diff options
| author | naruse <naruse@b2dd03c8-39d4-4d8f-98ff-823fe69b080e> | 2010-11-22 09:35:08 +0000 |
|---|---|---|
| committer | naruse <naruse@b2dd03c8-39d4-4d8f-98ff-823fe69b080e> | 2010-11-22 09:35:08 +0000 |
| commit | 60dfa6b655239a176c4fef4067c409c8b3f86e5d (patch) | |
| tree | dcda60fd5d87756e42c6b454d71824b9c4b3699c /tool | |
| parent | 6ade3a42918e94be1c6d4b1bd9273ec1e3265525 (diff) | |
* enc/big5.c: split CP950 from Big5.
* enc/big5.c: split CP951 from Big5-HKSCS.
* enc/trans/big5.trans: import conversion table of Big5, Big5-HKSCS,
CP950, and CP951 from ICU. they need fallback conversions.
ref [ruby-core:33256]
http://source.icu-project.org/repos/icu/data/trunk/charset/data/ucm/
* tool/transcode-tblgen.rb (import_ucm): add to import ucm files.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@29869 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
Diffstat (limited to 'tool')
| -rwxr-xr-x | tool/transcode-tblgen.rb | 21 |
1 files changed, 21 insertions, 0 deletions
diff --git a/tool/transcode-tblgen.rb b/tool/transcode-tblgen.rb index edc8db72a9..833fc219a8 100755 --- a/tool/transcode-tblgen.rb +++ b/tool/transcode-tblgen.rb @@ -748,6 +748,27 @@ def citrus_decode_mapsrc(ces, csid, mapsrcs) return table end +def import_ucm(path) + to_ucs = [] + from_ucs = [] + File.foreach(File.join($srcdir, "ucm", path)) do |line| + uc, bs, fb = nil + if /^<U(\h+)>\s*([\+\hx\\]+)\s*\|(\d)/ =~ line + uc = $1.hex + bs = $2.delete('x\\') + fb = $3.to_i + next if uc < 128 && uc == bs.hex + elsif /^([<U\h>+]+)\s*([\+\hx\\]+)\s*\|(\d)/ =~ line + uc = $1.scan(/\h+>/).map(&:hex).pack("U*").unpack("H*")[0] + bs = $2.delete('x\\') + fb = $3.to_i + end + to_ucs << [bs, uc] if fb == 0 || fb == 3 + from_ucs << [uc, bs] if fb == 0 || fb == 1 + end + [to_ucs, from_ucs] +end + def encode_utf8(map) r = [] map.each {|k, v| |