Timerix/utf8proc
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

remove requirement that get_property and decompose_char argument be in range 0x0 to 0x10ffff

Browse Source
This commit is contained in:
Steven G. Johnson 2015-03-12 14:17:27 -04:00
parent 128c04e3d0
commit 3822984606
2 changed files with 14 additions and 14 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

24
utf8proc.c
View File

@ -182,7 +182,8 @@ DLLEXPORT ssize_t utf8proc_encode_char(int32_t uc, uint8_t *dst) {
} else return 0; } else return 0;
} }
DLLEXPORT const utf8proc_property_t *utf8proc_get_property(int32_t uc) { /* internal "unsafe" version that does not check whether uc is in range */
static const utf8proc_property_t *get_property(int32_t uc) {
/* ASSERT: uc >= 0 && uc < 0x110000 */ /* ASSERT: uc >= 0 && uc < 0x110000 */
return utf8proc_properties + ( return utf8proc_properties + (
utf8proc_stage2table[ utf8proc_stage2table[
@ -191,6 +192,10 @@ DLLEXPORT const utf8proc_property_t *utf8proc_get_property(int32_t uc) {
); );
} }
DLLEXPORT const utf8proc_property_t *utf8proc_get_property(int32_t uc) {
return uc < 0 || uc >= 0x110000 ? utf8proc_properties : get_property(uc);
}
/* return whether there is a grapheme break between boundclasses lbc and tbc */ /* return whether there is a grapheme break between boundclasses lbc and tbc */
static bool grapheme_break(int lbc, int tbc) { static bool grapheme_break(int lbc, int tbc) {
return return
@ -242,13 +247,12 @@ DLLEXPORT const char *utf8proc_category_string(int32_t c) {
return utf8proc_decompose_char((replacement_uc), dst, bufsize, \ return utf8proc_decompose_char((replacement_uc), dst, bufsize, \
options & ~UTF8PROC_LUMP, last_boundclass) options & ~UTF8PROC_LUMP, last_boundclass)
DLLEXPORT ssize_t utf8proc_decompose_char(int32_t uc, int32_t *dst, ssize_t bufsize, DLLEXPORT ssize_t utf8proc_decompose_char(int32_t uc, int32_t *dst, ssize_t bufsize, int options, int *last_boundclass) {
int options, int *last_boundclass) {
/* ASSERT: uc >= 0 && uc < 0x110000 */
const utf8proc_property_t *property; const utf8proc_property_t *property;
utf8proc_propval_t category; utf8proc_propval_t category;
int32_t hangul_sindex; int32_t hangul_sindex;
property = utf8proc_get_property(uc); if (uc < 0 || uc >= 0x110000) return UTF8PROC_ERROR_NOTASSIGNED;
property = get_property(uc);
category = property->category; category = property->category;
hangul_sindex = uc - UTF8PROC_HANGUL_SBASE; hangul_sindex = uc - UTF8PROC_HANGUL_SBASE;
if (options & (UTF8PROC_COMPOSE|UTF8PROC_DECOMPOSE)) { if (options & (UTF8PROC_COMPOSE|UTF8PROC_DECOMPOSE)) {
@ -394,8 +398,8 @@ DLLEXPORT ssize_t utf8proc_decompose(
const utf8proc_property_t *property1, *property2; const utf8proc_property_t *property1, *property2;
uc1 = buffer[pos]; uc1 = buffer[pos];
uc2 = buffer[pos+1]; uc2 = buffer[pos+1];
property1 = utf8proc_get_property(uc1); property1 = get_property(uc1);
property2 = utf8proc_get_property(uc2); property2 = get_property(uc2);
if (property1->combining_class > property2->combining_class && if (property1->combining_class > property2->combining_class &&
property2->combining_class > 0) { property2->combining_class > 0) {
buffer[pos] = uc2; buffer[pos] = uc2;
@ -453,7 +457,7 @@ DLLEXPORT ssize_t utf8proc_reencode(int32_t *buffer, ssize_t length, int options
int32_t composition; int32_t composition;
for (rpos = 0; rpos < length; rpos++) { for (rpos = 0; rpos < length; rpos++) {
current_char = buffer[rpos]; current_char = buffer[rpos];
current_property = utf8proc_get_property(current_char); current_property = get_property(current_char);
if (starter && current_property->combining_class > max_combining_class) { if (starter && current_property->combining_class > max_combining_class) {
/* combination perhaps possible */ /* combination perhaps possible */
int32_t hangul_lindex; int32_t hangul_lindex;
@ -482,7 +486,7 @@ DLLEXPORT ssize_t utf8proc_reencode(int32_t *buffer, ssize_t length, int options
} }
} }
if (!starter_property) { if (!starter_property) {
starter_property = utf8proc_get_property(*starter); starter_property = get_property(*starter);
} }
if (starter_property->comb1st_index >= 0 && if (starter_property->comb1st_index >= 0 &&
current_property->comb2nd_index >= 0) { current_property->comb2nd_index >= 0) {
@ -491,7 +495,7 @@ DLLEXPORT ssize_t utf8proc_reencode(int32_t *buffer, ssize_t length, int options
current_property->comb2nd_index current_property->comb2nd_index
]; ];
if (composition >= 0 && (!(options & UTF8PROC_STABLE) || if (composition >= 0 && (!(options & UTF8PROC_STABLE) ||
!(utf8proc_get_property(composition)->comp_exclusion))) { !(get_property(composition)->comp_exclusion))) {
*starter = composition; *starter = composition;
starter_property = NULL; starter_property = NULL;
continue; continue;

4
utf8proc.h
View File

@ -310,8 +310,6 @@ DLLEXPORT const utf8proc_property_t *utf8proc_get_property(int32_t uc);
* the unicode char with the given code point 'uc'. * the unicode char with the given code point 'uc'.
* If the character is not existent a pointer to a special struct is * If the character is not existent a pointer to a special struct is
* returned, where 'category' is 0 (UTF8PROC_CATEGORY_CN). * returned, where 'category' is 0 (UTF8PROC_CATEGORY_CN).
* WARNING: The parameter 'uc' has to be in the range of 0x0000 to
* 0x10FFFF, otherwise the program might crash!
*/ */
DLLEXPORT ssize_t utf8proc_decompose_char( DLLEXPORT ssize_t utf8proc_decompose_char(
@ -338,8 +336,6 @@ DLLEXPORT ssize_t utf8proc_decompose_char(
* If the number of written chars would be bigger than 'bufsize', * If the number of written chars would be bigger than 'bufsize',
* the buffer (up to 'bufsize') has inpredictable data, and the needed * the buffer (up to 'bufsize') has inpredictable data, and the needed
* buffer size is returned. * buffer size is returned.
* WARNING: The parameter 'uc' has to be in the range of 0x0000 to
* 0x10FFFF, otherwise the program might crash!
*/ */
DLLEXPORT ssize_t utf8proc_decompose( DLLEXPORT ssize_t utf8proc_decompose(