintf(buf, "exact '%c' 0x%x", exp.buffer[pos], + (unsigned char)exp.buffer[pos]); + pos++; + break; + case Canychar: + strcpy(buf, "anychar"); + break; + case Cstart_memory: + sprintf(buf, "start_memory %d", exp.buffer[pos++]); + break; + case Cend_memory: + sprintf(buf, "end_memory %d", exp.buffer[pos++]); + break; + case Cmatch_memory: + sprintf(buf, "match_memory %d", exp.buffer[pos++]); + break; + case Cjump: + case Cdummy_failure_jump: + case Cstar_jump: + case Cfailure_jump: + case Cupdate_failure_jump: + a = (unsigned char)exp.buffer[pos++]; + a += (unsigned char)exp.buffer[pos++] << 8; + a = (int)(short)a; + switch (exp.buffer[pos-3]) + { + case Cjump: + cp = "jump"; + break; + case Cstar_jump: + cp = "star_jump"; + break; + case Cfailure_jump: + cp = "failure_jump"; + break; + case Cupdate_failure_jump: + cp = "update_failure_jump"; + break; + case Cdummy_failure_jump: + cp = "dummy_failure_jump"; + break; + default: + cp = "unknown jump"; + break; + } + sprintf(buf, "%s %d", cp, a + pos); + break; + case Cbegbuf: + strcpy(buf,"begbuf"); + break; + case Cendbuf: + strcpy(buf,"endbuf"); + break; + case Cwordbeg: + strcpy(buf,"wordbeg"); + break; + case Cwordend: + strcpy(buf,"wordend"); + break; + case Cwordbound: + strcpy(buf,"wordbound"); + break; + case Cnotwordbound: + strcpy(buf,"notwordbound"); + break; + default: + sprintf(buf, "unknown code %d", + (unsigned char)exp.buffer[pos - 1]); + break; + } + printf("%s\n", buf); + } + printf("can_be_null = %d uses_registers = %d anchor = %d\n", + exp.can_be_null, exp.uses_registers, exp.anchor); + + printf("fastmap:"); + for (a = 0; a < 256; a++) + if (exp.fastmap[a]) + printf(" %d", a); + printf("\n"); + printf("Enter strings. An empty line terminates.\n"); + while (fgets(buf, sizeof(buf), stdin)) + { + if (buf[0] == '\n') + break; + a = re_search(&exp, buf, strlen(buf), 0, strlen(buf), ®s); + printf("search returns %d\n", a); + if (a != -1) + { + for (a = 0; a < RE_NREGS; a++) + { + printf("buf %d: %d to %d\n", a, regs.start[a], regs.end[a]); + } + } + } + } +} + +#endif /* TEST_REGEXP */ diff --git a/lib/contrib/regexpr.h b/lib/contrib/regexpr.h new file mode 100644 index 00000000..b6927357 --- /dev/null +++ b/lib/contrib/regexpr.h @@ -0,0 +1,139 @@ +/* + +regexpr.h + +Author: Tatu Ylonen + +Copyright (c) 1991 Tatu Ylonen, Espoo, Finland + +Permission to use, copy, modify, distribute, and sell this software +and its documentation is hereby granted without fee, provided that the +above copyright notice appears in all source code copies, the name of +Tatu Ylonen is not used to advertise products containing this software +or a derivation thereof, and all modified versions are clearly marked +as such. + +This software is provided "as is" without express or implied warranty. + +Created: Thu Sep 26 17:15:36 1991 ylo +Last modified: Fri Jan 3 12:05:45 1992 ylo + +*/ + +/* $Id$ */ + +#ifndef REGEXPR_H +#define REGEXPR_H + +#define RE_NREGS 10 /* number of registers available */ + +typedef struct re_pattern_buffer +{ + char *buffer; /* compiled pattern */ + int allocated; /* allocated size of compiled pattern */ + int used; /* actual length of compiled pattern */ + char *fastmap; /* fastmap[ch] is true if ch can start pattern */ + char *translate; /* translation to apply during compilation/matching */ + char fastmap_accurate; /* true if fastmap is valid */ + char can_be_null; /* true if can match empty string */ + char uses_registers; /* registers are used and need to be initialized */ + char anchor; /* anchor: 0=none 1=begline 2=begbuf */ +} *regexp_t; + +typedef struct re_registers +{ + int start[RE_NREGS]; /* start offset of region */ + int end[RE_NREGS]; /* end offset of region */ +} *regexp_registers_t; + +/* bit definitions for syntax */ +#define RE_NO_BK_PARENS 1 /* no quoting for parentheses */ +#define RE_NO_BK_VBAR 2 /* no quoting for vertical bar */ +#define RE_BK_PLUS_QM 4 /* quoting needed for + and ? */ +#define RE_TIGHT_VBAR 8 /* | binds tighter than ^ and $ */ +#define RE_NEWLINE_OR 16 /* treat newline as or */ +#define RE_CONTEXT_INDEP_OPS 32 /* ^$?*+ are special in all contexts */ +#define RE_ANSI_HEX 64 /* ansi sequences (\n etc) and \xhh */ +#define RE_NO_GNU_EXTENSIONS 128 /* no gnu extensions */ + +/* definitions for some common regexp styles */ +#define RE_SYNTAX_AWK (RE_NO_BK_PARENS|RE_NO_BK_VBAR|RE_CONTEXT_INDEP_OPS) +#define RE_SYNTAX_EGREP (RE_SYNTAX_AWK|RE_NEWLINE_OR) +#define RE_SYNTAX_GREP (RE_BK_PLUS_QM|RE_NEWLINE_OR) +#define RE_SYNTAX_EMACS 0 + +int re_set_syntax(int syntax); +/* This sets the syntax to use and returns the previous syntax. The + syntax is specified by a bit mask of the above defined bits. */ + +char *re_compile_pattern(char *regex, int regex_size, regexp_t compiled); +/* This compiles the regexp (given in regex and length in regex_size). + This returns NULL if the regexp compiled successfully, and an error + message if an error was encountered. The buffer field must be + initialized to a memory area allocated by malloc (or to NULL) before + use, and the allocated field must be set to its length (or 0 if buffer is + NULL). Also, the translate field must be set to point to a valid + translation table, or NULL if it is not used. */ + +int re_match(regexp_t compiled, char *string, int size, int pos, + regexp_registers_t regs); +/* This tries to match the regexp against the string. This returns the + length of the matched portion, or -1 if the pattern could not be + matched and -2 if an error (such as failure stack overflow) is + encountered. */ + +int re_match_2(regexp_t compiled, char *string1, int size1, + char *string2, int size2, int pos, regexp_registers_t regs, + int mstop); +/* This tries to match the regexp to the concatenation of string1 and + string2. This returns the length of the matched portion, or -1 if the + pattern could not be matched and -2 if an error (such as failure stack + overflow) is encountered. */ + +int re_search(regexp_t compiled, char *string, int size, int startpos, + int range, regexp_registers_t regs); +/* This rearches for a substring matching the regexp. This returns the first + index at which a match is found. range specifies at how many positions to + try matching; positive values indicate searching forwards, and negative + values indicate searching backwards. mstop specifies the offset beyond + which a match must not go. This returns -1 if no match is found, and + -2 if an error (such as failure stack overflow) is encountered. */ + +int re_search_2(regexp_t compiled, char *string1, int size1, + char *string2, int size2, int startpos, int range, + regexp_registers_t regs, int mstop); +/* This is like re_search, but search from the concatenation of string1 and + string2. */ + +void re_compile_fastmap(regexp_t compiled); +/* This computes the fastmap for the regexp. For this to have any effect, + the calling program must have initialized the fastmap field to point + to an array of 256 characters. */ + +char *re_comp(char *s); +/* BSD 4.2 regex library routine re_comp. This compiles the regexp into + an internal buffer. This returns NULL if the regexp was compiled + successfully, and an error message if there was an error. */ + +int re_exec(char *s); +/* BSD 4.2 regexp library routine re_exec. This returns true if the string + matches the regular expression (that is, a matching part is found + anywhere in the string). */ + +/* POSIX Compatibility */ +#define regex_t struct re_pattern_buffer +#define regmatch_t struct re_registers +#define REG_EXTENDED 1 +#define REG_ICASE (REG_EXTENDED << 1) +#define REG_NEWLINE (REG_ICASE << 1) +#define REG_NOSUB (REG_NEWLINE << 1) +#define REG_NOTBOL 1 +#define REG_NOTEOL (REG_NOTBOL << 1) +int regcomp(regex_t *preg, const char *regex, int cflags); +int regexec(const regex_t *preg, const char *string, size_t nmatch, + regmatch_t pmatch[], int eflags); +size_t regerror(int errcode, const regex_t *preg, char *errbuf, + size_t errbuf_size); +void regfree(regex_t *preg); + +#endif /* REGEXPR_H */