• branch=trunk inherited from [704b122e53]
• sym-trunk inherited from [704b122e53]

Changes to src/test_regexp.c

   22  **     X|Y     X or Y                                                                 22  **     X|Y     X or Y
   23  **     ^X      X occurring at the beginning of the string                             23  **     ^X      X occurring at the beginning of the string
   24  **     X$      X occurring at the end of the string                                   24  **     X$      X occurring at the end of the string
   25  **     .       Match any single character                                             25  **     .       Match any single character
   26  **     \c      Character c where c is one of \{}()[]|*+?.                             26  **     \c      Character c where c is one of \{}()[]|*+?.
   27  **     \c      C-language escapes for c in afnrtv.  ex: \t or \n                      27  **     \c      C-language escapes for c in afnrtv.  ex: \t or \n
   28  **     \uXXXX  Where XXXX is exactly 4 hex digits, unicode value XXXX                 28  **     \uXXXX  Where XXXX is exactly 4 hex digits, unicode value XXXX
   29  **     \xXXX   Where XXX is any number of hex digits, unicode value XXX          |    29  **     \xXX    Where XX is exactly 2 hex digits, unicode value XX
   30  **     [abc]   Any single character from the set abc                                  30  **     [abc]   Any single character from the set abc
   31  **     [^abc]  Any single character not in the set abc                                31  **     [^abc]  Any single character not in the set abc
   32  **     [a-z]   Any single character in the range a-z                                  32  **     [a-z]   Any single character in the range a-z
   33  **     [^a-z]  Any single character not in the range a-z                              33  **     [^a-z]  Any single character not in the range a-z
   34  **     \b      Word boundary                                                          34  **     \b      Word boundary
   35  **     \w      Word character.  [A-Za-z0-9_]                                          35  **     \w      Word character.  [A-Za-z0-9_]
   36  **     \W      Non-word character                                                     36  **     \W      Non-word character
................................................................................................................................................................................

  383  static unsigned re_esc_char(ReCompiled *p){                                          383  static unsigned re_esc_char(ReCompiled *p){
  384    static const char zEsc[] = "afnrtv\\()*.+?[$^{|}]";                                384    static const char zEsc[] = "afnrtv\\()*.+?[$^{|}]";
  385    static const char zTrans[] = "\a\f\n\r\t\v";                                       385    static const char zTrans[] = "\a\f\n\r\t\v";
  386    int i, v = 0;                                                                      386    int i, v = 0;
  387    char c;                                                                            387    char c;
  388    if( p->sIn.i>=p->sIn.mx ) return 0;                                                388    if( p->sIn.i>=p->sIn.mx ) return 0;
  389    c = p->sIn.z[p->sIn.i];                                                            389    c = p->sIn.z[p->sIn.i];
  390    if( c=='u' && p->sIn.i+5<p->sIn.mx ){                                          |   390    if( c=='u' && p->sIn.i+4<p->sIn.mx ){
  391      const unsigned char *zIn = p->sIn.z + p->sIn.i;                                  391      const unsigned char *zIn = p->sIn.z + p->sIn.i;
  392      v = 0;                                                                       <
  393      if( re_hex(zIn[1],&v)                                                            392      if( re_hex(zIn[1],&v)
  394       && re_hex(zIn[2],&v)                                                            393       && re_hex(zIn[2],&v)
  395       && re_hex(zIn[3],&v)                                                            394       && re_hex(zIn[3],&v)
  396       && re_hex(zIn[4],&v)                                                            395       && re_hex(zIn[4],&v)
  397      ){                                                                               396      ){
  398        p->sIn.i += 5;                                                                 397        p->sIn.i += 5;
  399        return v;                                                                      398        return v;
  400      }                                                                                399      }
  401    }                                                                                  400    }
  402    if( c=='x' ){                                                                  |   401    if( c=='x' && p->sIn.i+2<p->sIn.mx ){
  403      v = 0;                                                                       |   402      const unsigned char *zIn = p->sIn.z + p->sIn.i;
  404      for(i=1; p->sIn.i<p->sIn.mx && re_hex(p->sIn.z[p->sIn.i+i], &v); i++){}      |   403      if( re_hex(zIn[1],&v)
  405      if( i>1 ){                                                                   |   404       && re_hex(zIn[2],&v)
                                                                                        >   405      ){
  406        p->sIn.i += i;                                                             |   406        p->sIn.i += 3;
  407        return v;                                                                      407        return v;
  408      }                                                                                408      }
  409    }                                                                                  409    }
  410    for(i=0; zEsc[i] && zEsc[i]!=c; i++){}                                             410    for(i=0; zEsc[i] && zEsc[i]!=c; i++){}
  411    if( zEsc[i] ){                                                                     411    if( zEsc[i] ){
  412      if( i<6 ) c = zTrans[i];                                                         412      if( i<6 ) c = zTrans[i];
  413      p->sIn.i++;                                                                      413      p->sIn.i++;

Changes to test/regexp1.test

  193           'abc{4}x' REGEXP '^abc\{4\}x$',                                             193           'abc{4}x' REGEXP '^abc\{4\}x$',
  194           'abc|def' REGEXP '^abc\|def$'                                               194           'abc|def' REGEXP '^abc\|def$'
  195  } {1 1 1 1 1 1 1 1 1 1 1 1}                                                          195  } {1 1 1 1 1 1 1 1 1 1 1 1}
  196                                                                                       196  
  197  do_execsql_test regexp1-2.20 {                                                       197  do_execsql_test regexp1-2.20 {
  198    SELECT 'abc$¢€xyz' REGEXP '^abc\u0024\u00a2\u20acxyz$',                            198    SELECT 'abc$¢€xyz' REGEXP '^abc\u0024\u00a2\u20acxyz$',
  199           'abc$¢€xyz' REGEXP '^abc\u0024\u00A2\u20ACxyz$',                            199           'abc$¢€xyz' REGEXP '^abc\u0024\u00A2\u20ACxyz$',
  200           'abc$¢€xyz' REGEXP '^abc\x24\xa2\x20acxyz$'                             |   200           'abc$¢€xyz' REGEXP '^abc\x24\xa2\u20acxyz$'
  201  } {1 1 1}                                                                            201  } {1 1 1}
  202  do_execsql_test regexp1-2.21 {                                                       202  do_execsql_test regexp1-2.21 {
  203    SELECT 'abc$¢€xyz' REGEXP '^abc[\u0024][\u00a2][\u20ac]xyz$',                      203    SELECT 'abc$¢€xyz' REGEXP '^abc[\u0024][\u00a2][\u20ac]xyz$',
  204           'abc$¢€xyz' REGEXP '^abc[\u0024\u00A2\u20AC]{3}xyz$',                       204           'abc$¢€xyz' REGEXP '^abc[\u0024\u00A2\u20AC]{3}xyz$',
  205           'abc$¢€xyz' REGEXP '^abc[\x24][\xa2\x20ac]+xyz$'                        |   205           'abc$¢€xyz' REGEXP '^abc[\x24][\xa2\u20ac]+xyz$'
  206  } {1 1 1}                                                                            206  } {1 1 1}
  207  do_execsql_test regexp1-2.22 {                                                       207  do_execsql_test regexp1-2.22 {
  208    SELECT 'abc$¢€xyz' REGEXP '^abc[^\u0025-X][^ -\u007f][^\u20ab]xyz$'                208    SELECT 'abc$¢€xyz' REGEXP '^abc[^\u0025-X][^ -\u007f][^\u20ab]xyz$'
  209  } {1}                                                                                209  } {1}
  210                                                                                       210  
  211  finish_test                                                                          211  finish_test