Subject: Re: Patch: Abiword Bugs 1090 & 1166
From: Andrew Dunbar (hippietrail@yahoo.com)
Date: Mon Jun 25 2001 - 01:32:31 CDT
Vlad Harchev wrote:
>
> On Mon, 25 Jun 2001, Andrew Dunbar wrote:
>
> Hi Andrew!
Hi Vlad. I'm breaking my prime directive of never replying at the
top of a page (: I looked further into these problems and it turns
out that the .strings files work 100% fine but the .strings file
does not contain the menu strings! They are hard-coded in the
source without declaring their encodings. They are assumed to be
in the correct encoding to match the locale at runtime. This
breaks.
So is there a good reason we separate localized menus from the rest
of the localized strings?
Andrew.
> > 1166 on Windows says Russian strings are in KOI8-R but need to be
> > in CP1251.
> > The ru-RU.strings file declares KOI8-R and really is in KOI8-R.
> > This one should be working.
>
> This shouldn't matter - since I remember adding the code that converts text
> from charset specified in .strings file to the one returned by
> EncodingManager.
> If the bug 1166 is opened for 0.7.14 (latest release) and not for current
> CVS, then everything was possible on Windows since EncodingManager always
> returned "ISO8859-1" as a system's encoding (in fact, this made all russian
> strings in dialogs invisible (!) since no russian characters are in
> iso-8859-1, at least on win95). Thanks to your work, EncodingManager now
> correctly returns system's charset and everything should work fine under
> windows without recoding data in .strings file
>
> It would be nice to check the correctness, could you please send screenshot
> of any (perferably small - such as "find") dialog with any russian string
> visible to the person who opened #1166 or me and ask whether russian is in
> proper encoding on the screenshot. Or another option (if you don't have the
> russian locale for windows installed) is to place for downloading zipped
> abiword.exe (I hope that person who opened #1166 will check it - I can't
> afford downloading such quantities of data).
>
> > My understanding of Abi is that the .strings files can be in any
> > (currently 8 bit) encoding. Abi reads them in at startup and
> > converts them to Unicode. At display time Abi converts them from
> > Unicode to the system's encoding. So the source encoding should
> > be irrelevant.
>
> Exactly. But AFAIR they were being converted to locale's charset, not
> unicode.
>
> > I don't speak Russian or Czech but I'm 99% confident they
> > both render 100% correctly in Windows with the appropriate locale.
> > Please check whether these bugs are fixed under both OSes.
> > If not, please post an example in *Unicode* showing the correct
> > string and the string as you see it so that I can be sure what
> > I'm looking for (:
>
> Hmm, I checked recoding of .strings many times on unix - it works perfectly.
> Only windows is affected and needs to be checked.
>
> Thanks!
>
> Best regards,
> -Vlad
-- http://linguaphile.sourceforge.net _________________________________________________________ Do You Yahoo!? Get your free @yahoo.com address at http://mail.yahoo.com
This archive was generated by hypermail 2b25 : Mon Jun 25 2001 - 01:30:34 CDT