src/port/pgstrcasecmp.c

   1 /*-------------------------------------------------------------------------
   2  *
   3  * pgstrcasecmp.c
   4  *         Portable SQL-like case-independent comparisons and conversions.
   5  *
   6  * SQL99 specifies Unicode-aware case normalization, which we don't yet
   7  * have the infrastructure for.  Instead we use tolower() to provide a
   8  * locale-aware translation.  However, there are some locales where this
   9  * is not right either (eg, Turkish may do strange things with 'i' and
  10  * 'I').  Our current compromise is to use tolower() for characters with
  11  * the high bit set, and use an ASCII-only downcasing for 7-bit
  12  * characters.
  13  *
  14  * NB: this code should match downcase_truncate_identifier() in scansup.c.
  15  *
  16  * We also provide strict ASCII-only case conversion functions, which can
  17  * be used to implement C/POSIX case folding semantics no matter what the
  18  * C library thinks the locale is.
  19  *
  20  *
  21  * Portions Copyright (c) 1996-2024, PostgreSQL Global Development Group
  22  *
  23  * src/port/pgstrcasecmp.c
  24  *
  25  *-------------------------------------------------------------------------
  26  */
  27 #include "c.h"
  28
  29 #include <ctype.h>
  30
  31
  32 /*
  33  * Case-independent comparison of two null-terminated strings.
  34  */
  35 int
  36 pg_strcasecmp(const char *s1, const char *s2)
  37 {
  38         for (;;)
  39         {
  40                 unsigned char ch1 = (unsigned char) *s1++;
  41                 unsigned char ch2 = (unsigned char) *s2++;
  42
  43                 if (ch1 != ch2)
  44                 {
  45                         if (ch1 >= 'A' && ch1 <= 'Z')
  46                                 ch1 += 'a' - 'A';
  47                         else if (IS_HIGHBIT_SET(ch1) && isupper(ch1))
  48                                 ch1 = tolower(ch1);
  49
  50                         if (ch2 >= 'A' && ch2 <= 'Z')
  51                                 ch2 += 'a' - 'A';
  52                         else if (IS_HIGHBIT_SET(ch2) && isupper(ch2))
  53                                 ch2 = tolower(ch2);
  54
  55                         if (ch1 != ch2)
  56                                 return (int) ch1 - (int) ch2;
  57                 }
  58                 if (ch1 == 0)
  59                         break;
  60         }
  61         return 0;
  62 }
  63
  64 /*
  65  * Case-independent comparison of two not-necessarily-null-terminated strings.
  66  * At most n bytes will be examined from each string.
  67  */
  68 int
  69 pg_strncasecmp(const char *s1, const char *s2, size_t n)
  70 {
  71         while (n-- > 0)
  72         {
  73                 unsigned char ch1 = (unsigned char) *s1++;
  74                 unsigned char ch2 = (unsigned char) *s2++;
  75
  76                 if (ch1 != ch2)
  77                 {
  78                         if (ch1 >= 'A' && ch1 <= 'Z')
  79                                 ch1 += 'a' - 'A';
  80                         else if (IS_HIGHBIT_SET(ch1) && isupper(ch1))
  81                                 ch1 = tolower(ch1);
  82
  83                         if (ch2 >= 'A' && ch2 <= 'Z')
  84                                 ch2 += 'a' - 'A';
  85                         else if (IS_HIGHBIT_SET(ch2) && isupper(ch2))
  86                                 ch2 = tolower(ch2);
  87
  88                         if (ch1 != ch2)
  89                                 return (int) ch1 - (int) ch2;
  90                 }
  91                 if (ch1 == 0)
  92                         break;
  93         }
  94         return 0;
  95 }
  96
  97 /*
  98  * Fold a character to upper case.
  99  *
 100  * Unlike some versions of toupper(), this is safe to apply to characters
 101  * that aren't lower case letters.  Note however that the whole thing is
 102  * a bit bogus for multibyte character sets.
 103  */
 104 unsigned char
 105 pg_toupper(unsigned char ch)
 106 {
 107         if (ch >= 'a' && ch <= 'z')
 108                 ch += 'A' - 'a';
 109         else if (IS_HIGHBIT_SET(ch) && islower(ch))
 110                 ch = toupper(ch);
 111         return ch;
 112 }
 113
 114 /*
 115  * Fold a character to lower case.
 116  *
 117  * Unlike some versions of tolower(), this is safe to apply to characters
 118  * that aren't upper case letters.  Note however that the whole thing is
 119  * a bit bogus for multibyte character sets.
 120  */
 121 unsigned char
 122 pg_tolower(unsigned char ch)
 123 {
 124         if (ch >= 'A' && ch <= 'Z')
 125                 ch += 'a' - 'A';
 126         else if (IS_HIGHBIT_SET(ch) && isupper(ch))
 127                 ch = tolower(ch);
 128         return ch;
 129 }
 130
 131 /*
 132  * Fold a character to upper case, following C/POSIX locale rules.
 133  */
 134 unsigned char
 135 pg_ascii_toupper(unsigned char ch)
 136 {
 137         if (ch >= 'a' && ch <= 'z')
 138                 ch += 'A' - 'a';
 139         return ch;
 140 }
 141
 142 /*
 143  * Fold a character to lower case, following C/POSIX locale rules.
 144  */
 145 unsigned char
 146 pg_ascii_tolower(unsigned char ch)
 147 {
 148         if (ch >= 'A' && ch <= 'Z')
 149                 ch += 'a' - 'A';
 150         return ch;
 151 }