Build: 03-14 trunk build Steps: 1. Load a page that contains surrogate characters, e.g.: http://warp/u/ftang/utf8test/gb18030.cgi?page=596 2. Highlight a character and copy/paste into MS WinXP word. Result: Will be pasted as a small square by default(original source format). A work around is paste as text only. (there are 3 options, only this one will work) IE can be pasted in all 3 options. Also with gb2312 or non-surrogate 18030 characters can be pasted as original source format.
I think we somehow does not generate NCR correctly for surroagte, that may cause the problem. reassign to ftang for now.
give to shanjian. I think this is a serialization issue. We save into a wrong &#; NCR (two NCR for one surrogate pair instead of one). I think it is a dup of one bug already assign to shanjian This is NOT a high priority item. But we should fix it by Q4 2002
if frank's guess is true, this is most likely a duplicate of one of my existing bug. I have no way to verify it because I don't have the latest word software. I am aware of the problem mentioned by frank and it should have been fixed in my local tree.
this might be dup of 137657
I want to see we solve all surrogate editing problem by m1.2final. mark it as nsbeta1+ and tm=m1.2beta
I didn't see any problem with my local build. I will keep it open until I check in my other surrogate patch.
batch: adding topembed per Gecko2 document http://rocknroll.mcom.com/users/marek/publish/Gecko/Gecko2Tasks.html
It works fine on build 2002100409, after those surrogate conversion patches checked in.
This actually still exising in 06-10 1.4 branch build, re-open...
Original demonstration page has disapeared. I did some test, but unfortunately without the needed fonts so I only saw questions marks. With http://bugzilla.mozilla.org/attachment.cgi?id=67079&action=view All three paste method are problematic. Raw and unicode give twice more question mark than the number of selected character. HTML insert gives one question mark less than the number of character copied if 3 or more character were selected. When selecting only one or two character, I get nothing. With http://www.unicode.org/charts/collation/chart_Cypriot.html HTML seems correct. Raw and unicode still give twice more question mark than the number of selected character.
shanjian is no longer working on mozilla for 2 years and these bugs are still here. Mark them won't fix. If you want to reopen it, find a good owner first.
Perhaps, depends on bug 239279
I don't have XP Word here to test, but as far as I can tell from the clipboard viewer this may be WORKSFORME. Note that an unrelated bug (bug 120114) affecting Cypriot characters was fixed a little while ago.
All 3 paste methods seem correct in current trunk.