Date: Thu, 30 Jun 2011 14:10:49 +0000 (UTC) From: Gabor Kovesdan <gabor@FreeBSD.org> To: src-committers@freebsd.org, svn-src-user@freebsd.org Subject: svn commit: r223693 - in user/gabor/tre-integration: contrib/tre/lib include Message-ID: <201106301410.p5UEAnPJ046950@svn.freebsd.org>
next in thread | raw e-mail | index | archive | help
Author: gabor Date: Thu Jun 30 14:10:49 2011 New Revision: 223693 URL: http://svn.freebsd.org/changeset/base/223693 Log: - Plug in the fixed string matching code Modified: user/gabor/tre-integration/contrib/tre/lib/fastmatch.c user/gabor/tre-integration/contrib/tre/lib/fastmatch.h user/gabor/tre-integration/contrib/tre/lib/regexec.c user/gabor/tre-integration/contrib/tre/lib/tre-compile.c user/gabor/tre-integration/contrib/tre/lib/tre.h user/gabor/tre-integration/include/regex.h Modified: user/gabor/tre-integration/contrib/tre/lib/fastmatch.c ============================================================================== --- user/gabor/tre-integration/contrib/tre/lib/fastmatch.c Thu Jun 30 10:56:02 2011 (r223692) +++ user/gabor/tre-integration/contrib/tre/lib/fastmatch.c Thu Jun 30 14:10:49 2011 (r223693) @@ -235,7 +235,7 @@ tre_fastcomp(fastmatch_t *fg, const tre_ int tre_fastexec(const fastmatch_t *fg, const tre_char_t *data, size_t len, - int nmatch, regmatch_t *pmatch) + int nmatch, regmatch_t pmatch[]) { unsigned int j; int cnt = 0; @@ -253,8 +253,10 @@ tre_fastexec(const fastmatch_t *fg, cons j = fg->eol ? len - fg->len : 0; if (fastcmp(fg->pattern, data + j, fg->len) == -1) { - pmatch->rm_so = j; - pmatch->rm_eo = j + fg->len; + if (!(fg->cflags & REG_NOSUB) || (nmatch < 1)) + return 0; + pmatch[cnt].rm_so = j; + pmatch[cnt].rm_eo = j + fg->len; ret = 0; } } @@ -264,7 +266,7 @@ tre_fastexec(const fastmatch_t *fg, cons do { if (fastcmp(fg->pattern, data + j - fg->len, fg->len) == -1) { - if (!(fg->cflags & REG_NOSUB)) + if (!(fg->cflags & REG_NOSUB) || (nmatch < 1)) return (0); pmatch[cnt++].rm_so = j - fg->len; pmatch[cnt++].rm_eo = j; @@ -296,7 +298,7 @@ tre_fastexec(const fastmatch_t *fg, cons j = 0; do { if (fastcmp(fg->pattern, data + j, fg->len) == -1) { - if (!(fg->cflags & REG_NOSUB)) + if (!(fg->cflags & REG_NOSUB) || (nmatch < 1)) return (0); pmatch[cnt++].rm_so = j; pmatch[cnt++].rm_eo = j + fg->len; Modified: user/gabor/tre-integration/contrib/tre/lib/fastmatch.h ============================================================================== --- user/gabor/tre-integration/contrib/tre/lib/fastmatch.h Thu Jun 30 10:56:02 2011 (r223692) +++ user/gabor/tre-integration/contrib/tre/lib/fastmatch.h Thu Jun 30 14:10:49 2011 (r223693) @@ -28,6 +28,8 @@ #ifndef FASTMATCH_H #define FASTMATCH_H 1 +#include <stdbool.h> + #include "hashtable.h" #include "tre.h" @@ -53,7 +55,7 @@ int tre_fastcomp_literal(fastmatch_t *pr int tre_fastcomp(fastmatch_t *preg, const tre_char_t *regex, size_t, int cflags); int tre_fastexec(const fastmatch_t *fg, const tre_char_t *data, - size_t len, int nmatch, regmatch_t *pmatch); + size_t len, int nmatch, regmatch_t pmatch[]); void tre_fastfree(fastmatch_t *preg); #endif /* FASTMATCH_H */ Modified: user/gabor/tre-integration/contrib/tre/lib/regexec.c ============================================================================== --- user/gabor/tre-integration/contrib/tre/lib/regexec.c Thu Jun 30 10:56:02 2011 (r223692) +++ user/gabor/tre-integration/contrib/tre/lib/regexec.c Thu Jun 30 14:10:49 2011 (r223693) @@ -44,6 +44,7 @@ char *alloca (); #endif /* HAVE_MALLOC_H */ #include <limits.h> +#include "fastmatch.h" #include "tre-internal.h" #include "tre.h" #include "xmalloc.h" @@ -150,10 +151,16 @@ tre_have_approx(const regex_t *preg) static int tre_match(const tre_tnfa_t *tnfa, const void *string, size_t len, tre_str_type_t type, size_t nmatch, regmatch_t pmatch[], - int eflags) + int eflags, void *shortcut) { reg_errcode_t status; int *tags = NULL, eo; + + /* Check if we can cheat with a fixed string */ + if (shortcut != NULL) + return tre_fastexec((fastmatch_t *)shortcut, (const tre_char_t *)string, + len, nmatch, pmatch); + if (tnfa->num_tags > 0 && nmatch > 0) { #ifdef TRE_USE_ALLOCA @@ -222,7 +229,8 @@ tre_regnexec(const regex_t *preg, const size_t slen = (size_t)(pmatch[0].rm_eo - pmatch[0].rm_so); size_t offset = pmatch[0].rm_so; str = &str[offset]; - int ret = tre_match(tnfa, str, slen, type, nmatch, pmatch, eflags); + int ret = tre_match(tnfa, str, slen, type, nmatch, pmatch, eflags, + preg->shortcut); if (!(eflags & REG_NOSUB)) { for (unsigned i = 0; i < nmatch; i++) @@ -235,7 +243,8 @@ tre_regnexec(const regex_t *preg, const } else { - return tre_match(tnfa, str, len, type, nmatch, pmatch, eflags); + return tre_match(tnfa, str, len, type, nmatch, pmatch, eflags, + preg->shortcut); } } @@ -260,7 +269,8 @@ tre_regwnexec(const regex_t *preg, const size_t slen = (size_t)(pmatch[0].rm_eo - pmatch[0].rm_so); size_t offset = pmatch[0].rm_so; str = &str[offset]; - int ret = tre_match(tnfa, str, slen, STR_WIDE, nmatch, pmatch, eflags); + int ret = tre_match(tnfa, str, slen, STR_WIDE, nmatch, pmatch, eflags, + preg->shortcut); if (!(eflags & REG_NOSUB)) { for (unsigned i = 0; i < nmatch; i++) @@ -273,7 +283,8 @@ tre_regwnexec(const regex_t *preg, const } else { - return tre_match(tnfa, str, len, STR_WIDE, nmatch, pmatch, eflags); + return tre_match(tnfa, str, len, STR_WIDE, nmatch, pmatch, eflags, + preg->shortcut); } } @@ -291,7 +302,8 @@ tre_reguexec(const regex_t *preg, const size_t nmatch, regmatch_t pmatch[], int eflags) { tre_tnfa_t *tnfa = (void *)preg->TRE_REGEX_T_FIELD; - return tre_match(tnfa, str, (unsigned)-1, STR_USER, nmatch, pmatch, eflags); + return tre_match(tnfa, str, (unsigned)-1, STR_USER, nmatch, pmatch, eflags, + preg->shortcut); } @@ -315,7 +327,7 @@ tre_match_approx(const tre_tnfa_t *tnfa, if (params.max_cost == 0 && !tnfa->have_approx && !(eflags & REG_APPROX_MATCHER)) return tre_match(tnfa, string, len, type, match->nmatch, match->pmatch, - eflags); + eflags, NULL); /* Back references are not supported by the approximate matcher. */ if (tnfa->have_backrefs) Modified: user/gabor/tre-integration/contrib/tre/lib/tre-compile.c ============================================================================== --- user/gabor/tre-integration/contrib/tre/lib/tre-compile.c Thu Jun 30 10:56:02 2011 (r223692) +++ user/gabor/tre-integration/contrib/tre/lib/tre-compile.c Thu Jun 30 14:10:49 2011 (r223693) @@ -20,6 +20,7 @@ #include <assert.h> #include <string.h> +#include "fastmatch.h" #include "tre-internal.h" #include "tre-mem.h" #include "tre-stack.h" @@ -1858,17 +1859,30 @@ tre_compile(regex_t *preg, const tre_cha tre_ast_node_t *tree, *tmp_ast_l, *tmp_ast_r; tre_pos_and_tags_t *p; int *counts = NULL, *offs = NULL; - int i, add = 0; + int i, add = 0, ret; tre_tnfa_transition_t *transitions, *initial; tre_tnfa_t *tnfa = NULL; tre_submatch_data_t *submatch_data; tre_tag_direction_t *tag_directions = NULL; reg_errcode_t errcode; tre_mem_t mem; + fastmatch_t shortcut; /* Parse context. */ tre_parse_ctx_t parse_ctx; + /* Check if we can cheat with a fixed string algorithm. */ + ret = (cflags & REG_LITERAL) + ? tre_fastcomp_literal(&shortcut, regex, n, cflags) + : tre_fastcomp(&shortcut, regex, n, cflags); + if (!ret) + { + preg->shortcut = &shortcut; + return REG_OK; + } + else + preg->shortcut = NULL; + /* Allocate a stack used throughout the compilation process for various purposes. */ stack = tre_stack_new(512, 10240, 128); Modified: user/gabor/tre-integration/contrib/tre/lib/tre.h ============================================================================== --- user/gabor/tre-integration/contrib/tre/lib/tre.h Thu Jun 30 10:56:02 2011 (r223692) +++ user/gabor/tre-integration/contrib/tre/lib/tre.h Thu Jun 30 14:10:49 2011 (r223693) @@ -48,6 +48,7 @@ typedef int regoff_t; typedef struct { size_t re_nsub; /* Number of parenthesized subexpressions. */ void *value; /* For internal use only. */ + void *shortcut; /* For internal use only. */ const char *re_endp; } regex_t; Modified: user/gabor/tre-integration/include/regex.h ============================================================================== --- user/gabor/tre-integration/include/regex.h Thu Jun 30 10:56:02 2011 (r223692) +++ user/gabor/tre-integration/include/regex.h Thu Jun 30 14:10:49 2011 (r223693) @@ -40,6 +40,7 @@ typedef int regoff_t; typedef struct { size_t re_nsub; /* Number of parenthesized subexpressions. */ void *value; /* For internal use only. */ + void *shortcut; /* For internal use only. */ const char *re_endp; } regex_t;
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?201106301410.p5UEAnPJ046950>