X-Git-Url: http://git.indexdata.com/?p=idzebra-moved-to-github.git;a=blobdiff_plain;f=rset%2Frstemp.c;h=b06a1fa2b0729fb18e4183cf573957a647637479;hp=e7a0a4c3abcfead619f94f224944af97421319cb;hb=a8b86c35a39ce9eb5279c46dd99ef7e91b83fb7b;hpb=7eefa059dbe0c2d4ba24060421940df637d9cca8 diff --git a/rset/rstemp.c b/rset/rstemp.c index e7a0a4c..b06a1fa 100644 --- a/rset/rstemp.c +++ b/rset/rstemp.c @@ -1,159 +1,196 @@ -/* - * Copyright (C) 1994-1995, Index Data I/S - * All rights reserved. - * Sebastian Hammer, Adam Dickmeiss - * - * $Log: rstemp.c,v $ - * Revision 1.8 1995-09-08 14:52:42 adam - * Work on relevance feedback. - * - * Revision 1.7 1995/09/07 13:58:44 adam - * New parameter: result-set file descriptor (RSFD) to support multiple - * positions within the same result-set. - * Boolean operators: and, or, not implemented. - * - * Revision 1.6 1995/09/06 16:11:56 adam - * More work on boolean sets. - * - * Revision 1.5 1995/09/05 16:36:59 adam - * Minor changes. - * - * Revision 1.4 1995/09/05 11:43:24 adam - * Complete version of temporary sets. Not tested yet though. - * - * Revision 1.3 1995/09/04 15:20:40 adam - * More work on temp sets. is_open member removed. - * - * Revision 1.2 1995/09/04 09:10:56 adam - * Minor changes. - * - * Revision 1.1 1994/11/04 13:21:30 quinn - * Working. - * - */ +/* $Id: rstemp.c,v 1.42 2004-08-06 12:55:03 adam Exp $ + Copyright (C) 1995,1996,1997,1998,1999,2000,2001,2002,2003 + 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 +#ifdef WIN32 +#include +#else #include +#endif +#include #include #include -#include +#include #include -static rset_control *r_create(const struct rset_control *sel, void *parms); -static RSFD r_open (rset_control *ct, int wflag); +static void *r_create(RSET ct, const struct rset_control *sel, void *parms); +static RSFD r_open (RSET ct, int flag); static void r_close (RSFD rfd); -static void r_delete (rset_control *ct); +static void r_delete (RSET ct); static void r_rewind (RSFD rfd); -static int r_count (rset_control *ct); -static int r_read (RSFD rfd, void *buf); +/* static int r_count (RSET ct);*/ +static int r_read (RSFD rfd, void *buf, int *term_index); static int r_write (RSFD rfd, const void *buf); +static void r_pos (RSFD rfd, double *current, double *total); -static const rset_control control = +static const struct rset_control control = { - "Temporary set", - 0, + "temp", r_create, r_open, r_close, r_delete, r_rewind, - r_count, + rset_default_forward, + r_pos, r_read, - r_write + r_write, }; -const rset_control *rset_kind_temp = &control; +const struct rset_control *rset_kind_temp = &control; struct rset_temp_info { int fd; char *fname; - size_t key_size; - char *buf_mem; - size_t buf_size; - size_t pos_end; - size_t pos_cur; - size_t pos_buf; - size_t pos_border; - int dirty; + size_t key_size; /* key size */ + char *buf_mem; /* window buffer */ + size_t buf_size; /* size of window */ + size_t pos_end; /* last position in set */ + size_t pos_buf; /* position of first byte in window */ + size_t pos_border; /* position of last byte+1 in window */ + int dirty; /* window is dirty */ + zint hits; /* no of hits */ + char *temp_path; + int (*cmp)(const void *p1, const void *p2); + struct rset_temp_rfd *rfd_list; }; struct rset_temp_rfd { struct rset_temp_info *info; struct rset_temp_rfd *next; + zint *countp; + void *buf; + size_t pos_cur; /* current position in set */ + zint cur; /* number of the current hit */ }; -static struct rset_control *r_create(const struct rset_control *sel, - void *parms) +static void *r_create(RSET ct, const struct rset_control *sel, void *parms) { - rset_control *newct; - rset_temp_parms *temp_parms = parms; + rset_temp_parms *temp_parms = (rset_temp_parms *) parms; struct rset_temp_info *info; - - newct = xmalloc(sizeof(*newct)); - memcpy(newct, sel, sizeof(*sel)); - newct->buf = xmalloc (sizeof(struct rset_temp_info)); - info = newct->buf; - + + info = (struct rset_temp_info *) xmalloc (sizeof(struct rset_temp_info)); info->fd = -1; info->fname = NULL; info->key_size = temp_parms->key_size; - info->buf_size = 1024; - info->buf_mem = xmalloc (info->buf_size); - info->pos_cur = 0; + info->buf_size = 4096; + info->buf_mem = (char *) xmalloc (info->buf_size); info->pos_end = 0; info->pos_buf = 0; info->dirty = 0; + info->hits = 0; + info->cmp = temp_parms->cmp; + info->rfd_list = NULL; - return newct; + if (!temp_parms->temp_path) + info->temp_path = NULL; + else + { + info->temp_path = (char *) xmalloc (strlen(temp_parms->temp_path)+1); + strcpy (info->temp_path, temp_parms->temp_path); + } + ct->no_rset_terms = 1; + ct->rset_terms = (RSET_TERM *) xmalloc (sizeof(*ct->rset_terms)); + ct->rset_terms[0] = temp_parms->rset_term; + + return info; } -static RSFD r_open (struct rset_control *ct, int wflag) +static RSFD r_open (RSET ct, int flag) { - struct rset_temp_info *info = ct->buf; + struct rset_temp_info *info = (struct rset_temp_info *) ct->buf; struct rset_temp_rfd *rfd; - assert (info->fd == -1); - if (info->fname) + if (info->fd == -1 && info->fname) { - if (wflag) - info->fd = open (info->fname, O_RDWR|O_CREAT, 0666); + if (flag & RSETF_WRITE) + info->fd = open (info->fname, O_BINARY|O_RDWR|O_CREAT, 0666); else - info->fd = open (info->fname, O_RDONLY); + info->fd = open (info->fname, O_BINARY|O_RDONLY); if (info->fd == -1) { logf (LOG_FATAL|LOG_ERRNO, "open %s", info->fname); exit (1); } } - rfd = xmalloc (sizeof(*rfd)); + rfd = (struct rset_temp_rfd *) xmalloc (sizeof(*rfd)); + rfd->next = info->rfd_list; + info->rfd_list = rfd; rfd->info = info; - r_rewind (ct); + r_rewind (rfd); + + rfd->countp = &ct->rset_terms[0]->count; + *rfd->countp = 0; + rfd->buf = xmalloc (info->key_size); + return rfd; } +/* r_flush: + flush current window to file if file is assocated with set + */ static void r_flush (RSFD rfd, int mk) { struct rset_temp_info *info = ((struct rset_temp_rfd*) rfd)->info; if (!info->fname && mk) { - char *s = (char*) tempnam (NULL, "zrs"); +#if HAVE_MKSTEMP + char template[1024]; + + if (info->temp_path) + sprintf (template, "%s/zrsXXXXXX", info->temp_path); + else + sprintf (template, "zrsXXXXXX"); - info->fname = xmalloc (strlen(s)+1); + info->fd = mkstemp (template); + + if (info->fd == -1) + { + logf (LOG_FATAL|LOG_ERRNO, "mkstemp %s", template); + exit (1); + } + info->fname = (char *) xmalloc (strlen(template)+1); + strcpy (info->fname, template); +#else + char *s = (char*) tempnam (info->temp_path, "zrs"); + info->fname = (char *) xmalloc (strlen(s)+1); strcpy (info->fname, s); - info->fd = open (info->fname, O_RDWR|O_CREAT, 0666); + logf (LOG_DEBUG, "creating tempfile %s", info->fname); + info->fd = open (info->fname, O_BINARY|O_RDWR|O_CREAT, 0666); if (info->fd == -1) { logf (LOG_FATAL|LOG_ERRNO, "open %s", info->fname); exit (1); } +#endif } if (info->fname && info->fd != -1 && info->dirty) { - size_t r, count; + size_t count; + int r; if (lseek (info->fd, info->pos_buf, SEEK_SET) == -1) { @@ -163,7 +200,7 @@ static void r_flush (RSFD rfd, int mk) count = info->buf_size; if (count > info->pos_end - info->pos_buf) count = info->pos_end - info->pos_buf; - if ((r = write (info->fd, info->buf_mem, count)) < count) + if ((r = write (info->fd, info->buf_mem, count)) < (int) count) { if (r == -1) logf (LOG_FATAL|LOG_ERRNO, "read %s", info->fname); @@ -179,41 +216,75 @@ static void r_flush (RSFD rfd, int mk) static void r_close (RSFD rfd) { struct rset_temp_info *info = ((struct rset_temp_rfd*)rfd)->info; + struct rset_temp_rfd **rfdp; - r_flush (rfd, 0); - if (info->fname && info->fd != -1) - { - close (info->fd); - info->fd = -1; - } + for (rfdp = &info->rfd_list; *rfdp; rfdp = &(*rfdp)->next) + if (*rfdp == rfd) + { + r_flush (*rfdp, 0); + xfree ((*rfdp)->buf); + + *rfdp = (*rfdp)->next; + xfree (rfd); + + if (!info->rfd_list && info->fname && info->fd != -1) + { + close (info->fd); + info->fd = -1; + } + return; + } + logf (LOG_FATAL, "r_close but no rfd match!"); + assert (0); } -static void r_delete (struct rset_control *ct) +static void r_delete (RSET ct) { - struct rset_temp_info *info = ct->buf; + struct rset_temp_info *info = (struct rset_temp_info*) ct->buf; - r_close (ct); if (info->fname) unlink (info->fname); - free (info->buf_mem); - free (info->fname); - free (info); + xfree (info->buf_mem); + logf (LOG_DEBUG, "r_delete: set size %ld", (long) info->pos_end); + if (info->fname) + { + logf (LOG_DEBUG, "r_delete: unlink %s", info->fname); + unlink (info->fname); + xfree (info->fname); + } + if (info->temp_path) + xfree (info->temp_path); + rset_term_destroy (ct->rset_terms[0]); + xfree (ct->rset_terms); + xfree (info); } +/* r_reread: + read from file to window if file is assocated with set - + indicated by fname + */ static void r_reread (RSFD rfd) { struct rset_temp_info *info = ((struct rset_temp_rfd*)rfd)->info; if (info->fname) { - size_t r, count; + size_t count; + int r; - info->pos_border = info->pos_cur + info->buf_size; + info->pos_border = ((struct rset_temp_rfd *)rfd)->pos_cur + + info->buf_size; if (info->pos_border > info->pos_end) info->pos_border = info->pos_end; count = info->pos_border - info->pos_buf; if (count > 0) - if ((r = read (info->fd, info->buf_mem, count)) < count) + { + if (lseek (info->fd, info->pos_buf, SEEK_SET) == -1) + { + logf (LOG_FATAL|LOG_ERRNO, "lseek %s", info->fname); + exit (1); + } + if ((r = read (info->fd, info->buf_mem, count)) < (int) count) { if (r == -1) logf (LOG_FATAL|LOG_ERRNO, "read %s", info->fname); @@ -222,6 +293,7 @@ static void r_reread (RSFD rfd) (long) count, (long) r); exit (1); } + } } else info->pos_border = info->pos_end; @@ -232,54 +304,76 @@ static void r_rewind (RSFD rfd) struct rset_temp_info *info = ((struct rset_temp_rfd*)rfd)->info; r_flush (rfd, 0); - info->pos_cur = 0; + ((struct rset_temp_rfd *)rfd)->pos_cur = 0; info->pos_buf = 0; r_reread (rfd); + ((struct rset_temp_rfd *)rfd)->cur=0; } -static int r_count (struct rset_control *ct) +/* +static int r_count (RSET ct) { - struct rset_temp_info *info = ct->buf; + struct rset_temp_info *info = (struct rset_temp_info *) ct->buf; return info->pos_end / info->key_size; } - -static int r_read (RSFD rfd, void *buf) +*/ +static int r_read (RSFD rfd, void *buf, int *term_index) { - struct rset_temp_info *info = ((struct rset_temp_rfd*)rfd)->info; + struct rset_temp_rfd *mrfd = (struct rset_temp_rfd*) rfd; + struct rset_temp_info *info = mrfd->info; - size_t nc = info->pos_cur + info->key_size; + size_t nc = mrfd->pos_cur + info->key_size; - if (nc > info->pos_border) + if (mrfd->pos_cur < info->pos_buf || nc > info->pos_border) { if (nc > info->pos_end) return 0; r_flush (rfd, 0); - info->pos_buf = info->pos_cur; + info->pos_buf = mrfd->pos_cur; r_reread (rfd); } - memcpy (buf, info->buf_mem + (info->pos_cur - info->pos_buf), + memcpy (buf, info->buf_mem + (mrfd->pos_cur - info->pos_buf), info->key_size); - info->pos_cur = nc; + mrfd->pos_cur = nc; + *term_index = 0; + + if (*mrfd->countp == 0 || (*info->cmp)(buf, mrfd->buf) > 1) + { + memcpy (mrfd->buf, buf, mrfd->info->key_size); + (*mrfd->countp)++; + } + mrfd->cur++; return 1; } static int r_write (RSFD rfd, const void *buf) { - struct rset_temp_info *info = ((struct rset_temp_rfd*)rfd)->info; + struct rset_temp_rfd *mrfd = (struct rset_temp_rfd*) rfd; + struct rset_temp_info *info = mrfd->info; - size_t nc = info->pos_cur + info->key_size; + size_t nc = mrfd->pos_cur + info->key_size; if (nc > info->pos_buf + info->buf_size) { r_flush (rfd, 1); - info->pos_buf = info->pos_cur; - r_reread (rfd); + info->pos_buf = mrfd->pos_cur; + if (info->pos_buf < info->pos_end) + r_reread (rfd); } - memcpy (info->buf_mem + (info->pos_cur - info->pos_buf), buf, - info->key_size); info->dirty = 1; - info->pos_border = info->pos_cur = nc; + memcpy (info->buf_mem + (mrfd->pos_cur - info->pos_buf), buf, + info->key_size); + mrfd->pos_cur = nc; + if (nc > info->pos_end) + info->pos_border = info->pos_end = nc; + info->hits++; return 1; } +static void r_pos (RSFD rfd, double *current, double *total) +{ + struct rset_temp_rfd *mrfd = (struct rset_temp_rfd*) rfd; + *current=(double) mrfd->cur; + *total=(double) mrfd->info->hits; +}