Upgrade to Unicode 3.2.

diff --git a/lib/ChangeLog b/lib/ChangeLog
index 6c52356b193c3d4cf6cf30dcf734794d8e384686..f7ca99942a08c27022ccfa42565d3d55dd2cd9f6 100644 (file)
--- a/lib/ChangeLog
+++ b/lib/ChangeLog
@@ -1,3 +1,8 @@
+2002-05-29  Bruno Haible  <bruno@clisp.org>
+
+       * linebreak.c (nonspacing_table_data, uc_width): Upgrade to
+       Unicode 3.2.0.
+
 2002-05-18  Bruno Haible  <bruno@clisp.org>
 
        * basename.h: Use __GLIBC__ >= 2, because BeOS (which defines __GLIBC__

diff --git a/lib/linebreak.c b/lib/linebreak.c
index 1edc69f3aa40a741a64e3d2154b9dfbd5ccaeee4..f62826b2f73df5f770b27031baeaa19c2ff4fb34 100644 (file)
--- a/lib/linebreak.c
+++ b/lib/linebreak.c
@@ -226,9 +226,13 @@ int uc_width PARAMS ((unsigned int uc, const char *encoding));
 
 /*
  * Non-spacing attribute table.
- * See PropList.txt, or grep '^[^;]*;[^;]*;[^;]*;[^;]*;NSM;' UnicodeData.txt
- * Control characters are also marked non-spacing here, because they are not
- * printable. Zero width characters are also marked non-spacing here.
+ * Consists of:
+ * - Non-spacing characters; generated from PropList.txt or
+ *   "grep '^[^;]*;[^;]*;[^;]*;[^;]*;NSM;' UnicodeData.txt"
+ * - Format control characters; generated from
+ *   "grep '^[^;]*;[^;]*;Cf;' UnicodeData.txt"
+ * - Zero width characters; generated from
+ *   "grep '^[^;]*;ZERO WIDTH ' UnicodeData.txt"
  */
 static const unsigned char nonspacing_table_data[16*64] = {
   /* 0x0000-0x01ff */
@@ -246,7 +250,7 @@ static const unsigned char nonspacing_table_data[16*64] = {
   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x0280-0x02bf */
   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x02c0-0x02ff */
   0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, /* 0x0300-0x033f */
-  0xff, 0x7f, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, /* 0x0340-0x037f */
+  0xff, 0xff, 0x00, 0x00, 0xff, 0xff, 0x00, 0x00, /* 0x0340-0x037f */
   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x0380-0x03bf */
   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x03c0-0x03ff */
   /* 0x0400-0x05ff */
@@ -317,8 +321,8 @@ static const unsigned char nonspacing_table_data[16*64] = {
   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x1640-0x167f */
   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x1680-0x16bf */
   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x16c0-0x16ff */
-  0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x1700-0x173f */
-  0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x1740-0x177f */
+  0x00, 0x00, 0x1c, 0x00, 0x00, 0x00, 0x1c, 0x00, /* 0x1700-0x173f */
+  0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x0c, 0x00, /* 0x1740-0x177f */
   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x80, 0x3f, /* 0x1780-0x17bf */
   0x40, 0xfe, 0x0f, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x17c0-0x17ff */
   /* 0x1800-0x19ff */
@@ -332,9 +336,9 @@ static const unsigned char nonspacing_table_data[16*64] = {
   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x19c0-0x19ff */
   /* 0x2000-0x21ff */
   0x00, 0xf8, 0x00, 0x00, 0x00, 0x7c, 0x00, 0x00, /* 0x2000-0x203f */
-  0x00, 0x00, 0x00, 0x00, 0x00, 0xfc, 0x00, 0x00, /* 0x2040-0x207f */
+  0x00, 0x00, 0x00, 0x00, 0x0f, 0xfc, 0x00, 0x00, /* 0x2040-0x207f */
   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x2080-0x20bf */
-  0x00, 0x00, 0xff, 0xff, 0x0f, 0x00, 0x00, 0x00, /* 0x20c0-0x20ff */
+  0x00, 0x00, 0xff, 0xff, 0xff, 0x07, 0x00, 0x00, /* 0x20c0-0x20ff */
   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x2100-0x213f */
   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x2140-0x217f */
   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x2180-0x21bf */
@@ -358,7 +362,7 @@ static const unsigned char nonspacing_table_data[16*64] = {
   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0xfb80-0xfbbf */
   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0xfbc0-0xfbff */
   /* 0xfe00-0xffff */
-  0x00, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, /* 0xfe00-0xfe3f */
+  0xff, 0xff, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, /* 0xfe00-0xfe3f */
   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0xfe40-0xfe7f */
   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0xfe80-0xfebf */
   0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x80, /* 0xfec0-0xfeff */
@@ -428,6 +432,11 @@ uc_width (uc, encoding)
              return 0;
          }
     }
+  else if ((uc >> 9) == (0xe0000 >> 9))
+    {
+      if (uc >= 0xe0020 ? uc <= 0xe007f : uc == 0xe0001)
+       return 0;
+    }
   /* Test for double-width character.
    * Generated from "grep '^....;[WF]' EastAsianWidth.txt"
    * and            "grep '^....;[^WF]' EastAsianWidth.txt"
@@ -435,18 +444,19 @@ uc_width (uc, encoding)
   if (uc >= 0x1100
       && ((uc < 0x1160) /* Hangul Jamo */
          || (uc >= 0x2e80 && uc < 0xa4d0  /* CJK ... Yi */
-             && !((uc & ~0x0013) == 0x3008
-                  || (uc & ~0x0001) == 0x3014
-                  || uc == 0x303f))
+             && !(uc == 0x303f))
          || (uc >= 0xac00 && uc < 0xd7a4) /* Hangul Syllables */
          || (uc >= 0xf900 && uc < 0xfb00) /* CJK Compatibility Ideographs */
          || (uc >= 0xfe30 && uc < 0xfe70) /* CJK Compatibility Forms */
-         || (uc >= 0xff00 && uc < 0xff60) /* Fullwidth Forms */
-         || (uc >= 0xffe0 && uc < 0xffe7)))
+         || (uc >= 0xff00 && uc < 0xff61) /* Fullwidth Forms */
+         || (uc >= 0xffe0 && uc < 0xffe7)
+         || (uc >= 0x20000 && uc <= 0x2a6d6) /* CJK */
+         || (uc >= 0x2f800 && uc <= 0x2fa1d) /* CJK Compatibility Ideographs */
+     )   )
     return 2;
   /* In ancient CJK encodings, Cyrillic and most other characters are
      double-width as well.  */
-  if (uc >= 0x00A1 && uc < 0xFF60 && uc != 0x20A9
+  if (uc >= 0x00A1 && uc < 0xFF61 && uc != 0x20A9
       && is_cjk_encoding (encoding))
     return 2;
   return 1;