Re: [PATCH] libunistring: update to Unicode 8.0.0

2015-06-19 Thread Daiki Ueno 
Daiki Ueno  writes:

> Daiki Ueno  writes:
>
>> Due to the size limitation, I'm sending manually adjusted files only.
>> Generated files have been already committed too.
>
> Also updated uniname/uniname, which contains only generated content.

One more thing: I've bumped the minimum libstring version of affected
modules to 0.9.6, so libunistring-optional should work:

2015-06-19  Daiki Ueno  

libunistring: bump minimum version to 0.9.6
* all modules depending on updated Unicode data: Regenerate.
The modules are listed by a script that does:
- for each file listed by: git show --oneline --name-only 705f4efc
  - deduce the containing modules, based on "Files:"
- deduce the modules which depend on the containing modules, based
  on "Depends-on:"

It's tempting to automate those steps.  Is it okay to add some scripts
for that to Gnulib?

Regards,
-- 
Daiki Ueno

Re: [PATCH] libunistring: update to Unicode 8.0.0

2015-06-18 Thread Daiki Ueno 
Daiki Ueno  writes:

> Due to the size limitation, I'm sending manually adjusted files only.
> Generated files have been already committed too.

Also updated uniname/uniname, which contains only generated content.

Regards,
-- 
Daiki Ueno

[PATCH] libunistring: update to Unicode 8.0.0

2015-06-18 Thread Daiki Ueno 
* lib/gen-uni-tables.c (SIZEOF): New macro.
(output_numeric): Increase the maximum number of fractions from
128 to 160.  Increase the level3 value width from 7 bits to 8
bits.  Use SIZEOF instead of a hard-coded integer.
(output_blocks): Decrease the cut-off threshold from 0x3 to
0x28000.
(fill_blocks): Increase the maximum number of blocks from 256 to
384.  Use SIZEOF instead of a hard-coded integer.
(get_lbp): Adjust to new characters added in Unicode 8.0.0.
* lib/unictype/numeric.c (uc_numeric_value): Adjust the level3
value width.
* lib/unilbrk/lbrktables.c (unilbrk_table): Implement LBP21b and
a new case added to LBP22.
* lib/uniwidth/width.c (nonspacing_table_data): Add U+08E3,
U+A69E, U+FE2E..U+FE2F, U+111CA..U+111CC, U+11300,
U+115DC..U+115DD, U+1171D..U+1171F, U+11722..U+11725,
U+11727..U+1172B, U+1DA00..U+1DA36, U+1DA3B..U+1DA6C, U+1DA75,
U+1DA84, U+1DA9B..U+1DA9F, and U+1DAA1..U+1DAAF.
* tests/uniwidth/test-uc_width2.sh: Same updates as in
lib/uniwidth/width.c.
* all generated files under lib/uni* and tests/uni*: Regenerate.
---
 ChangeLog| 25 +++
 lib/gen-uni-tables.c | 69 ++--
 lib/unictype/numeric.c   |  6 ++--
 lib/unilbrk/lbrktables.c |  4 +--
 lib/uniwidth/width.c | 29 +++--
 tests/uniwidth/test-uc_width2.sh | 45 +++---
 6 files changed, 127 insertions(+), 51 deletions(-)

Due to the size limitation, I'm sending manually adjusted files only.
Generated files have been already committed too.

diff --git a/ChangeLog b/ChangeLog
index 53fc032..8cd1336 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,28 @@
+2015-06-18  Daiki Ueno  
+
+   libunistring: update to Unicode 8.0.0
+   * lib/gen-uni-tables.c (SIZEOF): New macro.
+   (output_numeric): Increase the maximum number of fractions from
+   128 to 160.  Increase the level3 value width from 7 bits to 8
+   bits.  Use SIZEOF instead of a hard-coded integer.
+   (output_blocks): Decrease the cut-off threshold from 0x3 to
+   0x28000.
+   (fill_blocks): Increase the maximum number of blocks from 256 to
+   384.  Use SIZEOF instead of a hard-coded integer.
+   (get_lbp): Adjust to new characters added in Unicode 8.0.0.
+   * lib/unictype/numeric.c (uc_numeric_value): Adjust the level3
+   value width.
+   * lib/unilbrk/lbrktables.c (unilbrk_table): Implement LBP21b and
+   a new case added to LBP22.
+   * lib/uniwidth/width.c (nonspacing_table_data): Add U+08E3,
+   U+A69E, U+FE2E..U+FE2F, U+111CA..U+111CC, U+11300,
+   U+115DC..U+115DD, U+1171D..U+1171F, U+11722..U+11725,
+   U+11727..U+1172B, U+1DA00..U+1DA36, U+1DA3B..U+1DA6C, U+1DA75,
+   U+1DA84, U+1DA9B..U+1DA9F, and U+1DAA1..U+1DAAF.
+   * tests/uniwidth/test-uc_width2.sh: Same updates as in
+   lib/uniwidth/width.c.
+   * all generated files under lib/uni* and tests/uni*: Regenerate.
+
 2015-06-16  Pádraig Brady  
 
gnu-web-doc-update: add --mirror to remove stale files
diff --git a/lib/gen-uni-tables.c b/lib/gen-uni-tables.c
index 03fac9f..b69229f 100644
--- a/lib/gen-uni-tables.c
+++ b/lib/gen-uni-tables.c
@@ -43,6 +43,8 @@
 #include 
 #include 
 
+#define SIZEOF(a) (sizeof(a) / sizeof(a[0]))
+
 /* = */
 
 /* Reading UnicodeData.txt.  */
@@ -2130,7 +2132,7 @@ static void
 output_numeric (const char *filename, const char *version)
 {
   FILE *stream;
-  uc_fraction_t fractions[128];
+  uc_fraction_t fractions[160];
   unsigned int nfractions;
   unsigned int ch, i, j;
   struct numeric_table t;
@@ -2161,7 +2163,7 @@ output_numeric (const char *filename, const char *version)
   break;
   if (i == nfractions)
 {
-  assert (nfractions != 128);
+  assert (nfractions != SIZEOF (fractions));
   for (i = 0; i < nfractions; i++)
 if (value.denominator < fractions[i].denominator
 || (value.denominator == fractions[i].denominator
@@ -2226,7 +2228,7 @@ output_numeric (const char *filename, const char *version)
   fprintf (stream, "int level1[%zu];\n", t.level1_size);
   fprintf (stream, "short level2[%zu << %d];\n", t.level2_size, t.q);
   fprintf (stream, "unsigned short level3[%zu * %d + 1];\n", t.level3_size,
-   (1 << t.p) * 7 / 16);
+   (1 << t.p) * 8 / 16);
   fprintf (stream, "  }\n");
   fprintf (stream, "u_numeric =\n");
   fprintf (stream, "{\n");
@@ -2270,32 +2272,32 @@ output_numeric (const char *filename, const char 
*version)
   if (t.level2_size << t.q > 8)
 fprintf (stream, "\n ");
   fprintf (stream, " },\n");
-  /* Pack the level3 array.  Each entry needs 7 bits only.  Use 16-bit units,
+  /* Pack the level3 array.  Each entry needs 8 bits only.  Use 16-bit units,
  not 32-bit units, in order to make the lookup function easier.  */
   level3_packed =
 (uint