From 4216de6a3336cbc6dddb572cb7e6ab6193bf3729 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?J=C3=B6rg=20Frings-F=C3=BCrst?= Date: Fri, 29 Nov 2019 11:26:35 +0100 Subject: New upstream version 6.9.4 --- src/gb18030.c | 28 +++++++++++++++++----------- 1 file changed, 17 insertions(+), 11 deletions(-) (limited to 'src/gb18030.c') diff --git a/src/gb18030.c b/src/gb18030.c index 8d415b0..50898eb 100644 --- a/src/gb18030.c +++ b/src/gb18030.c @@ -3,7 +3,7 @@ **********************************************************************/ /*- * Copyright (c) 2005-2019 KUBO Takehiro - * K.Kosako + * K.Kosako * All rights reserved. * * Redistribution and use in source and binary forms, with or without @@ -33,6 +33,7 @@ #if 1 #define DEBUG_GB18030(arg) #else +#include #define DEBUG_GB18030(arg) printf arg #endif @@ -75,6 +76,20 @@ gb18030_mbc_enc_len(const UChar* p) return 2; } +static int +gb18030_code_to_mbclen(OnigCodePoint code) +{ + if ((code & 0xff000000) != 0) return 4; + else if ((code & 0xff0000) != 0) return ONIGERR_INVALID_CODE_POINT_VALUE; + else if ((code & 0xff00) != 0) return 2; + else { + if (GB18030_MAP[(int )(code & 0xff)] == CM) + return ONIGERR_INVALID_CODE_POINT_VALUE; + + return 1; + } +} + static int is_valid_mbc_string(const UChar* p, const UChar* end) { @@ -135,15 +150,6 @@ gb18030_mbc_case_fold(OnigCaseFoldType flag, const UChar** pp, const UChar* end, pp, end, lower); } -#if 0 -static int -gb18030_is_mbc_ambiguous(OnigCaseFoldType flag, - const UChar** pp, const UChar* end) -{ - return onigenc_mbn_is_mbc_ambiguous(ONIG_ENCODING_GB18030, flag, pp, end); -} -#endif - static int gb18030_is_code_ctype(OnigCodePoint code, unsigned int ctype) { @@ -522,7 +528,7 @@ OnigEncodingType OnigEncodingGB18030 = { 1, /* min enc length */ onigenc_is_mbc_newline_0x0a, gb18030_mbc_to_code, - onigenc_mb4_code_to_mbclen, + gb18030_code_to_mbclen, gb18030_code_to_mbc, gb18030_mbc_case_fold, onigenc_ascii_apply_all_case_fold, -- cgit v1.2.3