From a67d4fa01c43124048be957f7d8e4090b7255393 Mon Sep 17 00:00:00 2001 From: akr Date: Tue, 16 Sep 2008 16:48:05 +0000 Subject: * include/ruby/oniguruma.h (OnigEncodingTypeST): add precise_ret argument for mbc_to_code. (ONIGENC_MBC_TO_CODE): provide NULL for precise_ret. (ONIGENC_MBC_PRECISE_CODEPOINT): defined. * include/ruby/encoding.h (rb_enc_mbc_precise_codepoint): defined. * regenc.h (onigenc_single_byte_mbc_to_code): precise_ret argument added. (onigenc_mbn_mbc_to_code): ditto. * regenc.c (onigenc_single_byte_mbc_to_code): precise_ret argument added. (onigenc_mbn_mbc_to_code): ditto. * string.c (count_utf8_lead_bytes_with_word): removed. (str_utf8_nth): removed. (str_utf8_offset): removed. (str_strlen): UTF-8 codepoint oriented optimization removed. (rb_str_substr): ditto. (enc_succ_char): use rb_enc_mbc_precise_codepoint. (enc_pred_char): ditto. (rb_str_succ): ditto. * encoding.c (rb_enc_ascget): check length with rb_enc_mbc_precise_codepoint. (rb_enc_codepoint): use rb_enc_mbc_precise_codepoint. * regexec.c (string_cmp_ic): add text_end argument. (match_at): check end of character after exact string matches. * enc/utf_8.c (graphme_table): defined for extended graphme cluster boundary. (grapheme_cmp): defined. (get_grapheme_properties): defined. (grapheme_boundary_p): defined. (MAX_BYTES_LENGTH): defined. (comb_char_enc_len): defined. (mbc_to_code0): extracted from mbc_to_code. (mbc_to_code): use mbc_to_code0. (left_adjust_combchar_head): defined. (utf_8): use a extended graphme cluster as a unit. * enc/unicode.c (onigenc_unicode_mbc_case_fold): use ONIGENC_MBC_PRECISE_CODEPOINT to extract codepoints. (onigenc_unicode_get_case_fold_codes_by_str): ditto. * enc/euc_jp.c (mbc_to_code): follow mbc_to_code field change. use onigenc_mbn_mbc_to_code. * enc/shift_jis.c (mbc_to_code): ditto. * enc/emacs_mule.c (mbc_to_code): ditto. * enc/gbk.c (gbk_mbc_to_code): follow mbc_to_code field and onigenc_mbn_mbc_to_code change. * enc/cp949.c (cp949_mbc_to_code): ditto. * enc/big5.c (big5_mbc_to_code): ditto. * enc/euc_tw.c (euctw_mbc_to_code): ditto. * enc/euc_kr.c (euckr_mbc_to_code): ditto. * enc/gb18030.c (gb18030_mbc_to_code): ditto. * enc/utf_32be.c (utf32be_mbc_to_code): follow mbc_to_code field change. * enc/utf_16be.c (utf16be_mbc_to_code): ditto. * enc/utf_32le.c (utf32le_mbc_to_code): ditto. * enc/utf_16le.c (utf16le_mbc_to_code): ditto. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@19389 b2dd03c8-39d4-4d8f-98ff-823fe69b080e --- string.c | 109 +++++++-------------------------------------------------------- 1 file changed, 12 insertions(+), 97 deletions(-) (limited to 'string.c') diff --git a/string.c b/string.c index 501320ae25..e6b2b6265c 100644 --- a/string.c +++ b/string.c @@ -807,24 +807,6 @@ rb_enc_strlen_cr(const char *p, const char *e, rb_encoding *enc, int *cr) return c; } -#ifdef NONASCII_MASK -#define is_utf8_lead_byte(c) (((c)&0xC0) != 0x80) -static inline VALUE -count_utf8_lead_bytes_with_word(const VALUE *s) -{ - VALUE d = *s; - d |= ~(d>>1); - d >>= 6; - d &= NONASCII_MASK >> 7; - d += (d>>8); - d += (d>>16); -#if SIZEOF_VALUE == 8 - d += (d>>32); -#endif - return (d&0xF); -} -#endif - static long str_strlen(VALUE str, rb_encoding *enc) { @@ -835,32 +817,6 @@ str_strlen(VALUE str, rb_encoding *enc) if (!enc) enc = STR_ENC_GET(str); p = RSTRING_PTR(str); e = RSTRING_END(str); -#ifdef NONASCII_MASK - if (ENC_CODERANGE(str) == ENC_CODERANGE_VALID && - enc == rb_utf8_encoding()) { - VALUE len = 0; - if (sizeof(VALUE) * 2 < e - p) { - const VALUE *s, *t; - const VALUE lowbits = sizeof(VALUE) - 1; - s = (const VALUE*)(~lowbits & ((VALUE)p + lowbits)); - t = (const VALUE*)(~lowbits & (VALUE)e); - while (p < (const char *)s) { - if (is_utf8_lead_byte(*p)) len++; - p++; - } - while (s < t) { - len += count_utf8_lead_bytes_with_word(s); - s++; - } - p = (const char *)s; - } - while (p < e) { - if (is_utf8_lead_byte(*p)) len++; - p++; - } - return (long)len; - } -#endif n = rb_enc_strlen_cr(p, e, enc, &cr); if (cr) { ENC_CODERANGE_SET(str, cr); @@ -1227,44 +1183,6 @@ str_offset(const char *p, const char *e, int nth, rb_encoding *enc, int singleby return pp - p; } -#ifdef NONASCII_MASK -static char * -str_utf8_nth(const char *p, const char *e, int nth) -{ - if (sizeof(VALUE) * 2 < nth) { - const VALUE *s, *t; - const VALUE lowbits = sizeof(VALUE) - 1; - s = (const VALUE*)(~lowbits & ((VALUE)p + lowbits)); - t = (const VALUE*)(~lowbits & (VALUE)e); - while (p < (const char *)s) { - if (is_utf8_lead_byte(*p)) nth--; - p++; - } - do { - nth -= count_utf8_lead_bytes_with_word(s); - s++; - } while (s < t && sizeof(VALUE) <= nth); - p = (char *)s; - } - while (p < e) { - if (is_utf8_lead_byte(*p)) { - if (nth == 0) break; - nth--; - } - p++; - } - return (char *)p; -} - -static int -str_utf8_offset(const char *p, const char *e, int nth) -{ - const char *pp = str_utf8_nth(p, e, nth); - if (!pp) return e - p; - return pp - p; -} -#endif - /* byte offset to char offset */ long rb_str_sublen(VALUE str, long pos) @@ -1338,13 +1256,6 @@ rb_str_substr(VALUE str, long beg, long len) if (len == 0) { p = 0; } -#ifdef NONASCII_MASK - else if (ENC_CODERANGE(str) == ENC_CODERANGE_VALID && - enc == rb_utf8_encoding()) { - p = str_utf8_nth(s, e, beg); - len = str_utf8_offset(p, e, len); - } -#endif else if (rb_enc_mbmaxlen(enc) == rb_enc_mbminlen(enc)) { int char_sz = rb_enc_mbmaxlen(enc); @@ -2454,7 +2365,7 @@ enc_succ_char(char *p, int len, rb_encoding *enc) if (i < 0) return NEIGHBOR_WRAPPED; ++((unsigned char*)p)[i]; - l = rb_enc_precise_mbclen(p, p+len, enc); + rb_enc_mbc_precise_codepoint(p, p+len, &l, enc); if (MBCLEN_CHARFOUND_P(l)) { l = MBCLEN_CHARFOUND_LEN(l); if (l == len) { @@ -2467,7 +2378,7 @@ enc_succ_char(char *p, int len, rb_encoding *enc) if (MBCLEN_INVALID_P(l) && i < len-1) { int len2, l2; for (len2 = len-1; 0 < len2; len2--) { - l2 = rb_enc_precise_mbclen(p, p+len2, enc); + rb_enc_mbc_precise_codepoint(p, p+len2, &l2, enc); if (!MBCLEN_INVALID_P(l2)) break; } @@ -2486,7 +2397,7 @@ enc_pred_char(char *p, int len, rb_encoding *enc) if (i < 0) return NEIGHBOR_WRAPPED; --((unsigned char*)p)[i]; - l = rb_enc_precise_mbclen(p, p+len, enc); + rb_enc_mbc_precise_codepoint(p, p+len, &l, enc); if (MBCLEN_CHARFOUND_P(l)) { l = MBCLEN_CHARFOUND_LEN(l); if (l == len) { @@ -2499,7 +2410,7 @@ enc_pred_char(char *p, int len, rb_encoding *enc) if (MBCLEN_INVALID_P(l) && i < len-1) { int len2, l2; for (len2 = len-1; 0 < len2; len2--) { - l2 = rb_enc_precise_mbclen(p, p+len2, enc); + rb_enc_mbc_precise_codepoint(p, p+len2, &l2, enc); if (!MBCLEN_INVALID_P(l2)) break; } @@ -2606,7 +2517,7 @@ rb_str_succ(VALUE orig) VALUE str; char *sbeg, *s, *e, *last_alnum = 0; int c = -1; - long l; + int l; char carry[ONIGENC_CODE_TO_MBC_MAXLEN] = "\1"; int carry_pos = 0, carry_len = 1; enum neighbor_char neighbor = NEIGHBOR_FOUND; @@ -2628,7 +2539,8 @@ rb_str_succ(VALUE orig) break; } } - if ((l = rb_enc_precise_mbclen(s, e, enc)) <= 0) continue; + rb_enc_mbc_precise_codepoint(s, e, &l, enc); + if (l <= 0) continue; neighbor = enc_succ_alnum_char(s, l, enc, carry); switch (neighbor) { case NEIGHBOR_NOT_CHAR: @@ -2647,11 +2559,14 @@ rb_str_succ(VALUE orig) s = e; while ((s = rb_enc_prev_char(sbeg, s, e, enc)) != 0) { enum neighbor_char neighbor; - if ((l = rb_enc_precise_mbclen(s, e, enc)) <= 0) continue; + int l2; + rb_enc_mbc_precise_codepoint(s, e, &l, enc); + if (l <= 0) continue; neighbor = enc_succ_char(s, l, enc); if (neighbor == NEIGHBOR_FOUND) return str; - if (rb_enc_precise_mbclen(s, s+l, enc) != l) { + rb_enc_mbc_precise_codepoint(s, s+l, &l2, enc); + if (l2 != l) { /* wrapped to \0...\0. search next valid char. */ enc_succ_char(s, l, enc); } -- cgit v1.2.3