X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;f=recctrl%2Fregxread.c;h=7c9919ff0367fe5348fc26ba4bd2995a78560bb1;hb=a8b86c35a39ce9eb5279c46dd99ef7e91b83fb7b;hp=76f8e59d785022c1f6567698ce568ecf32407760;hpb=4d6d0d754cc1f0b1f7f814a2268e679d26b890f7;p=idzebra-moved-to-github.git diff --git a/recctrl/regxread.c b/recctrl/regxread.c index 76f8e59..7c9919f 100644 --- a/recctrl/regxread.c +++ b/recctrl/regxread.c @@ -1,188 +1,42 @@ -/* - * Copyright (C) 1994-1999, Index Data - * All rights reserved. - * Sebastian Hammer, Adam Dickmeiss - * - * $Log: regxread.c,v $ - * Revision 1.27 1999-06-28 13:25:40 quinn - * Improved diagnostics for Tcl - * - * Revision 1.26 1999/05/26 07:49:14 adam - * C++ compilation. - * - * Revision 1.25 1999/05/25 12:33:32 adam - * Fixed bug in Tcl filter. - * - * Revision 1.24 1999/05/21 11:08:46 adam - * Tcl filter attempts to read .tflt. Improvements to configure - * script so that it reads uninstalled Tcl source. - * - * Revision 1.23 1999/05/20 12:57:18 adam - * Implemented TCL filter. Updated recctrl system. - * - * Revision 1.22 1998/11/03 16:07:13 adam - * Yet another fix. - * - * Revision 1.21 1998/11/03 15:43:39 adam - * Fixed bug introduced by previous commit. - * - * Revision 1.20 1998/11/03 14:51:28 adam - * Changed code so that it creates as few data1 nodes as possible. - * - * Revision 1.19 1998/11/03 10:22:39 adam - * Fixed memory leak that could occur for when large data1 node were - * concatenated. Data-type data1_nodes may have multiple nodes. - * - * Revision 1.18 1998/10/15 13:11:47 adam - * Added support for option -record for "end element". When specified - * end element will mark end-of-record when at outer-level. - * - * Revision 1.17 1998/07/01 10:13:51 adam - * Minor fix. - * - * Revision 1.16 1998/06/30 15:15:09 adam - * Tags are trimmed: white space removed before- and after the tag. - * - * Revision 1.15 1998/06/30 12:55:45 adam - * Bug fix. - * - * Revision 1.14 1998/03/05 08:41:00 adam - * Implemented rule contexts. - * - * Revision 1.13 1997/12/12 06:33:58 adam - * Fixed bug that showed up when multiple filter where used. - * Made one routine thread-safe. - * - * Revision 1.12 1997/11/18 10:03:24 adam - * Member num_children removed from data1_node. - * - * Revision 1.11 1997/11/06 11:41:01 adam - * Implemented "begin variant" for the sgml.regx filter. - * - * Revision 1.10 1997/10/31 12:36:12 adam - * Minor change that avoids compiler warning. - * - * Revision 1.9 1997/09/29 09:02:49 adam - * Fixed small bug (introduced by previous commit). - * - * Revision 1.8 1997/09/17 12:19:22 adam - * Zebra version corresponds to YAZ version 1.4. - * Changed Zebra server so that it doesn't depend on global common_resource. - * - * Revision 1.7 1997/07/15 16:33:07 adam - * Check for zero length in execData. - * - * Revision 1.6 1997/02/24 10:41:51 adam - * Cleanup of code and commented out the "end element-end-record" code. - * - * Revision 1.5 1997/02/19 16:22:33 adam - * Fixed "end element" to terminate record in outer-most level. - * - * Revision 1.4 1997/02/12 20:42:58 adam - * Changed some log messages. - * - * Revision 1.3 1996/11/08 14:05:33 adam - * Bug fix: data1 node member u.tag.get_bytes weren't initialized. - * - * Revision 1.2 1996/10/29 14:02:09 adam - * Doesn't use the global data1_tabpath (from YAZ). Instead the function - * data1_get_tabpath is used. - * - * Revision 1.1 1996/10/11 10:57:30 adam - * New module recctrl. Used to manage records (extract/retrieval). - * - * Revision 1.24 1996/06/17 14:25:31 adam - * Removed LOG_DEBUG logs; can still be enabled by setting REGX_DEBUG. - * - * Revision 1.23 1996/06/04 10:19:00 adam - * Minor changes - removed include of ctype.h. - * - * Revision 1.22 1996/06/03 15:23:13 adam - * Bug fix: /../ BODY /../ - pattern didn't match EOF. - * - * Revision 1.21 1996/05/14 16:58:38 adam - * Minor change. - * - * Revision 1.20 1996/05/01 13:46:36 adam - * First work on multiple records in one file. - * New option, -offset, to the "unread" command in the filter module. - * - * Revision 1.19 1996/02/12 16:18:20 adam - * Yet another bug fix in implementation of unread command. - * - * Revision 1.18 1996/02/12 16:07:54 adam - * Bug fix in new unread command. - * - * Revision 1.17 1996/02/12 15:56:11 adam - * New code command: unread. - * - * Revision 1.16 1996/01/17 14:57:51 adam - * Prototype changed for reader functions in extract/retrieve. File - * is identified by 'void *' instead of 'int. - * - * Revision 1.15 1996/01/08 19:15:47 adam - * New input filter that works! - * - * Revision 1.14 1996/01/08 09:10:38 adam - * Yet another complete rework on this module. - * - * Revision 1.13 1995/12/15 17:21:50 adam - * This version is able to set data.formatted_text in data1-nodes. - * - * Revision 1.12 1995/12/15 16:20:10 adam - * The filter files (*.flt) are read from the path given by data1_tabpath. - * - * Revision 1.11 1995/12/15 12:35:16 adam - * Better logging. - * - * Revision 1.10 1995/12/15 10:35:36 adam - * Misc. bug fixes. - * - * Revision 1.9 1995/12/14 16:38:48 adam - * Completely new attempt to make regular expression parsing. - * - * Revision 1.8 1995/12/13 17:16:59 adam - * Small changes. - * - * Revision 1.7 1995/12/13 16:51:58 adam - * Modified to set last_child in data1_nodes. - * Uses destroy handler to free up data text nodes. - * - * Revision 1.6 1995/12/13 13:45:37 quinn - * Changed data1 to use nmem. - * - * Revision 1.5 1995/12/11 09:12:52 adam - * The rec_get function returns NULL if record doesn't exist - will - * happen in the server if the result set records have been deleted since - * the creation of the set (i.e. the search). - * The server saves a result temporarily if it is 'volatile', i.e. the - * set is register dependent. - * - * Revision 1.4 1995/12/05 16:57:40 adam - * More work on regular patterns. - * - * Revision 1.3 1995/12/05 09:37:09 adam - * One malloc was renamed to xmalloc. - * - * Revision 1.2 1995/12/04 17:59:24 adam - * More work on regular expression conversion. - * - * Revision 1.1 1995/12/04 14:25:30 adam - * Started work on regular expression parsed input to structured records. - * - */ +/* $Id: regxread.c,v 1.51 2004-08-04 08:35:25 adam Exp $ + Copyright (C) 1995,1996,1997,1998,1999,2000,2001,2002,2003,2004 + Index Data Aps + +This file is part of the Zebra server. + +Zebra 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. + +Zebra 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 Zebra; see the file LICENSE.zebra. If not, write to the +Free Software Foundation, 59 Temple Place - Suite 330, Boston, MA +02111-1307, USA. +*/ + + #include #include #include #include -#include +#include #include #include #include "grsread.h" #if HAVE_TCL_H #include + +#if MAJOR_VERSION >= 8 +#define HAVE_TCL_OBJECTS +#endif #endif #define REGX_DEBUG 0 @@ -201,6 +55,9 @@ struct regxCode { char *str; +#if HAVE_TCL_OBJECTS + Tcl_Obj *tcl_obj; +#endif }; struct lexRuleAction { @@ -240,7 +97,6 @@ struct lexContext { }; struct lexConcatBuf { - int len; int max; char *buf; }; @@ -269,7 +125,7 @@ struct lexSpec { int (*f_win_rf)(void *, char *, size_t); off_t (*f_win_sf)(void *, off_t); - struct lexConcatBuf **concatBuf; + struct lexConcatBuf *concatBuf; int maxLevel; data1_node **d1_stack; int d1_level; @@ -347,6 +203,10 @@ static void regxCodeDel (struct regxCode **pp) struct regxCode *p = *pp; if (p) { +#if HAVE_TCL_OBJECTS + if (p->tcl_obj) + Tcl_DecrRefCount (p->tcl_obj); +#endif xfree (p->str); xfree (p); *pp = NULL; @@ -361,6 +221,11 @@ static void regxCodeMk (struct regxCode **pp, const char *buf, int len) p->str = (char *) xmalloc (len+1); memcpy (p->str, buf, len); p->str[len] = '\0'; +#if HAVE_TCL_OBJECTS + p->tcl_obj = Tcl_NewStringObj ((char *) buf, len); + if (p->tcl_obj) + Tcl_IncrRefCount (p->tcl_obj); +#endif *pp = p; } @@ -417,6 +282,7 @@ static void lexContextDestroy (struct lexContext *p) { struct lexRule *rp, *rp1; + dfa_delete (&p->dfa); xfree (p->fastRule); for (rp = p->rules; rp; rp = rp1) { @@ -426,6 +292,7 @@ static void lexContextDestroy (struct lexContext *p) } actionListDel (&p->beginActionList); actionListDel (&p->endActionList); + actionListDel (&p->initActionList); xfree (p->name); xfree (p); } @@ -450,14 +317,12 @@ static struct lexSpec *lexSpecCreate (const char *name, data1_handle dh) p->f_win_buf = NULL; p->maxLevel = 128; - p->concatBuf = (struct lexConcatBuf **) + p->concatBuf = (struct lexConcatBuf *) xmalloc (sizeof(*p->concatBuf) * p->maxLevel); for (i = 0; i < p->maxLevel; i++) { - p->concatBuf[i] = (struct lexConcatBuf *) - xmalloc (sizeof(**p->concatBuf)); - p->concatBuf[i]->len = p->concatBuf[i]->max = 0; - p->concatBuf[i]->buf = 0; + p->concatBuf[i].max = 0; + p->concatBuf[i].buf = 0; } p->d1_stack = (data1_node **) xmalloc (sizeof(*p->d1_stack) * p->maxLevel); p->d1_level = 0; @@ -476,7 +341,7 @@ static void lexSpecDestroy (struct lexSpec **pp) return ; for (i = 0; i < p->maxLevel; i++) - xfree (p->concatBuf[i]); + xfree (p->concatBuf[i].buf); xfree (p->concatBuf); lt = p->context; @@ -486,7 +351,7 @@ static void lexSpecDestroy (struct lexSpec **pp) lexContextDestroy (lt); lt = lt_next; } -#if HAVE_TCL_H +#if HAVE_TCL_OBJECTS if (p->tcl_interp) Tcl_DeleteInterp (p->tcl_interp); #endif @@ -504,7 +369,7 @@ static int readParseToken (const char **cpp, int *len) char cmd[32]; int i, level; - while (*cp == ' ' || *cp == '\t' || *cp == '\n') + while (*cp == ' ' || *cp == '\t' || *cp == '\n' || *cp == '\r') cp++; switch (*cp) { @@ -548,7 +413,8 @@ static int readParseToken (const char **cpp, int *len) { logf (LOG_WARN, "bad character %d %c", *cp, *cp); cp++; - while (*cp && *cp != ' ' && *cp != '\t' && *cp != '\n') + while (*cp && *cp != ' ' && *cp != '\t' && + *cp != '\n' && *cp != '\r') cp++; *cpp = cp; return 0; @@ -607,6 +473,8 @@ static int actionListMk (struct lexSpec *spec, const char *s, logf (LOG_WARN, "regular expression error '%.*s'", s-s0, s0); return -1; } + if (debug_dfa_tran) + printf ("pattern: %.*s\n", s-s0, s0); dfa_mkstate ((*ap)->u.pattern.dfa); s++; break; @@ -671,7 +539,7 @@ int readOneSpec (struct lexSpec *spec, const char *s) break; case REGX_PATTERN: #if REGX_DEBUG - logf (LOG_DEBUG, "rule %d %s", spec->context->ruleNo, s); + logf (LOG_LOG, "rule %d %s", spec->context->ruleNo, s); #endif r = dfa_parse (spec->context->dfa, &s); if (r) @@ -697,41 +565,48 @@ int readOneSpec (struct lexSpec *spec, const char *s) int readFileSpec (struct lexSpec *spec) { struct lexContext *lc; - char *lineBuf; - int lineSize = 512; int c, i, errors = 0; FILE *spec_inf = 0; + WRBUF lineBuf; + char fname[256]; - lineBuf = (char *) xmalloc (1+lineSize); #if HAVE_TCL_H if (spec->tcl_interp) { - sprintf (lineBuf, "%s.tflt", spec->name); - spec_inf = yaz_path_fopen (data1_get_tabpath(spec->dh), lineBuf, "r"); + sprintf (fname, "%s.tflt", spec->name); + spec_inf = data1_path_fopen (spec->dh, fname, "r"); } #endif if (!spec_inf) { - sprintf (lineBuf, "%s.flt", spec->name); - spec_inf = yaz_path_fopen (data1_get_tabpath(spec->dh), lineBuf, "r"); + sprintf (fname, "%s.flt", spec->name); + spec_inf = data1_path_fopen (spec->dh, fname, "r"); } if (!spec_inf) { logf (LOG_ERRNO|LOG_WARN, "cannot read spec file %s", spec->name); - xfree (lineBuf); return -1; } - logf (LOG_LOG, "reading regx filter %s", lineBuf); + logf (LOG_LOG, "reading regx filter %s", fname); #if HAVE_TCL_H if (spec->tcl_interp) logf (LOG_LOG, "Tcl enabled"); #endif + +#if 0 + debug_dfa_trav = 0; + debug_dfa_tran = 1; + debug_dfa_followpos = 0; + dfa_verbose = 1; +#endif + + lineBuf = wrbuf_alloc(); spec->lineNo = 0; c = getc (spec_inf); while (c != EOF) { - int off = 0; - if (c == '#' || c == '\n' || c == ' ' || c == '\t') + wrbuf_rewind (lineBuf); + if (c == '#' || c == '\n' || c == ' ' || c == '\t' || c == '\r') { while (c != '\n' && c != EOF) c = getc (spec_inf); @@ -742,12 +617,14 @@ int readFileSpec (struct lexSpec *spec) else { int addLine = 0; - - lineBuf[off++] = c; + while (1) { int c1 = c; + wrbuf_putc(lineBuf, c); c = getc (spec_inf); + while (c == '\r') + c = getc (spec_inf); if (c == EOF) break; if (c1 == '\n') @@ -756,24 +633,15 @@ int readFileSpec (struct lexSpec *spec) break; addLine++; } - lineBuf[off] = c; - if (off < lineSize) - off++; } - lineBuf[off] = '\0'; - readOneSpec (spec, lineBuf); + wrbuf_putc(lineBuf, '\0'); + readOneSpec (spec, wrbuf_buf(lineBuf)); spec->lineNo += addLine; } } fclose (spec_inf); - xfree (lineBuf); + wrbuf_free(lineBuf, 1); -#if 0 - debug_dfa_trav = 1; - debug_dfa_tran = 1; - debug_dfa_followpos = 1; - dfa_verbose = 1; -#endif for (lc = spec->context; lc; lc = lc->next) { struct lexRule *rp; @@ -804,13 +672,17 @@ static void execData (struct lexSpec *spec, if (elen == 0) /* shouldn't happen, but it does! */ return ; #if REGX_DEBUG - if (elen > 40) - logf (LOG_DEBUG, "data (%d bytes) %.15s ... %.*s", elen, - ebuf, 15, ebuf + elen-15); + if (elen > 80) + logf (LOG_LOG, "data(%d bytes) %.40s ... %.*s", elen, + ebuf, 40, ebuf + elen-40); + else if (elen == 1 && ebuf[0] == '\n') + { + logf (LOG_LOG, "data(new line)"); + } else if (elen > 0) - logf (LOG_DEBUG, "data (%d bytes) %.*s", elen, elen, ebuf); + logf (LOG_LOG, "data(%d bytes) %.*s", elen, elen, ebuf); else - logf (LOG_DEBUG, "data (%d bytes)", elen); + logf (LOG_LOG, "data(%d bytes)", elen); #endif if (spec->d1_level <= 1) @@ -825,45 +697,30 @@ static void execData (struct lexSpec *spec, { org_len = 0; - res = data1_mk_node (spec->dh, spec->m); - res->parent = parent; - res->which = DATA1N_data; + res = data1_mk_node2 (spec->dh, spec->m, DATA1N_data, parent); res->u.data.what = DATA1I_text; res->u.data.len = 0; res->u.data.formatted_text = formatted_text; -#if 0 - if (elen > DATA1_LOCALDATA) - res->u.data.data = nmem_malloc (spec->m, elen); - else - res->u.data.data = res->lbuf; - memcpy (res->u.data.data, ebuf, elen); -#else res->u.data.data = 0; -#endif - res->root = parent->root; - parent->last_child = res; if (spec->d1_stack[spec->d1_level]) spec->d1_stack[spec->d1_level]->next = res; - else - parent->child = res; spec->d1_stack[spec->d1_level] = res; } - if (org_len + elen >= spec->concatBuf[spec->d1_level]->max) + if (org_len + elen >= spec->concatBuf[spec->d1_level].max) { char *old_buf, *new_buf; - spec->concatBuf[spec->d1_level]->max = org_len + elen + 256; - new_buf = (char *) xmalloc (spec->concatBuf[spec->d1_level]->max); - if ((old_buf = spec->concatBuf[spec->d1_level]->buf)) + spec->concatBuf[spec->d1_level].max = org_len + elen + 256; + new_buf = (char *) xmalloc (spec->concatBuf[spec->d1_level].max); + if ((old_buf = spec->concatBuf[spec->d1_level].buf)) { memcpy (new_buf, old_buf, org_len); xfree (old_buf); } - spec->concatBuf[spec->d1_level]->buf = new_buf; + spec->concatBuf[spec->d1_level].buf = new_buf; } - assert (spec->concatBuf[spec->d1_level]); - memcpy (spec->concatBuf[spec->d1_level]->buf + org_len, ebuf, elen); + memcpy (spec->concatBuf[spec->d1_level].buf + org_len, ebuf, elen); res->u.data.len += elen; } @@ -887,7 +744,7 @@ static void tagDataRelease (struct lexSpec *spec) res->u.data.data = (char *) nmem_malloc (spec->m, res->u.data.len); else res->u.data.data = res->lbuf; - memcpy (res->u.data.data, spec->concatBuf[spec->d1_level]->buf, + memcpy (res->u.data.data, spec->concatBuf[spec->d1_level].buf, res->u.data.len); } } @@ -919,7 +776,7 @@ static void variantBegin (struct lexSpec *spec, ttype[type_len] = '\0'; #if REGX_DEBUG - logf (LOG_DEBUG, "variant begin %s %s (%d)", tclass, ttype, + logf (LOG_LOG, "variant begin(%s,%s,%d)", tclass, ttype, spec->d1_level); #endif @@ -930,21 +787,9 @@ static void variantBegin (struct lexSpec *spec, if (parent->which != DATA1N_variant) { - res = data1_mk_node (spec->dh, spec->m); - res->parent = parent; - res->which = DATA1N_variant; - res->u.variant.type = 0; - res->u.variant.value = 0; - res->root = parent->root; - - parent->last_child = res; + res = data1_mk_node2 (spec->dh, spec->m, DATA1N_variant, parent); if (spec->d1_stack[spec->d1_level]) - { tagDataRelease (spec); - spec->d1_stack[spec->d1_level]->next = res; - } - else - parent->child = res; spec->d1_stack[spec->d1_level] = res; spec->d1_stack[++(spec->d1_level)] = NULL; } @@ -956,13 +801,10 @@ static void variantBegin (struct lexSpec *spec, } #if REGX_DEBUG - logf (LOG_DEBUG, "variant node (%d)", spec->d1_level); + logf (LOG_LOG, "variant node(%d)", spec->d1_level); #endif parent = spec->d1_stack[spec->d1_level-1]; - res = data1_mk_node (spec->dh, spec->m); - res->parent = parent; - res->which = DATA1N_variant; - res->root = parent->root; + res = data1_mk_node2 (spec->dh, spec->m, DATA1N_variant, parent); res->u.variant.type = tp; if (value_len >= DATA1_LOCALDATA) @@ -972,14 +814,8 @@ static void variantBegin (struct lexSpec *spec, res->u.variant.value = res->lbuf; - parent->last_child = res; if (spec->d1_stack[spec->d1_level]) - { tagDataRelease (spec); - spec->d1_stack[spec->d1_level]->next = res; - } - else - parent->child = res; spec->d1_stack[spec->d1_level] = res; spec->d1_stack[++(spec->d1_level)] = NULL; } @@ -1000,60 +836,21 @@ static void tagStrip (const char **tag, int *len) static void tagBegin (struct lexSpec *spec, const char *tag, int len) { - struct data1_node *parent = spec->d1_stack[spec->d1_level -1]; - data1_element *elem = NULL; - data1_node *partag = get_parent_tag(spec->dh, parent); - data1_node *res; - data1_element *e = NULL; - int localtag = 0; - if (spec->d1_level == 0) { logf (LOG_WARN, "in element begin. No record type defined"); return ; } tagStrip (&tag, &len); - - res = data1_mk_node (spec->dh, spec->m); - res->parent = parent; - res->which = DATA1N_tag; - res->u.tag.get_bytes = -1; - - if (len >= DATA1_LOCALDATA) - res->u.tag.tag = (char *) nmem_malloc (spec->m, len+1); - else - res->u.tag.tag = res->lbuf; + if (spec->d1_stack[spec->d1_level]) + tagDataRelease (spec); - memcpy (res->u.tag.tag, tag, len); - res->u.tag.tag[len] = '\0'; - #if REGX_DEBUG - logf (LOG_DEBUG, "begin tag %s (%d)", res->u.tag.tag, spec->d1_level); + logf (LOG_LOG, "begin tag(%.*s, %d)", len, tag, spec->d1_level); #endif - if (parent->which == DATA1N_variant) - return ; - if (partag) - if (!(e = partag->u.tag.element)) - localtag = 1; - - elem = data1_getelementbytagname (spec->dh, - spec->d1_stack[0]->u.root.absyn, - e, res->u.tag.tag); - res->u.tag.element = elem; - res->u.tag.node_selected = 0; - res->u.tag.make_variantlist = 0; - res->u.tag.no_data_requested = 0; - res->root = parent->root; - - parent->last_child = res; - if (spec->d1_stack[spec->d1_level]) - { - tagDataRelease (spec); - spec->d1_stack[spec->d1_level]->next = res; - } - else - parent->child = res; - spec->d1_stack[spec->d1_level] = res; + + spec->d1_stack[spec->d1_level] = data1_mk_tag_n ( + spec->dh, spec->m, tag, len, 0, spec->d1_stack[spec->d1_level -1]); spec->d1_stack[++(spec->d1_level)] = NULL; } @@ -1075,27 +872,39 @@ static void tagEnd (struct lexSpec *spec, int min_level, break; } #if REGX_DEBUG - logf (LOG_DEBUG, "end tag (%d)", spec->d1_level); + logf (LOG_LOG, "end tag(%d)", spec->d1_level); #endif } static int tryMatch (struct lexSpec *spec, int *pptr, int *mptr, - struct DFA *dfa) + struct DFA *dfa, int greedy) { struct DFA_state *state = dfa->states[0]; struct DFA_tran *t; - unsigned char c; + unsigned char c = 0; unsigned char c_prev = 0; int ptr = *pptr; /* current pointer */ int start_ptr = *pptr; /* first char of match */ int last_ptr = 0; /* last char of match */ int last_rule = 0; /* rule number of current match */ + int restore_ptr = 0; int i; + if (ptr) + { + --ptr; + c = f_win_advance (spec, &ptr); + } while (1) { + if (dfa->states[0] == state) + { + c_prev = c; + restore_ptr = ptr; + } c = f_win_advance (spec, &ptr); + if (ptr == F_WIN_EOF) { if (last_rule) @@ -1106,10 +915,11 @@ static int tryMatch (struct lexSpec *spec, int *pptr, int *mptr, } break; } + t = state->trans; i = state->tran_no; while (1) - if (--i < 0) + if (--i < 0) /* no transition for character c */ { if (last_rule) { @@ -1118,27 +928,28 @@ static int tryMatch (struct lexSpec *spec, int *pptr, int *mptr, return 1; } state = dfa->states[0]; + + ptr = restore_ptr; + c = f_win_advance (spec, &ptr); + start_ptr = ptr; - c_prev = c; + break; } else if (c >= t->ch[0] && c <= t->ch[1]) { state = dfa->states[t->to]; - if (state->rule_no) - { - if (c_prev == '\n') - { - last_rule = state->rule_no; - last_ptr = ptr; - } - else - { - last_rule = state->rule_nno; - last_ptr = ptr; - } - } - break; + if (state->rule_no && c_prev == '\n') + { + last_rule = state->rule_no; + last_ptr = ptr; + } + else if (state->rule_nno) + { + last_rule = state->rule_nno; + last_ptr = ptr; + } + break; } else t++; @@ -1192,7 +1003,8 @@ static int execTok (struct lexSpec *spec, const char **src, else if (*s == '-') { *tokBuf = s++; - while (*s && *s != ' ' && *s != '\t' && *s != '\n' && *s != ';') + while (*s && *s != ' ' && *s != '\t' && *s != '\n' && *s != '\r' && + *s != ';') s++; *tokLen = s - *tokBuf; *src = s; @@ -1201,7 +1013,8 @@ static int execTok (struct lexSpec *spec, const char **src, else { *tokBuf = s++; - while (*s && *s != ' ' && *s != '\t' && *s != '\n' && *s != ';') + while (*s && *s != ' ' && *s != '\t' && *s != '\n' && *s != '\r' && + *s != ';') s++; *tokLen = s - *tokBuf; } @@ -1220,34 +1033,30 @@ static char *regxStrz (const char *src, int len, char *str) #if HAVE_TCL_H static int cmd_tcl_begin (ClientData clientData, Tcl_Interp *interp, - int argc, char **argv) + int argc, const char **argv) { struct lexSpec *spec = (struct lexSpec *) clientData; if (argc < 2) return TCL_ERROR; if (!strcmp(argv[1], "record") && argc == 3) { - char *absynName = argv[2]; - data1_absyn *absyn; + const char *absynName = argv[2]; + data1_node *res; #if REGX_DEBUG - logf (LOG_DEBUG, "begin record %s", absynName); + logf (LOG_LOG, "begin record %s", absynName); #endif - if (!(absyn = data1_get_absyn (spec->dh, absynName))) - logf (LOG_WARN, "Unknown tagset: %s", absynName); - else - { - data1_node *res; - - res = data1_mk_node (spec->dh, spec->m); - res->which = DATA1N_root; - res->u.root.type = absynName; - res->u.root.absyn = absyn; - res->root = res; - - spec->d1_stack[spec->d1_level] = res; - spec->d1_stack[++(spec->d1_level)] = NULL; - } + res = data1_mk_root (spec->dh, spec->m, absynName); + + spec->d1_level = 0; + + spec->d1_stack[spec->d1_level++] = res; + + res = data1_mk_tag (spec->dh, spec->m, absynName, 0, res); + + spec->d1_stack[spec->d1_level++] = res; + + spec->d1_stack[spec->d1_level] = NULL; } else if (!strcmp(argv[1], "element") && argc == 3) { @@ -1263,7 +1072,7 @@ static int cmd_tcl_begin (ClientData clientData, Tcl_Interp *interp, { struct lexContext *lc = spec->context; #if REGX_DEBUG - logf (LOG_DEBUG, "begin context %s",argv[2]); + logf (LOG_LOG, "begin context %s",argv[2]); #endif while (lc && strcmp (argv[2], lc->name)) lc = lc->next; @@ -1280,7 +1089,7 @@ static int cmd_tcl_begin (ClientData clientData, Tcl_Interp *interp, } static int cmd_tcl_end (ClientData clientData, Tcl_Interp *interp, - int argc, char **argv) + int argc, const char **argv) { struct lexSpec *spec = (struct lexSpec *) clientData; if (argc < 2) @@ -1294,14 +1103,14 @@ static int cmd_tcl_end (ClientData clientData, Tcl_Interp *interp, (spec->d1_level)--; } #if REGX_DEBUG - logf (LOG_DEBUG, "end record"); + logf (LOG_LOG, "end record"); #endif spec->stop_flag = 1; } else if (!strcmp (argv[1], "element")) { - int min_level = 1; - char *element = 0; + int min_level = 2; + const char *element = 0; if (argc >= 3 && !strcmp(argv[2], "-record")) { min_level = 0; @@ -1312,10 +1121,10 @@ static int cmd_tcl_end (ClientData clientData, Tcl_Interp *interp, if (argc == 3) element = argv[2]; tagEnd (spec, min_level, element, (element ? strlen(element) : 0)); - if (spec->d1_level == 0) + if (spec->d1_level <= 1) { #if REGX_DEBUG - logf (LOG_DEBUG, "end element end records"); + logf (LOG_LOG, "end element end records"); #endif spec->stop_flag = 1; } @@ -1323,7 +1132,7 @@ static int cmd_tcl_end (ClientData clientData, Tcl_Interp *interp, else if (!strcmp (argv[1], "context")) { #if REGX_DEBUG - logf (LOG_DEBUG, "end context"); + logf (LOG_LOG, "end context"); #endif if (spec->context_stack_top) (spec->context_stack_top)--; @@ -1334,7 +1143,7 @@ static int cmd_tcl_end (ClientData clientData, Tcl_Interp *interp, } static int cmd_tcl_data (ClientData clientData, Tcl_Interp *interp, - int argc, char **argv) + int argc, const char **argv) { int argi = 1; int textFlag = 0; @@ -1362,16 +1171,23 @@ static int cmd_tcl_data (ClientData clientData, Tcl_Interp *interp, while (argi < argc) { +#if TCL_MAJOR_VERSION > 8 || (TCL_MAJOR_VERSION == 8 && TCL_MINOR_VERSION > 0) + Tcl_DString ds; + char *native = Tcl_UtfToExternalDString(0, argv[argi], -1, &ds); + execData (spec, native, strlen(native), textFlag); + Tcl_DStringFree (&ds); +#else execData (spec, argv[argi], strlen(argv[argi]), textFlag); +#endif argi++; } if (element) - tagEnd (spec, 1, NULL, 0); + tagEnd (spec, 2, NULL, 0); return TCL_OK; } static int cmd_tcl_unread (ClientData clientData, Tcl_Interp *interp, - int argc, char **argv) + int argc, const char **argv) { struct lexSpec *spec = (struct lexSpec *) clientData; int argi = 1; @@ -1404,6 +1220,7 @@ static int cmd_tcl_unread (ClientData clientData, Tcl_Interp *interp, static void execTcl (struct lexSpec *spec, struct regxCode *code) { int i; + int ret; for (i = 0; i < spec->arg_no; i++) { char var_name[10], *var_buf; @@ -1420,7 +1237,12 @@ static void execTcl (struct lexSpec *spec, struct regxCode *code) var_buf[var_len] = ch; } } - if (Tcl_Eval (spec->tcl_interp, code->str) != TCL_OK) +#if HAVE_TCL_OBJECTS + ret = Tcl_GlobalEvalObj(spec->tcl_interp, code->tcl_obj); +#else + ret = Tcl_GlobalEval (spec->tcl_interp, code->str); +#endif + if (ret != TCL_OK) { const char *err = Tcl_GetVar(spec->tcl_interp, "errorInfo", 0); logf(LOG_FATAL, "Tcl error, line=%d, \"%s\"\n%s", @@ -1428,7 +1250,6 @@ static void execTcl (struct lexSpec *spec, struct regxCode *code) spec->tcl_interp->result, err ? err : "[NO ERRORINFO]"); } - } /* HAVE_TCL_H */ #endif @@ -1464,34 +1285,29 @@ static void execCode (struct lexSpec *spec, struct regxCode *code) r = execTok (spec, &s, &cmd_str, &cmd_len); if (r < 2) continue; - if (spec->d1_level == 0) + if (spec->d1_level <= 1) { static char absynName[64]; - data1_absyn *absyn; + data1_node *res; if (cmd_len > 63) cmd_len = 63; memcpy (absynName, cmd_str, cmd_len); absynName[cmd_len] = '\0'; - #if REGX_DEBUG - logf (LOG_DEBUG, "begin record %s", absynName); + logf (LOG_LOG, "begin record %s", absynName); #endif - if (!(absyn = data1_get_absyn (spec->dh, absynName))) - logf (LOG_WARN, "Unknown tagset: %s", absynName); - else - { - data1_node *res; - - res = data1_mk_node (spec->dh, spec->m); - res->which = DATA1N_root; - res->u.root.type = absynName; - res->u.root.absyn = absyn; - res->root = res; - - spec->d1_stack[spec->d1_level] = res; - spec->d1_stack[++(spec->d1_level)] = NULL; - } + res = data1_mk_root (spec->dh, spec->m, absynName); + + spec->d1_level = 0; + + spec->d1_stack[spec->d1_level++] = res; + + res = data1_mk_tag (spec->dh, spec->m, absynName, 0, res); + + spec->d1_stack[spec->d1_level++] = res; + + spec->d1_stack[spec->d1_level] = NULL; } r = execTok (spec, &s, &cmd_str, &cmd_len); } @@ -1542,7 +1358,7 @@ static void execCode (struct lexSpec *spec, struct regxCode *code) r = execTok (spec, &s, &cmd_str, &cmd_len); p = regxStrz (cmd_str, cmd_len, ptmp); #if REGX_DEBUG - logf (LOG_DEBUG, "begin context %s", p); + logf (LOG_LOG, "begin context %s", p); #endif while (lc && strcmp (p, lc->name)) lc = lc->next; @@ -1577,13 +1393,13 @@ static void execCode (struct lexSpec *spec, struct regxCode *code) } r = execTok (spec, &s, &cmd_str, &cmd_len); #if REGX_DEBUG - logf (LOG_DEBUG, "end record"); + logf (LOG_LOG, "end record"); #endif spec->stop_flag = 1; } else if (!strcmp (p, "element")) { - int min_level = 1; + int min_level = 2; while ((r = execTok (spec, &s, &cmd_str, &cmd_len)) == 3) { if (cmd_len==7 && !memcmp ("-record", cmd_str, cmd_len)) @@ -1596,10 +1412,10 @@ static void execCode (struct lexSpec *spec, struct regxCode *code) } else tagEnd (spec, min_level, NULL, 0); - if (spec->d1_level == 0) + if (spec->d1_level <= 1) { #if REGX_DEBUG - logf (LOG_DEBUG, "end element end records"); + logf (LOG_LOG, "end element end records"); #endif spec->stop_flag = 1; } @@ -1608,7 +1424,7 @@ static void execCode (struct lexSpec *spec, struct regxCode *code) else if (!strcmp (p, "context")) { #if REGX_DEBUG - logf (LOG_DEBUG, "end context"); + logf (LOG_LOG, "end context"); #endif if (spec->context_stack_top) (spec->context_stack_top)--; @@ -1650,7 +1466,7 @@ static void execCode (struct lexSpec *spec, struct regxCode *code) r = execTok (spec, &s, &cmd_str, &cmd_len); } while (r > 1); if (element_str) - tagEnd (spec, 1, NULL, 0); + tagEnd (spec, 2, NULL, 0); } else if (!strcmp (p, "unread")) { @@ -1747,13 +1563,14 @@ static int execAction (struct lexSpec *spec, struct lexRuleAction *ap, if (ap->u.pattern.body) { arg_start[arg_no] = *pptr; - if (!tryMatch (spec, pptr, &sptr, ap->u.pattern.dfa)) + if (!tryMatch (spec, pptr, &sptr, ap->u.pattern.dfa, 0)) { arg_end[arg_no] = F_WIN_EOF; arg_no++; arg_start[arg_no] = F_WIN_EOF; arg_end[arg_no] = F_WIN_EOF; -/* return 1*/ + yaz_log(LOG_DEBUG, "Pattern match rest of record"); + *pptr = F_WIN_EOF; } else { @@ -1766,7 +1583,7 @@ static int execAction (struct lexSpec *spec, struct lexRuleAction *ap, else { arg_start[arg_no] = *pptr; - if (!tryMatch (spec, pptr, &sptr, ap->u.pattern.dfa)) + if (!tryMatch (spec, pptr, &sptr, ap->u.pattern.dfa, 1)) return 1; if (sptr != arg_start[arg_no]) return 1; @@ -1804,7 +1621,7 @@ static int execRule (struct lexSpec *spec, struct lexContext *context, int ruleNo, int start_ptr, int *pptr) { #if REGX_DEBUG - logf (LOG_DEBUG, "exec rule %d", ruleNo); + logf (LOG_LOG, "exec rule %d", ruleNo); #endif return execAction (spec, context->fastRule[ruleNo]->actionList, start_ptr, pptr); @@ -1882,7 +1699,7 @@ data1_node *lexNode (struct lexSpec *spec, int *ptr) if (spec->f_win_ef && *ptr != F_WIN_EOF) { #if REGX_DEBUG - logf (LOG_DEBUG, "regx: endf ptr=%d", *ptr); + logf (LOG_LOG, "regx: endf ptr=%d", *ptr); #endif (*spec->f_win_ef)(spec->f_win_fh, *ptr); } @@ -1994,7 +1811,7 @@ data1_node *grs_read_regx (struct grs_read_info *p) struct lexSpec **curLexSpec = &specs->spec; #if REGX_DEBUG - logf (LOG_DEBUG, "grs_read_regx"); + logf (LOG_LOG, "grs_read_regx"); #endif if (!*curLexSpec || strcmp ((*curLexSpec)->name, p->type)) { @@ -2040,7 +1857,7 @@ data1_node *grs_read_tcl (struct grs_read_info *p) struct lexSpec **curLexSpec = &specs->spec; #if REGX_DEBUG - logf (LOG_DEBUG, "grs_read_tcl"); + logf (LOG_LOG, "grs_read_tcl"); #endif if (!*curLexSpec || strcmp ((*curLexSpec)->name, p->type)) { @@ -2048,7 +1865,9 @@ data1_node *grs_read_tcl (struct grs_read_info *p) if (*curLexSpec) lexSpecDestroy (curLexSpec); *curLexSpec = lexSpecCreate (p->type, p->dh); + Tcl_FindExecutable(""); tcl_interp = (*curLexSpec)->tcl_interp = Tcl_CreateInterp(); + Tcl_Init(tcl_interp); Tcl_CreateCommand (tcl_interp, "begin", cmd_tcl_begin, *curLexSpec, 0); Tcl_CreateCommand (tcl_interp, "end", cmd_tcl_end, *curLexSpec, 0); Tcl_CreateCommand (tcl_interp, "data", cmd_tcl_data, *curLexSpec, 0);