> > BTW, will Mozilla have support for zh_CN's GBK (GB 13000.1) character set, > > a superset of GB 2312? Microsoft's definition of the CP936 codepage has > > been upgraded from GB 2312 to GBK for quite a while now (at least, > > starting with the mainland China version of Windows 95), and the > > above-mentioned MS Hei and MS Song are GBK fonts. Internet Explorer 3.x > > handles GBK fine, but I cannot even get Netscape 4.5 (yes, I know Netscape > > != Mozilla) to display GBK. > > This is a very good question. Frank, would it be possible to add GBK > support to our Unicode converters? > > Erik It should not be a problem, but we should 1) make it as a seperate charset from GB2312, 2) make the converter use two table, a. a GB2312 table, and b. a table which have the diff between GB2312 and GB13000 to do that so we want wast a lot of space.

Marking M6 for now.

moved to M8

CID have been defined. placeholder have been add in mozilla/intl/uconv/ucvcn. Need to make diff of GBK and GB2312 to generate smaller table and chain the diff table w/ gb2312 table to keep the size small enough. Leave it as M9.

push to M10

I probably will do this in the same time when I tune the GB2312 conversion performance so we can share the same table to reduce the size.

GBK is cp936, not the cp936 print on the Nadin's book but the cp936 document in the ftp.unicode.org mapping

*** Bug 7961 has been marked as a duplicate of this bug. ***

*** Bug 7961 has been marked as a duplicate of this bug. ***

Change to M14 since it is post beta

you may want to split this bug to two, one for nsIUnicodeEncoder, one for nsIUnicodeDecoder. Mark the M build, please. Thanks.

per email communication. Mark this for M11

I have check in the 1st prototype code for yueheng.xu@intel.com. The following thing need to resolve in M12: 1. Share the same mapping table with GB2312 2. Change the super class from the nsMultiTableDecoder/Encoder to BufferEncoder/Decoder 3. Remove the referenced to the shift table. Move this to M12

New code got check in before M11. Now it is clean. But yueheng.xu@intel.com need more unitest. Teruko, can we push some of our GB2312 and HZ test cases out so he can use them ?

A optimized decoder is check-in. Of course we can further optimize the speed if we use more memory for the mapping table.

Changed QA contact to erik@netscape.com

I believe ftang would be a better QA agent for this task.

change teh QA contact back to teruko. Teruko- you should create test cases for GBK and verify the fix.

I created test case in http://babel/tests/browser/charset/gbx/gbx1.txt, http://babel/tests/browser/charset/gbx/gbx1.txt Both works fine. Tested 2000032417 US beta1 build.