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.
* all generated files under lib/uni* and tests/uni*: Regenerate.
* all the dependant modules: Bump version.

1 files changed, 21 insertions, 8 deletions

diff --git a/lib/unigbrk/uc-is-grapheme-break.c b/lib/unigbrk/uc-is-grapheme-break.c
index f14a01c51a..8ac975b336 100644
--- a/lib/unigbrk/uc-is-grapheme-break.c
+++ b/lib/unigbrk/uc-is-grapheme-break.c
@@ -47,19 +47,22 @@
    /* GB8 */                                                            \
    ((A) == GBP_LVT || (A) == GBP_T) && (B) == GBP_T ? false :           \
                                                                         \
-   /* GB8a */								\
-   (A) == GBP_RI && (B) == GBP_RI ? false :				\
-									\
    /* GB9 */                                                            \
-   (B) == GBP_EXTEND ? false :                                          \
+   (B) == GBP_EXTEND || (B) == GBP_ZWJ ? false :                        \
                                                                         \
    /* GB9a */                                                           \
    (B) == GBP_SPACINGMARK ? false :                                     \
                                                                         \
    /* GB9b */                                                           \
-   (A) == GBP_PREPEND ? false                                           \
+   (A) == GBP_PREPEND ? false :                                         \
                                                                         \
-   /* GB10 */                                                           \
+   /* GB10 -- incomplete */                                             \
+   ((A) == GBP_EB || (A) == GBP_EBG) && (B) == GBP_EM ? false :         \
+                                                                        \
+   /* GB11 */                                                           \
+   (A) == GBP_ZWJ && ((B) == GBP_GAZ || (B) == GBP_EBG) ? false         \
+                                                                        \
+   /* GB999 */                                                          \
    : true)
 
 #define UC_GRAPHEME_BREAKS_FOR(A)                                       \
@@ -75,9 +78,14 @@
    | (UC_IS_GRAPHEME_BREAK(A, GBP_T)           << GBP_T)                \
    | (UC_IS_GRAPHEME_BREAK(A, GBP_LV)          << GBP_LV)               \
    | (UC_IS_GRAPHEME_BREAK(A, GBP_LVT)         << GBP_LVT)              \
-   | (UC_IS_GRAPHEME_BREAK(A, GBP_RI)          << GBP_RI))
+   | (UC_IS_GRAPHEME_BREAK(A, GBP_RI)          << GBP_RI)               \
+   | (UC_IS_GRAPHEME_BREAK(A, GBP_ZWJ)         << GBP_ZWJ)              \
+   | (UC_IS_GRAPHEME_BREAK(A, GBP_EB)          << GBP_EB)               \
+   | (UC_IS_GRAPHEME_BREAK(A, GBP_EM)          << GBP_EM)               \
+   | (UC_IS_GRAPHEME_BREAK(A, GBP_GAZ)         << GBP_GAZ)              \
+   | (UC_IS_GRAPHEME_BREAK(A, GBP_EBG)         << GBP_EBG))
 
-static const unsigned short int gb_table[13] =
+static const unsigned long int gb_table[18] =
   {
     UC_GRAPHEME_BREAKS_FOR(0),  /* GBP_OTHER */
     UC_GRAPHEME_BREAKS_FOR(1),  /* GBP_CR */
@@ -92,6 +100,11 @@ static const unsigned short int gb_table[13] =
     UC_GRAPHEME_BREAKS_FOR(10), /* GBP_LV */
     UC_GRAPHEME_BREAKS_FOR(11), /* GBP_LVT */
     UC_GRAPHEME_BREAKS_FOR(12), /* GBP_RI */
+    UC_GRAPHEME_BREAKS_FOR(13), /* GBP_ZWJ */
+    UC_GRAPHEME_BREAKS_FOR(14), /* GBP_EB */
+    UC_GRAPHEME_BREAKS_FOR(15), /* GBP_EM */
+    UC_GRAPHEME_BREAKS_FOR(16), /* GBP_GAZ */
+    UC_GRAPHEME_BREAKS_FOR(17), /* GBP_EBG */
   };
 
 bool