src/backend/parser/kwlookup.c

   1 /*-------------------------------------------------------------------------
   2  *
   3  * kwlookup.c
   4  *        lexical token lookup for key words in PostgreSQL
   5  *
   6  * NB - this file is also used by ECPG and several frontend programs in
   7  * src/bin/ including pg_dump and psql
   8  *
   9  * Note that this file expects that the ScanKeywords array is defined
  10  * and that LastScanKeyword points to its element one past the last.
  11  *
  12  * Portions Copyright (c) 1996-2009, PostgreSQL Global Development Group
  13  * Portions Copyright (c) 1994, Regents of the University of California
  14  *
  15  *
  16  * IDENTIFICATION
  17  *        $PostgreSQL$
  18  *
  19  *-------------------------------------------------------------------------
  20  */
  21
  22 /* use c.h so this can be built as either frontend or backend */
  23 #include "c.h"
  24
  25 #include <ctype.h>
  26
  27 #include "parser/keywords.h"
  28
  29 /*
  30  * ScanKeywordLookup - see if a given word is a keyword
  31  *
  32  * Returns a pointer to the ScanKeyword table entry, or NULL if no match.
  33  *
  34  * The match is done case-insensitively.  Note that we deliberately use a
  35  * dumbed-down case conversion that will only translate 'A'-'Z' into 'a'-'z',
  36  * even if we are in a locale where tolower() would produce more or different
  37  * translations.  This is to conform to the SQL99 spec, which says that
  38  * keywords are to be matched in this way even though non-keyword identifiers
  39  * receive a different case-normalization mapping.
  40  */
  41 const ScanKeyword *
  42 ScanKeywordLookup(const char *text)
  43 {
  44         int                     len,
  45                                 i;
  46         char            word[NAMEDATALEN];
  47         const ScanKeyword *low;
  48         const ScanKeyword *high;
  49
  50         len = strlen(text);
  51         /* We assume all keywords are shorter than NAMEDATALEN. */
  52         if (len >= NAMEDATALEN)
  53                 return NULL;
  54
  55         /*
  56          * Apply an ASCII-only downcasing.      We must not use tolower() since it may
  57          * produce the wrong translation in some locales (eg, Turkish).
  58          */
  59         for (i = 0; i < len; i++)
  60         {
  61                 char            ch = text[i];
  62
  63                 if (ch >= 'A' && ch <= 'Z')
  64                         ch += 'a' - 'A';
  65                 word[i] = ch;
  66         }
  67         word[len] = '\0';
  68
  69         /*
  70          * Now do a binary search using plain strcmp() comparison.
  71          */
  72         low = &ScanKeywords[0];
  73         high = LastScanKeyword - 1;
  74         while (low <= high)
  75         {
  76                 const ScanKeyword *middle;
  77                 int                     difference;
  78
  79                 middle = low + (high - low) / 2;
  80                 difference = strcmp(middle->name, word);
  81                 if (difference == 0)
  82                         return middle;
  83                 else if (difference < 0)
  84                         low = middle + 1;
  85                 else
  86                         high = middle - 1;
  87         }
  88
  89         return NULL;
  90 }