Fix script for characters in 3300..3357 range

* admin/unidata/blocks.awk: The characters in the 3300..3357 range
are Katakana according to Unicode's Scripts.txt.

(cherry picked from commit c70ac0fef12f5c84b2a62a4a8b1bdf78f69300ee)

diff --git a/admin/unidata/blocks.awk b/admin/unidata/blocks.awk
index 122164ce5b632480759d00969c86a7edc922c5c1..40cefe0d56303a926407cbf79cbdccc4b2a47a29 100755 (executable)
--- a/admin/unidata/blocks.awk
+++ b/admin/unidata/blocks.awk
@@ -146,6 +146,19 @@ FILENAME ~ "Blocks.txt" && /^[0-9A-F]/ {
     end[i] = fix_end[e] ? fix_end[e]: e
     name[i] = $0
 
+    # Hard-coded splits that must be processed before name2alias and
+    # before combining same-named adjacent ranges.
+    if (start[i] == "3300") # See Scripts.txt
+    {
+       end[i] = "3357"
+       name[i] = "Katakana"
+       alt[i] = "kana"
+       i++
+       start[i] = "3358"
+       end[i] = "33FF"
+       name[i] = "CJK Compatibility"
+    }
+
     alt[i] = name2alias(name[i])
 
     if (!alt[i])