diff lib/unigbrk/uc-gbrk-prop.c @ 19273:638b6d1fdf36 ueno/unicode-9.0.0

libunistring: update to Unicode 9.0.0 * lib/gen-uni-tables.c (fill_properties): Recognize Sentence_Terminal and Prepended_Concatenation_Mark. (is_property_default_ignorable_code_point): Exclude U+08E2. (fill_arabicshaping): Allow missing whitespace when parsing; recognize "AFRICAN FEH", "AFRICAN QAF", and "AFRICAN MOON". (output_blocks): Increase the element size of the level1 table to accommodate more blocks. (get_lbp): Recognize ZWJ, E_Base, and E_Modifier characters; Update each class according to the standard. (get_wbp): Recognize ZWJ, E_Base, E_Modifier, Glue_After_Zwj, and E_Base_GAZ characters. (output_gbp_table): Recognize ZWJ, E_Base, E_Modifier, Glue_After_Zwj, and E_Base_GAZ characters. * lib/unictype.in.h (UC_JOINING_GROUP_AFRICAN_FEH, UC_JOINING_GROUP_AFRICAN_QAF, UC_JOINING_GROUP_AFRICAN_MOON): New enum value. * lib/unilbrk/lbrktables.h (LBP_ZWJ, LBP_EB, LBP_EM): New enum value. * lib/unilbrk/lbrktables.c (unilbrk_table): Extend the table with LBP_ZWJ, LBP_EB, and LBP_EM. * lib/uniwbrk.in.h (WBP_ZWJ, WBP_EB, WBP_EM, WBP_GAZ, WBP_EBG): New enum value. * lib/uniwbrk/u-wordbreaks.h: Implement WB3c, WB15, and WB16. * lib/uniwbrk/wbrktable.h (uniwbrk_prop_index): New variable declaration. * lib/uniwbrk/wbrktable.c (uniwbrk_prop_index): New variable. (uniwbrk_table): Implement WB14. * tests/uniwbrk/test-uc-wordbreaks.c (wordbreakproperty_to_string): Check WBP_ZWJ, WBP_EB, WBP_EM, WBP_GAZ, and WBP_EBG. * modules/unigbrk/u{32,16,8}-grapheme-breaks: No longer depend on uc-is-grapheme-break. * modules/unigbrk/uc-grapheme-breaks: New module. * modules/unigbrk/uc-grapheme-breaks-tests: New module. * lib/unigbrk.in.h (GBP_ZWJ, GBP_EB, GBP_EM, GBP_GAZ, GBP_EBG): New enum value. (uc_grapheme_breaks): New function, replacing uc_is_grapheme_break. * lib/unigbrk/u-grapheme-breaks.h: New file. * lib/unigbrk/u{32,16,8}-grapheme-breaks.c: Rewrite using u-grapheme-breaks.h instead of uc_is_grapheme_break. * lib/unigbrk/uc-grapheme-breaks.c: New file. * lib/unigbrk/uc-is-grapheme-break.c: Partially update to TR29 rev 29. * tests/unigbrk/test-uc-gbrk-prop.c (graphemebreakproperty_to_string): Check GBP_ZWJ, GBP_EB, GBP_EM, GBP_GAZ, and GBP_EBG. * tests/unigbrk/test-uc-grapheme-breaks.c: New test. * tests/unigbrk/test-uc-is-grapheme-break.c (graphemebreakproperty_to_string): Check GBP_ZWJ, GBP_EB, GBP_EM, GBP_GAZ, and GBP_EBG. (main): Skip unsupported rules involving 3 or more characters, namely GB10, GB12, and GB13. * lib/uniwidth/width.c (nonspacing_table_data): Update.
author Daiki Ueno <ueno@gnu.org>
date Wed, 12 Oct 2016 17:40:37 +0200
parents 9759915b2aca
children 10eb9086bea0
line wrap: on
line diff
--- a/lib/unigbrk/uc-gbrk-prop.c	Sun Oct 29 16:22:41 2017 -0700
+++ b/lib/unigbrk/uc-gbrk-prop.c	Wed Oct 12 17:40:37 2016 +0200
@@ -36,8 +36,7 @@
           if (lookup2 >= 0)
             {
               unsigned int index3 = uc & gbrkprop_header_4;
-              unsigned char lookup3 = unigbrkprop.level3[lookup2 + index3 / 2];
-              return (lookup3 >> ((uc & 1) << 2)) & 0x0f;
+              return unigbrkprop.level3[lookup2 + index3];
             }
         }
     }