diff options
| -rw-r--r-- | ChangeLog | 11 | ||||
| -rw-r--r-- | enc/unicode.c | 8 | ||||
| -rw-r--r-- | enc/unicode/casefold.h | 8 | ||||
| -rw-r--r-- | regparse.c | 14 |
4 files changed, 32 insertions, 9 deletions
@@ -1,3 +1,14 @@ +Thu Mar 1 01:25:43 2012 NARUSE, Yui <naruse@ruby-lang.org> + + * regparse.c (is_onechar_cclass): optimize character class + Merge Onigmo 27278c12e6674043cc8affca6507e20e119a86ee. + + * regparse.c (is_onechar_cclass): [bug] unexpected match occurs when a + char class contains no char + + * enc/unicode.c (init_case_fold_table): define the sizes of case + folding tables in casefold.h + Wed Feb 29 16:11:34 2012 NAKAMURA Usaku <usa@ruby-lang.org> * win32/Makefile.sub (MAKEDIRS): use mkdir of cmd.exe instead of ruby. diff --git a/enc/unicode.c b/enc/unicode.c index 1dda3f32f8..20990c1e54 100644 --- a/enc/unicode.c +++ b/enc/unicode.c @@ -238,7 +238,7 @@ static int init_case_fold_table(void) THREAD_ATOMIC_START; - FoldTable = st_init_numtable_with_size(1400); + FoldTable = st_init_numtable_with_size(FOLD_TABLE_SIZE); if (ONIG_IS_NULL(FoldTable)) return ONIGERR_MEMORY; for (i = 0; i < numberof(CaseFold); i++) { p = &CaseFold[i]; @@ -249,7 +249,7 @@ static int init_case_fold_table(void) st_add_direct(FoldTable, (st_data_t )p->from, (st_data_t )&(p->to)); } - Unfold1Table = st_init_numtable_with_size(1200); + Unfold1Table = st_init_numtable_with_size(UNFOLD1_TABLE_SIZE); if (ONIG_IS_NULL(Unfold1Table)) return ONIGERR_MEMORY; for (i = 0; i < numberof(CaseUnfold_11); i++) { @@ -261,7 +261,7 @@ static int init_case_fold_table(void) st_add_direct(Unfold1Table, (st_data_t )p1->from, (st_data_t )&(p1->to)); } - Unfold2Table = st_init_table_with_size(&type_code2_hash, 200); + Unfold2Table = st_init_table_with_size(&type_code2_hash, UNFOLD2_TABLE_SIZE); if (ONIG_IS_NULL(Unfold2Table)) return ONIGERR_MEMORY; for (i = 0; i < numberof(CaseUnfold_12); i++) { @@ -273,7 +273,7 @@ static int init_case_fold_table(void) st_add_direct(Unfold2Table, (st_data_t )p2->from, (st_data_t )(&p2->to)); } - Unfold3Table = st_init_table_with_size(&type_code3_hash, 30); + Unfold3Table = st_init_table_with_size(&type_code3_hash, UNFOLD3_TABLE_SIZE); if (ONIG_IS_NULL(Unfold3Table)) return ONIGERR_MEMORY; for (i = 0; i < numberof(CaseUnfold_13); i++) { diff --git a/enc/unicode/casefold.h b/enc/unicode/casefold.h index b4f2ee3c9e..b9559de4a3 100644 --- a/enc/unicode/casefold.h +++ b/enc/unicode/casefold.h @@ -1,3 +1,6 @@ +/* DO NOT EDIT THIS FILE. */ +/* Generated by tool/CaseFolding.py */ + static const CaseFold_11_Type CaseFold[] = { { 0x0041, {1, {0x0061}}}, { 0x0042, {1, {0x0062}}}, @@ -2228,3 +2231,8 @@ static const CaseUnfold_13_Type CaseUnfold_13[] = { { {0x03c5, 0x0313, 0x0342}, {1, {0x1f56 }}}, { {0x03c9, 0x0342, 0x03b9}, {1, {0x1ff7 }}}, }; + +#define FOLD_TABLE_SIZE 1357 +#define UNFOLD1_TABLE_SIZE 1207 +#define UNFOLD2_TABLE_SIZE 88 +#define UNFOLD3_TABLE_SIZE 23 diff --git a/regparse.c b/regparse.c index 08ffa8060d..f1e8e4f20a 100644 --- a/regparse.c +++ b/regparse.c @@ -1786,8 +1786,8 @@ add_code_range_to_buf0(BBuf** pbuf, ScanEnv* env, OnigCodePoint from, OnigCodePo if (inc_n != 1) { if (checkdup && from <= data[low*2+1] - && (data[low*2] <= from || data[low*2+1] <= to)) - CC_DUP_WARN(env); + && (data[low*2] <= from || data[low*2+1] <= to)) + CC_DUP_WARN(env); if (from > data[low*2]) from = data[low*2]; if (to < data[(high - 1)*2 + 1]) @@ -1799,7 +1799,7 @@ add_code_range_to_buf0(BBuf** pbuf, ScanEnv* env, OnigCodePoint from, OnigCodePo int to_pos = SIZE_CODE_POINT * (1 + (low + 1) * 2); if (inc_n > 0) { - if ((OnigCodePoint )high < n) { + if (high < n) { int size = (n - high) * 2 * SIZE_CODE_POINT; BBUF_MOVE_RIGHT(bbuf, from_pos, to_pos, size); } @@ -5722,9 +5722,13 @@ is_onechar_cclass(CClassNode* cc, OnigCodePoint* code) } } } + if (found == 0) { + /* the character class contains no char. */ + return 0; + } if (j >= 0) { /* only one char found in the bitset, calculate the code point. */ - c = BITS_IN_ROOM * j + (countbits(b2 - 1) & 0x1f); + c = BITS_IN_ROOM * j + countbits(b2 - 1); } *code = c; return 1; @@ -5746,7 +5750,7 @@ parse_exp(Node** np, OnigToken* tok, int term, switch (tok->type) { case TK_ALT: case TK_EOT: - end_of_token: + end_of_token: *np = node_new_empty(); return tok->type; break; |