Line data    Source code

       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    24131274 : pg_strcasecmp(const char *s1, const char *s2)
      37             : {
      38             :     for (;;)
      39     6873422 :     {
      40    24131274 :         unsigned char ch1 = (unsigned char) *s1++;
      41    24131274 :         unsigned char ch2 = (unsigned char) *s2++;
      42             : 
      43    24131274 :         if (ch1 != ch2)
      44             :         {
      45    16929988 :             if (ch1 >= 'A' && ch1 <= 'Z')
      46     6254990 :                 ch1 += 'a' - 'A';
      47    10674998 :             else if (IS_HIGHBIT_SET(ch1) && isupper(ch1))
      48           0 :                 ch1 = tolower(ch1);
      49             : 
      50    16929988 :             if (ch2 >= 'A' && ch2 <= 'Z')
      51     3432768 :                 ch2 += 'a' - 'A';
      52    13497220 :             else if (IS_HIGHBIT_SET(ch2) && isupper(ch2))
      53           0 :                 ch2 = tolower(ch2);
      54             : 
      55    16929988 :             if (ch1 != ch2)
      56    15548870 :                 return (int) ch1 - (int) ch2;
      57             :         }
      58     8582404 :         if (ch1 == 0)
      59     1708982 :             break;
      60             :     }
      61     1708982 :     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     9623282 : pg_strncasecmp(const char *s1, const char *s2, size_t n)
      70             : {
      71    13834718 :     while (n-- > 0)
      72             :     {
      73    10616318 :         unsigned char ch1 = (unsigned char) *s1++;
      74    10616318 :         unsigned char ch2 = (unsigned char) *s2++;
      75             : 
      76    10616318 :         if (ch1 != ch2)
      77             :         {
      78     6481718 :             if (ch1 >= 'A' && ch1 <= 'Z')
      79     4771740 :                 ch1 += 'a' - 'A';
      80     1709978 :             else if (IS_HIGHBIT_SET(ch1) && isupper(ch1))
      81           0 :                 ch1 = tolower(ch1);
      82             : 
      83     6481718 :             if (ch2 >= 'A' && ch2 <= 'Z')
      84      181432 :                 ch2 += 'a' - 'A';
      85     6300286 :             else if (IS_HIGHBIT_SET(ch2) && isupper(ch2))
      86           0 :                 ch2 = tolower(ch2);
      87             : 
      88     6481718 :             if (ch1 != ch2)
      89     6404882 :                 return (int) ch1 - (int) ch2;
      90             :         }
      91     4211436 :         if (ch1 == 0)
      92           0 :             break;
      93             :     }
      94     3218400 :     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      187122 : pg_toupper(unsigned char ch)
     106             : {
     107      187122 :     if (ch >= 'a' && ch <= 'z')
     108       56394 :         ch += 'A' - 'a';
     109      130728 :     else if (IS_HIGHBIT_SET(ch) && islower(ch))
     110           0 :         ch = toupper(ch);
     111      187122 :     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    11162402 : pg_tolower(unsigned char ch)
     123             : {
     124    11162402 :     if (ch >= 'A' && ch <= 'Z')
     125     8483782 :         ch += 'a' - 'A';
     126     2678620 :     else if (IS_HIGHBIT_SET(ch) && isupper(ch))
     127           0 :         ch = tolower(ch);
     128    11162402 :     return ch;
     129             : }
     130             : 
     131             : /*
     132             :  * Fold a character to upper case, following C/POSIX locale rules.
     133             :  */
     134             : unsigned char
     135      153500 : pg_ascii_toupper(unsigned char ch)
     136             : {
     137      153500 :     if (ch >= 'a' && ch <= 'z')
     138      114632 :         ch += 'A' - 'a';
     139      153500 :     return ch;
     140             : }
     141             : 
     142             : /*
     143             :  * Fold a character to lower case, following C/POSIX locale rules.
     144             :  */
     145             : unsigned char
     146      384490 : pg_ascii_tolower(unsigned char ch)
     147             : {
     148      384490 :     if (ch >= 'A' && ch <= 'Z')
     149        6738 :         ch += 'a' - 'A';
     150      384490 :     return ch;
     151             : }