src/backend/tsearch/ts_utils.c

   1 /*-------------------------------------------------------------------------
   2  *
   3  * ts_utils.c
   4  *              various support functions
   5  *
   6  * Portions Copyright (c) 1996-2008, PostgreSQL Global Development Group
   7  *
   8  *
   9  * IDENTIFICATION
  10  *        $PostgreSQL$
  11  *
  12  *-------------------------------------------------------------------------
  13  */
  14
  15 #include "postgres.h"
  16
  17 #include <ctype.h>
  18
  19 #include "miscadmin.h"
  20 #include "tsearch/ts_locale.h"
  21 #include "tsearch/ts_public.h"
  22 #include "tsearch/ts_utils.h"
  23 #include "utils/builtins.h"
  24
  25
  26 /*
  27  * Given the base name and extension of a tsearch config file, return
  28  * its full path name.  The base name is assumed to be user-supplied,
  29  * and is checked to prevent pathname attacks.  The extension is assumed
  30  * to be safe.
  31  *
  32  * The result is a palloc'd string.
  33  */
  34 char *
  35 get_tsearch_config_filename(const char *basename,
  36                                                         const char *extension)
  37 {
  38         char            sharepath[MAXPGPATH];
  39         char       *result;
  40
  41         /*
  42          * We limit the basename to contain a-z, 0-9, and underscores.  This may
  43          * be overly restrictive, but we don't want to allow access to anything
  44          * outside the tsearch_data directory, so for instance '/' *must* be
  45          * rejected, and on some platforms '\' and ':' are risky as well. Allowing
  46          * uppercase might result in incompatible behavior between case-sensitive
  47          * and case-insensitive filesystems, and non-ASCII characters create other
  48          * interesting risks, so on the whole a tight policy seems best.
  49          */
  50         if (strspn(basename, "abcdefghijklmnopqrstuvwxyz0123456789_") != strlen(basename))
  51                 ereport(ERROR,
  52                                 (errcode(ERRCODE_INVALID_PARAMETER_VALUE),
  53                                  errmsg("invalid text search configuration file name \"%s\"",
  54                                                 basename)));
  55
  56         get_share_path(my_exec_path, sharepath);
  57         result = palloc(MAXPGPATH);
  58         snprintf(result, MAXPGPATH, "%s/tsearch_data/%s.%s",
  59                          sharepath, basename, extension);
  60
  61         return result;
  62 }
  63
  64 static int
  65 comparestr(const void *a, const void *b)
  66 {
  67         return strcmp(*(char **) a, *(char **) b);
  68 }
  69
  70 /*
  71  * Reads a stop-word file. Each word is run through 'wordop'
  72  * function, if given.  wordop may either modify the input in-place,
  73  * or palloc a new version.
  74  */
  75 void
  76 readstoplist(const char *fname, StopList *s, char *(*wordop) (const char *))
  77 {
  78         char      **stop = NULL;
  79
  80         s->len = 0;
  81         if (fname && *fname)
  82         {
  83                 char       *filename = get_tsearch_config_filename(fname, "stop");
  84                 tsearch_readline_state trst;
  85                 char       *line;
  86                 int                     reallen = 0;
  87
  88                 if (!tsearch_readline_begin(&trst, filename))
  89                         ereport(ERROR,
  90                                         (errcode(ERRCODE_CONFIG_FILE_ERROR),
  91                                          errmsg("could not open stop-word file \"%s\": %m",
  92                                                         filename)));
  93
  94                 while ((line = tsearch_readline(&trst)) != NULL)
  95                 {
  96                         char       *pbuf = line;
  97
  98                         /* Trim trailing space */
  99                         while (*pbuf && !t_isspace(pbuf))
 100                                 pbuf += pg_mblen(pbuf);
 101                         *pbuf = '\0';
 102
 103                         /* Skip empty lines */
 104                         if (*line == '\0')
 105                         {
 106                                 pfree(line);
 107                                 continue;
 108                         }
 109
 110                         if (s->len >= reallen)
 111                         {
 112                                 if (reallen == 0)
 113                                 {
 114                                         reallen = 64;
 115                                         stop = (char **) palloc(sizeof(char *) * reallen);
 116                                 }
 117                                 else
 118                                 {
 119                                         reallen *= 2;
 120                                         stop = (char **) repalloc((void *) stop,
 121                                                                                           sizeof(char *) * reallen);
 122                                 }
 123                         }
 124
 125                         if (wordop)
 126                         {
 127                                 stop[s->len] = wordop(line);
 128                                 if (stop[s->len] != line)
 129                                         pfree(line);
 130                         }
 131                         else
 132                                 stop[s->len] = line;
 133
 134                         (s->len)++;
 135                 }
 136
 137                 tsearch_readline_end(&trst);
 138                 pfree(filename);
 139         }
 140
 141         s->stop = stop;
 142
 143         /* Sort to allow binary searching */
 144         if (s->stop && s->len > 0)
 145                 qsort(s->stop, s->len, sizeof(char *), comparestr);
 146 }
 147
 148 bool
 149 searchstoplist(StopList *s, char *key)
 150 {
 151         return (s->stop && s->len > 0 &&
 152                         bsearch(&key, s->stop, s->len,
 153                                         sizeof(char *), comparestr)) ? true : false;
 154 }