Moz needs to support Indic/Armenain/Georgian text input in Windows 2000 and WinXP. We may be able to use WM_UNICHAR

#define WM_UNICHAR 0x0109

batch: adding topembed per Gecko2 document http://rocknroll.mcom.com/users/marek/publish/Gecko/Gecko2Tasks.html

cc shanjian and kin

shanjian: can you review?

With the patch and MOZ_UNICODE flag on, we can support Indic/Armenain/Georgian. (tested in W2K-en with Hindi, Armenain and Georgian keyboard layout.)

So I'm curious, is there some reason why the nsToolkit function pointers are only defined with MOZ_UNICODE on? If we made them active, even when MOZ_UNICODE *isn't* defined, and point by default to the function versions we currently use when MOZ_UNICODE is not defined, we can avoid these ifdefs that are being added everywhere right?

>we can avoid these ifdefs Good point. I never thought of that. Technically, we should be able to do the function pointers without the use of MOZ_UNICODE. I think I was conservative since the impact of changes were very high (ie. all the windows msgs are affected) shanjian/kin: Can you review / super review for the patch or do you want me to remove the ifdef's this time? I kinda want to keep them. _being lazy_ :)

Comment on attachment 99396 [details] [diff] [review] define and use nsToolkit::mIsNT instead of mUseImeApiW to reflect real use r=shanjian, I am ok with your existing code. When do you plan to make MOZ_UNICODE default? Once that is done, I guess we can remove all those ifndef MOZ_UNICODE code. Of cause, after it stablizes for a while.

>When do you plan to make MOZ_UNICODE default? Yes, very soon. As a matter of fact, I'm in the process of getting the patch together and making sure other platform won't break. I also run performance tests (startup and pageload) on both Win-XP and Win ME and they look very promising. See bug 104934 for more info kin: can you super review?

we need this resolve to make use in parity to IE6 for the new market. Without this fix, we look very behind of IE in turn of window intergration.

Comment on attachment 99396 [details] [diff] [review] define and use nsToolkit::mIsNT instead of mUseImeApiW to reflect real use sr=kin@netscape.com

checked into the trunk. sujay: Please wait until we get the Unicode version of Mozilla to verify. (see 104934)

i was testing the unicode build 10-18-15 and i was able to input armenian, georgian and hebrew into the URl bar, compose window and forms, also i've noticed that hebrew look much different in forms than in other places.

>hebrew look much different in forms than in other places. marina: Is it unique to the Unicode build or in the regular trunk as well?

cc simon

*** Bug 120514 has been marked as a duplicate of this bug. ***