1 /*-------------------------------------------------------------------------
2 *
3 * kwlookup.c
4 * Key word lookup for PostgreSQL
5 *
6 *
7 * Portions Copyright (c) 2003-2020, PgPool Global Development Group
8 * Portions Copyright (c) 1996-2020, PostgreSQL Global Development Group
9 * Portions Copyright (c) 1994, Regents of the University of California
10 *
11 *
12 * IDENTIFICATION
13 * src/common/kwlookup.c
14 *
15 *-------------------------------------------------------------------------
16 */
17
18 #include <string.h>
19 #include "kwlookup.h"
20
21
22 /*
23 * ScanKeywordLookup - see if a given word is a keyword
24 *
25 * The list of keywords to be matched against is passed as a ScanKeywordList.
26 *
27 * Returns the keyword number (0..N-1) of the keyword, or -1 if no match.
28 * Callers typically use the keyword number to index into information
29 * arrays, but that is no concern of this code.
30 *
31 * The match is done case-insensitively. Note that we deliberately use a
32 * dumbed-down case conversion that will only translate 'A'-'Z' into 'a'-'z',
33 * even if we are in a locale where tolower() would produce more or different
34 * translations. This is to conform to the SQL99 spec, which says that
35 * keywords are to be matched in this way even though non-keyword identifiers
36 * receive a different case-normalization mapping.
37 */
38 int
ScanKeywordLookup(const char * str,const ScanKeywordList * keywords)39 ScanKeywordLookup(const char *str,
40 const ScanKeywordList *keywords)
41 {
42 size_t len;
43 int h;
44 const char *kw;
45
46 /*
47 * Reject immediately if too long to be any keyword. This saves useless
48 * hashing and downcasing work on long strings.
49 */
50 len = strlen(str);
51 if (len > keywords->max_kw_len)
52 return -1;
53
54 /*
55 * Compute the hash function. We assume it was generated to produce
56 * case-insensitive results. Since it's a perfect hash, we need only
57 * match to the specific keyword it identifies.
58 */
59 h = keywords->hash(str, len);
60
61 /* An out-of-range result implies no match */
62 if (h < 0 || h >= keywords->num_keywords)
63 return -1;
64
65 /*
66 * Compare character-by-character to see if we have a match, applying an
67 * ASCII-only downcasing to the input characters. We must not use
68 * tolower() since it may produce the wrong translation in some locales
69 * (eg, Turkish).
70 */
71 kw = GetScanKeyword(h, keywords);
72 while (*str != '\0')
73 {
74 char ch = *str++;
75
76 if (ch >= 'A' && ch <= 'Z')
77 ch += 'a' - 'A';
78 if (ch != *kw++)
79 return -1;
80 }
81 if (*kw != '\0')
82 return -1;
83
84 /* Success! */
85 return h;
86 }
87