diff options
-rw-r--r-- | ChangeLog | 7 | ||||
-rw-r--r-- | enc/gb18030.c | 15 |
2 files changed, 21 insertions, 1 deletions
@@ -1,3 +1,10 @@ +Mon Jun 2 03:23:25 2008 NARUSE, Yui <naruse@ruby-lang.org> + + * enc/gb18030.c (gb18030_mbc_to_code): mask by 0x7FFFFFFF + because OnigCodePoint will be used as 32bit signed int. + Masking by 0x7FFFFFFF is ok on GB18030; + Minumum 4bytes character is 0x81308130. + Sun Jun 1 22:29:35 2008 NARUSE, Yui <naruse@ruby-lang.org> * rational.c (string_to_r_internal): use rb_isdigit. diff --git a/enc/gb18030.c b/enc/gb18030.c index 231b961188..f62a2394c9 100644 --- a/enc/gb18030.c +++ b/enc/gb18030.c @@ -166,7 +166,20 @@ gb18030_mbc_enc_len(const UChar* p, const UChar* e, OnigEncoding enc ARG_UNUSED) static OnigCodePoint gb18030_mbc_to_code(const UChar* p, const UChar* end, OnigEncoding enc) { - return onigenc_mbn_mbc_to_code(enc, p, end); + int c, i, len; + OnigCodePoint n; + + len = enclen(enc, p, end); + n = (OnigCodePoint )(*p++); + if (len == 1) return n; + + for (i = 1; i < len; i++) { + if (p >= end) break; + c = *p++; + n <<= 8; n += c; + } + n &= 0x7FFFFFFF; + return n; } static int |