X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;f=recctrl%2Fregxread.c;h=299f2ef009d2dbd5775e531933eb9e05c1ff3371;hb=71647c93775e998a0e602c0371c34e1143d131dd;hp=bb26858b7fe941dc9913d4a5d4e1d60e83b62a91;hpb=0a50c507e32a11d9ef0d1f8458b02590044f25d5;p=idzebra-moved-to-github.git diff --git a/recctrl/regxread.c b/recctrl/regxread.c index bb26858..299f2ef 100644 --- a/recctrl/regxread.c +++ b/recctrl/regxread.c @@ -4,7 +4,39 @@ * Sebastian Hammer, Adam Dickmeiss * * $Log: regxread.c,v $ - * Revision 1.24 1999-05-21 11:08:46 adam + * Revision 1.34 2000-11-29 14:24:01 adam + * Script configure uses yaz pthreads options. Added locking for + * zebra_register_{lock,unlock}. + * + * Revision 1.33 1999/11/30 13:48:04 adam + * Improved installation. Updated for inclusion of YAZ header files. + * + * Revision 1.32 1999/09/07 07:19:21 adam + * Work on character mapping. Implemented replace rules. + * + * Revision 1.31 1999/07/14 13:05:29 adam + * Tcl filter works with objects when TCL is version 8 or later; filter + * works with strings otherwise (slow). + * + * Revision 1.30 1999/07/14 10:55:28 adam + * Fixed memory leak. + * + * Revision 1.29 1999/07/12 07:27:54 adam + * Improved speed of Tcl processing. Fixed one memory leak. + * + * Revision 1.28 1999/07/06 12:26:04 adam + * Fixed filters so that MS-DOS CR is ignored. + * + * 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. * @@ -167,13 +199,17 @@ #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 @@ -192,6 +228,9 @@ struct regxCode { char *str; +#if HAVE_TCL_OBJECTS + Tcl_Obj *tcl_obj; +#endif }; struct lexRuleAction { @@ -231,7 +270,6 @@ struct lexContext { }; struct lexConcatBuf { - int len; int max; char *buf; }; @@ -260,7 +298,7 @@ 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; @@ -292,7 +330,7 @@ static char *f_win_get (struct lexSpec *spec, off_t start_pos, off_t end_pos, spec->f_win_start = start_pos; if (!spec->f_win_buf) - spec->f_win_buf = xmalloc (spec->f_win_size); + spec->f_win_buf = (char *) xmalloc (spec->f_win_size); *size = (*spec->f_win_rf)(spec->f_win_fh, spec->f_win_buf, spec->f_win_size); spec->f_win_end = spec->f_win_start + *size; @@ -338,6 +376,10 @@ 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; @@ -348,10 +390,15 @@ static void regxCodeMk (struct regxCode **pp, const char *buf, int len) { struct regxCode *p; - p = xmalloc (sizeof(*p)); - p->str = xmalloc (len+1); + p = (struct regxCode *) xmalloc (sizeof(*p)); + 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; } @@ -389,7 +436,7 @@ static void actionListDel (struct lexRuleAction **rap) static struct lexContext *lexContextCreate (const char *name) { - struct lexContext *p = xmalloc (sizeof(*p)); + struct lexContext *p = (struct lexContext *) xmalloc (sizeof(*p)); p->name = xstrdup (name); p->ruleNo = 1; @@ -408,6 +455,7 @@ static void lexContextDestroy (struct lexContext *p) { struct lexRule *rp, *rp1; + dfa_delete (&p->dfa); xfree (p->fastRule); for (rp = p->rules; rp; rp = rp1) { @@ -417,6 +465,7 @@ static void lexContextDestroy (struct lexContext *p) } actionListDel (&p->beginActionList); actionListDel (&p->endActionList); + actionListDel (&p->initActionList); xfree (p->name); xfree (p); } @@ -426,8 +475,8 @@ static struct lexSpec *lexSpecCreate (const char *name, data1_handle dh) struct lexSpec *p; int i; - p = xmalloc (sizeof(*p)); - p->name = xmalloc (strlen(name)+1); + p = (struct lexSpec *) xmalloc (sizeof(*p)); + p->name = (char *) xmalloc (strlen(name)+1); strcpy (p->name, name); #if HAVE_TCL_H @@ -436,19 +485,19 @@ static struct lexSpec *lexSpecCreate (const char *name, data1_handle dh) p->dh = dh; p->context = NULL; p->context_stack_size = 100; - p->context_stack = xmalloc (sizeof(*p->context_stack) * - p->context_stack_size); + p->context_stack = (struct lexContext **) + xmalloc (sizeof(*p->context_stack) * p->context_stack_size); p->f_win_buf = NULL; p->maxLevel = 128; - p->concatBuf = xmalloc (sizeof(*p->concatBuf) * p->maxLevel); + p->concatBuf = (struct lexConcatBuf *) + xmalloc (sizeof(*p->concatBuf) * p->maxLevel); for (i = 0; i < p->maxLevel; i++) { - p->concatBuf[i] = 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 = xmalloc (sizeof(*p->d1_stack) * p->maxLevel); + p->d1_stack = (data1_node **) xmalloc (sizeof(*p->d1_stack) * p->maxLevel); p->d1_level = 0; return p; } @@ -465,7 +514,7 @@ 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; @@ -475,7 +524,7 @@ 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 @@ -493,7 +542,7 @@ 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) { @@ -528,7 +577,7 @@ static int readParseToken (const char **cpp, int *len) cmd[i] = *cp + 'a' - 'A'; else break; - if (i < sizeof(cmd)-2) + if (i < (int) sizeof(cmd)-2) i++; cp++; } @@ -537,7 +586,8 @@ 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; @@ -576,13 +626,13 @@ static int actionListMk (struct lexSpec *spec, const char *s, bodyMark = 1; continue; case REGX_CODE: - *ap = xmalloc (sizeof(**ap)); + *ap = (struct lexRuleAction *) xmalloc (sizeof(**ap)); (*ap)->which = tok; regxCodeMk (&(*ap)->u.code, s, len); s += len+1; break; case REGX_PATTERN: - *ap = xmalloc (sizeof(**ap)); + *ap = (struct lexRuleAction *) xmalloc (sizeof(**ap)); (*ap)->which = tok; (*ap)->u.pattern.body = bodyMark; bodyMark = 0; @@ -606,7 +656,7 @@ static int actionListMk (struct lexSpec *spec, const char *s, logf (LOG_WARN, "cannot use INIT here"); continue; case REGX_END: - *ap = xmalloc (sizeof(**ap)); + *ap = (struct lexRuleAction *) xmalloc (sizeof(**ap)); (*ap)->which = tok; break; } @@ -674,7 +724,7 @@ int readOneSpec (struct lexSpec *spec, const char *s) return -1; } s++; - rp = xmalloc (sizeof(*rp)); + rp = (struct lexRule *) xmalloc (sizeof(*rp)); rp->info.no = spec->context->ruleNo++; rp->next = spec->context->rules; spec->context->rules = rp; @@ -686,49 +736,40 @@ 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 = 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 = yaz_path_fopen (data1_get_tabpath(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 = yaz_path_fopen (data1_get_tabpath(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.flt", lineBuf); - sprintf (lineBuf, "%s.flt", spec->name); - if (!(spec_inf = yaz_path_fopen (data1_get_tabpath(spec->dh), - lineBuf, "r"))) - { - logf (LOG_ERRNO|LOG_WARN, "cannot read spec file %s", spec->name); - xfree (lineBuf); return -1; } + logf (LOG_LOG, "reading regx filter %s", fname); #if HAVE_TCL_H if (spec->tcl_interp) logf (LOG_LOG, "Tcl enabled"); #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); @@ -739,12 +780,14 @@ 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') @@ -753,17 +796,14 @@ 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; @@ -774,7 +814,8 @@ int readFileSpec (struct lexSpec *spec) for (lc = spec->context; lc; lc = lc->next) { struct lexRule *rp; - lc->fastRule = xmalloc (sizeof(*lc->fastRule) * lc->ruleNo); + lc->fastRule = (struct lexRuleInfo **) + xmalloc (sizeof(*lc->fastRule) * lc->ruleNo); for (i = 0; i < lc->ruleNo; i++) lc->fastRule[i] = NULL; for (rp = lc->rules; rp; rp = rp->next) @@ -845,21 +886,20 @@ static void execData (struct lexSpec *spec, 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 = 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; } @@ -880,10 +920,10 @@ static void tagDataRelease (struct lexSpec *spec) assert (!res->u.data.data); assert (res->u.data.len > 0); if (res->u.data.len > DATA1_LOCALDATA) - res->u.data.data = nmem_malloc (spec->m, res->u.data.len); + 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); } } @@ -996,9 +1036,9 @@ 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]; + struct data1_node *parent; data1_element *elem = NULL; - data1_node *partag = get_parent_tag(spec->dh, parent); + data1_node *partag; data1_node *res; data1_element *e = NULL; int localtag = 0; @@ -1009,14 +1049,15 @@ static void tagBegin (struct lexSpec *spec, return ; } tagStrip (&tag, &len); + + parent = spec->d1_stack[spec->d1_level -1]; + partag = get_parent_tag(spec->dh, parent); - res = data1_mk_node (spec->dh, spec->m); + res = data1_mk_node_type (spec->dh, spec->m, DATA1N_tag); res->parent = parent; - res->which = DATA1N_tag; - res->u.tag.get_bytes = -1; if (len >= DATA1_LOCALDATA) - res->u.tag.tag = nmem_malloc (spec->m, len+1); + res->u.tag.tag = (char *) nmem_malloc (spec->m, len+1); else res->u.tag.tag = res->lbuf; @@ -1036,9 +1077,6 @@ static void tagBegin (struct lexSpec *spec, 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; @@ -1188,7 +1226,8 @@ 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; @@ -1197,7 +1236,8 @@ 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; } @@ -1218,7 +1258,7 @@ static char *regxStrz (const char *src, int len, char *str) static int cmd_tcl_begin (ClientData clientData, Tcl_Interp *interp, int argc, char **argv) { - struct lexSpec *spec = clientData; + struct lexSpec *spec = (struct lexSpec *) clientData; if (argc < 2) return TCL_ERROR; if (!strcmp(argv[1], "record") && argc == 3) @@ -1278,7 +1318,7 @@ static int cmd_tcl_begin (ClientData clientData, Tcl_Interp *interp, static int cmd_tcl_end (ClientData clientData, Tcl_Interp *interp, int argc, char **argv) { - struct lexSpec *spec = clientData; + struct lexSpec *spec = (struct lexSpec *) clientData; if (argc < 2) return TCL_ERROR; @@ -1298,17 +1338,15 @@ static int cmd_tcl_end (ClientData clientData, Tcl_Interp *interp, { int min_level = 1; char *element = 0; - if (!strcmp(argv[2], "-record")) + if (argc >= 3 && !strcmp(argv[2], "-record")) { min_level = 0; if (argc == 4) element = argv[3]; } else - { if (argc == 3) element = argv[2]; - } tagEnd (spec, min_level, element, (element ? strlen(element) : 0)); if (spec->d1_level == 0) { @@ -1337,7 +1375,7 @@ static int cmd_tcl_data (ClientData clientData, Tcl_Interp *interp, int argi = 1; int textFlag = 0; const char *element = 0; - struct lexSpec *spec = clientData; + struct lexSpec *spec = (struct lexSpec *) clientData; while (argi < argc) { @@ -1371,7 +1409,7 @@ static int cmd_tcl_data (ClientData clientData, Tcl_Interp *interp, static int cmd_tcl_unread (ClientData clientData, Tcl_Interp *interp, int argc, char **argv) { - struct lexSpec *spec = clientData; + struct lexSpec *spec = (struct lexSpec *) clientData; int argi = 1; int offset = 0; int no; @@ -1402,6 +1440,7 @@ 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; @@ -1418,7 +1457,19 @@ static void execTcl (struct lexSpec *spec, struct regxCode *code) var_buf[var_len] = ch; } } - Tcl_Eval (spec->tcl_interp, code->str); +#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", + spec->tcl_interp->errorLine, + spec->tcl_interp->result, + err ? err : "[NO ERRORINFO]"); + } } /* HAVE_TCL_H */ #endif @@ -1962,7 +2013,7 @@ static data1_node *lexRoot (struct lexSpec *spec, off_t offset, void grs_destroy(void *clientData) { - struct lexSpecs *specs = clientData; + struct lexSpecs *specs = (struct lexSpecs *) clientData; if (specs->spec) { lexSpecDestroy(&specs->spec); @@ -1972,7 +2023,7 @@ void grs_destroy(void *clientData) void *grs_init(void) { - struct lexSpecs *specs = xmalloc (sizeof(*specs)); + struct lexSpecs *specs = (struct lexSpecs *) xmalloc (sizeof(*specs)); specs->spec = 0; return specs; } @@ -1980,7 +2031,7 @@ void *grs_init(void) data1_node *grs_read_regx (struct grs_read_info *p) { int res; - struct lexSpecs *specs = p->clientData; + struct lexSpecs *specs = (struct lexSpecs *) p->clientData; struct lexSpec **curLexSpec = &specs->spec; #if REGX_DEBUG @@ -2026,7 +2077,7 @@ RecTypeGrs recTypeGrs_regx = ®x_type; data1_node *grs_read_tcl (struct grs_read_info *p) { int res; - struct lexSpecs *specs = p->clientData; + struct lexSpecs *specs = (struct lexSpecs *) p->clientData; struct lexSpec **curLexSpec = &specs->spec; #if REGX_DEBUG