From: Richard L. <rl...@wi...> - 2006-02-14 16:21:53
|
On Tue, 2006-02-14 at 15:25 +0100, =D0=94=D0=B0=D0=BD=D0=B8=D0=BB=D0=BE =D0= =A8=D0=B5=D0=B3=D0=B0=D0=BD wrote: > =D0=88=D1=83=D1=87=D0=B5 =D1=83 21:51, Luke Schierer =D0=BD=D0=B0=D0=BF= =D0=B8=D1=81=D0=B0: >=20 > > I strongly agree with Ethan here. No reason why they should have to > > learn UTF-8 when we can automatically convert the encoding at compile > > time. >=20 > You actually can't (or not easily, anyway): intltool is reencoding > only into it's own cache so it can construct properly merged .desktop, > .xml, .schemas, .server and similar files which contain many > languages. Encoding ends up in the MO file as well, and translation > needs to be re-encoded on every run of the program in that locale. Why not just have an intltool-msgfmt script which checks the encoding, and if it's not UTF-8, converts it before passing the contents of the .po to msgfmt? Richard |