Gecko [ www.jdcpatala.org ]

Name	Size	Permission
bug_fix.c	2.3 KB	-rw-r--r--
callout.c	7.57 KB	-rw-r--r--
count.c	3.35 KB	-rw-r--r--
crnl.c	3.56 KB	-rw-r--r--
echo.c	3.18 KB	-rw-r--r--
encode.c	7.03 KB	-rw-r--r--
listcap.c	2.85 KB	-rw-r--r--
names.c	2.11 KB	-rw-r--r--
posix.c	2.52 KB	-rw-r--r--
regset.c	2.21 KB	-rw-r--r--
scan.c	1.94 KB	-rw-r--r--
simple.c	1.58 KB	-rw-r--r--
sql.c	2.44 KB	-rw-r--r--
syntax.c	1.99 KB	-rw-r--r--
user_property.c	2.41 KB	-rw-r--r--

Code Editor : encode.c

/*
 * encode.c
 */
#include <stdio.h>
#include "oniguruma.h"

static int
search(regex_t* reg, unsigned char* str, unsigned char* end)
{
  int r;
  unsigned char *start, *range;
  OnigRegion *region;

region = onig_region_new();

start = str;
  range = end;
  r = onig_search(reg, str, end, start, range, region, ONIG_OPTION_NONE);
  if (r >= 0) {
    int i;

fprintf(stderr, "match at %d  (%s)\n", r,
            ONIGENC_NAME(onig_get_encoding(reg)));
    for (i = 0; i < region->num_regs; i++) {
      fprintf(stderr, "%d: (%d-%d)\n", i, region->beg[i], region->end[i]);
    }
  }
  else if (r == ONIG_MISMATCH) {
    fprintf(stderr, "search fail (%s)\n",
            ONIGENC_NAME(onig_get_encoding(reg)));
  }
  else { /* error */
    char s[ONIG_MAX_ERROR_MESSAGE_LEN];
    onig_error_code_to_str((UChar* )s, r);
    fprintf(stderr, "ERROR: %s\n", s);
    fprintf(stderr, "  (%s)\n", ONIGENC_NAME(onig_get_encoding(reg)));
    onig_region_free(region, 1 /* 1:free self, 0:free contents only */);
    return -1;
  }

onig_region_free(region, 1 /* 1:free self, 0:free contents only */);
  return 0;
}

static int
exec(OnigEncoding enc, OnigOptionType options,
     char* apattern, char* astr)
{
  int r;
  unsigned char *end;
  regex_t* reg;
  OnigErrorInfo einfo;
  UChar* pattern = (UChar* )apattern;
  UChar* str     = (UChar* )astr;

onig_initialize(&enc, 1);

r = onig_new(&reg, pattern,
               pattern + onigenc_str_bytelen_null(enc, pattern),
               options, enc, ONIG_SYNTAX_DEFAULT, &einfo);
  if (r != ONIG_NORMAL) {
    char s[ONIG_MAX_ERROR_MESSAGE_LEN];
    onig_error_code_to_str((UChar* )s, r, &einfo);
    fprintf(stderr, "ERROR: %s\n", s);
    return -1;
  }

end   = str + onigenc_str_bytelen_null(enc, str);
  r = search(reg, str, end);

onig_free(reg);
  onig_end();
  return 0;
}

extern int main(int argc, char* argv[])
{
  int r;
  /* ISO 8859-1 test */
  static unsigned char str[] = { 0xc7, 0xd6, 0xfe, 0xea, 0xe0, 0xe2, 0x00 };
  static unsigned char pattern[] = { 0xe7, 0xf6, 0xde, '\\', 'w', '+', 0x00 };

r = exec(ONIG_ENCODING_SJIS, ONIG_OPTION_NONE,
           "^a\\p{Hiragana}c$", "a\202\274c");

r = exec(ONIG_ENCODING_EUC_JP, ONIG_OPTION_NONE,
           "^a\\p{Hiragana}c$", "a\244\276c");

r = exec(ONIG_ENCODING_CP1251, ONIG_OPTION_IGNORECASE,
           "aBc", " AbC");

r = exec(ONIG_ENCODING_ISO_8859_1, ONIG_OPTION_IGNORECASE,
           " [a-c\337z] ", "  SS  ");
  r = exec(ONIG_ENCODING_ISO_8859_1, ONIG_OPTION_IGNORECASE,
           " [\330-\341] ", "  SS  ");

r = exec(ONIG_ENCODING_ISO_8859_2, ONIG_OPTION_IGNORECASE,
           "\337          ", "          Ss          ");
  r = exec(ONIG_ENCODING_ISO_8859_2, ONIG_OPTION_IGNORECASE,
           "SS          ", "          \337          ");
  r = exec(ONIG_ENCODING_ISO_8859_2, ONIG_OPTION_IGNORECASE,
           "\\A\\S\\z", "ss");

r = exec(ONIG_ENCODING_ISO_8859_2, ONIG_OPTION_IGNORECASE,
           "[ac]+", "bbbaAaCCC");

r = exec(ONIG_ENCODING_ISO_8859_3, ONIG_OPTION_IGNORECASE,
           "[ac]+", "bbbaAaCCC");
  r = exec(ONIG_ENCODING_ISO_8859_4, ONIG_OPTION_IGNORECASE,
           "[ac]+", "bbbaAaCCC");
  r = exec(ONIG_ENCODING_ISO_8859_5, ONIG_OPTION_IGNORECASE,
           "[ac]+", "bbbaAaCCC");
  r = exec(ONIG_ENCODING_ISO_8859_6, ONIG_OPTION_IGNORECASE,
           "[ac]+", "bbbaAaCCC");
  r = exec(ONIG_ENCODING_ISO_8859_7, ONIG_OPTION_IGNORECASE,
           "[ac]+", "bbbaAaCCC");
  r = exec(ONIG_ENCODING_ISO_8859_8, ONIG_OPTION_IGNORECASE,
           "[ac]+", "bbbaAaCCC");
  r = exec(ONIG_ENCODING_ISO_8859_9, ONIG_OPTION_IGNORECASE,
           "[ac]+", "bbbaAaCCC");
  r = exec(ONIG_ENCODING_ISO_8859_10, ONIG_OPTION_IGNORECASE,
           "[ac]+", "bbbaAaCCC");
  r = exec(ONIG_ENCODING_ISO_8859_11, ONIG_OPTION_IGNORECASE,
           "[ac]+", "bbbaAaCCC");
  r = exec(ONIG_ENCODING_ISO_8859_13, ONIG_OPTION_IGNORECASE,
           "[ac]+", "bbbaAaCCC");
  r = exec(ONIG_ENCODING_ISO_8859_14, ONIG_OPTION_IGNORECASE,
           "[ac]+", "bbbaAaCCC");
  r = exec(ONIG_ENCODING_ISO_8859_15, ONIG_OPTION_IGNORECASE,
           (char* )pattern, (char* )str);
  r = exec(ONIG_ENCODING_ISO_8859_16, ONIG_OPTION_IGNORECASE,
           (char* )pattern, (char* )str);

r = exec(ONIG_ENCODING_KOI8_R, ONIG_OPTION_NONE, "a+", "bbbaaaccc");
  r = exec(ONIG_ENCODING_EUC_TW, ONIG_OPTION_NONE, "b*a+?c+", "bbbaaaccc");
  r = exec(ONIG_ENCODING_EUC_KR, ONIG_OPTION_NONE, "a+", "bbbaaaccc");
  r = exec(ONIG_ENCODING_EUC_CN, ONIG_OPTION_NONE, "c+", "bbbaaaccc");
  r = exec(ONIG_ENCODING_BIG5,   ONIG_OPTION_NONE, "a+", "bbbaaaccc");

r = exec(ONIG_ENCODING_ISO_8859_1, ONIG_OPTION_IGNORECASE,
           "\337", "SS");
  r = exec(ONIG_ENCODING_ISO_8859_1, ONIG_OPTION_IGNORECASE,
           "SS", "\337");
  r = exec(ONIG_ENCODING_ISO_8859_1, ONIG_OPTION_IGNORECASE,
           "SSb\337ssc", "a\337bSS\337cd");
  r = exec(ONIG_ENCODING_ISO_8859_1, ONIG_OPTION_IGNORECASE,
           "[a\337]{0,2}", "aSS");
  r = exec(ONIG_ENCODING_ISO_8859_1, ONIG_OPTION_IGNORECASE,
           "is", "iss");

r = exec(ONIG_ENCODING_UTF16_BE, ONIG_OPTION_NONE,
           "\000[\000[\000:\000a\000l\000n\000u\000m\000:\000]\000]\000+\000\000",
           "\000#\002\120\000a\000Z\012\077\012\076\012\075\000\000");
  /* 0x0a3d == \012\075 : is not alnum */
  /* 0x0a3e == \012\076 : is alnum */

r = exec(ONIG_ENCODING_UTF16_BE, ONIG_OPTION_NONE,
           "\000\\\000d\000+\000\000",
           "\0003\0001\377\020\377\031\377\032\000\000");

r = exec(ONIG_ENCODING_GB18030, ONIG_OPTION_IGNORECASE,
           "(Aa\\d)+", "BaA5Aa0234");

r = exec(ONIG_ENCODING_UTF16_BE, ONIG_OPTION_IGNORECASE,
           "\000[\000\337\000]\000\000", "\000S\000S\000\000");

r = exec(ONIG_ENCODING_UTF16_BE, ONIG_OPTION_IGNORECASE,
           "\000[\000\337\000]\000\000", "\000s\000S\000\000");

r = exec(ONIG_ENCODING_UTF16_BE, ONIG_OPTION_IGNORECASE,
           "\000^\000[\000\001\000-\377\375\000]\000$\000\000",
           "\000s\000S\000\000");

r = exec(ONIG_ENCODING_UTF16_BE, ONIG_OPTION_IGNORECASE,
           "\000S\000S\000\000",
           "\000S\000T\000\337\000\000");

r = exec(ONIG_ENCODING_UTF16_BE, ONIG_OPTION_IGNORECASE,
           "\000S\000T\000S\000S\000\000",
           "\000S\000t\000s\000S\000\000");

{
    UChar pat[]  = { 0x1f, 0xfc, 0x00, 0x00 };
    UChar str1[] = { 0x21, 0x26, 0x1f, 0xbe, 0x00, 0x00 };
    UChar str2[] = { 0x1f, 0xf3, 0x00, 0x00 };

r = exec(ONIG_ENCODING_UTF16_BE, ONIG_OPTION_IGNORECASE,
             (char* )pat, (char* )str1);

r = exec(ONIG_ENCODING_UTF16_BE, ONIG_OPTION_IGNORECASE,
             (char* )pat, (char* )str2);
  }

#if 0
  /* You should define USE_UNICODE_CASE_FOLD_TURKISH_AZERI in regenc.h. */

set_case_fold(ONIGENC_CASE_FOLD_TURKISH_AZERI);

r = exec(ONIG_ENCODING_UTF8, ONIG_ENCODING_UTF8, ONIG_OPTION_IGNORECASE,
           "Ii", "\304\261\304\260");

r = exec(ONIG_ENCODING_UTF16_BE, ONIG_OPTION_IGNORECASE,
           "\000I\000i\000\000", "\001\061\001\060\000\000");

r = exec(ONIG_ENCODING_UTF16_BE, ONIG_OPTION_IGNORECASE,
           "\001\061\001\060\000\000", "\000I\000i\000\000");

set_case_fold(ONIGENC_CASE_FOLD_MIN);
#endif

return r;
}

${this.title}

Code Editor : encode.c