You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by Erik Huelsmann <e....@gmx.net> on 2004/05/03 08:49:47 UTC
Using bind_textdomain_codeset() after all?
>
> >If the encoding of the po file (say UTF-8) does not match the active
> >locale (say iso-8859-1) then at least GNU gettext will translate the
> >UTF-8 from the po to iso-8859-1 in the presence of a 'charset=' string.
> >When this happens the output of *gettext() won't be
> >the required UTF-8...
> >That's why I don't see any other way than preventing gettext to know
> >what to translate from. You are ofcourse free to temporarily add them,
> >but they can't be in the Subversion repository or the 'compiled' mo...
> >
>
> I'm saying this:
>
> Implementations which convert charset and support
> bind_textdomain_codeset():
> so we call bind_textdomain_codeset() and everything is OK.
>
> Implementations which do not convert charsets and do not suppoer
> bind_textdomain_codeset():
> so we make all the .po files UTF-8 encoded.
>
> Is there an implementation which convert between charsets and do not
> support bind_textdomain_codeset()? Only if such an implementation exists
> it should be required to strip the encoding specification from the .po
> files, IMO.
I don't know any such system, but only have GNU gettext available and know
about the existence of the Solaris one. Although I have tried to find out I
have no idea if this completely describes the problem space. (Usually that
means it doesn't :-)
bye,
Erik.
--
"Sie haben neue Mails!" - Die GMX Toolbar informiert Sie beim Surfen!
Jetzt aktivieren unter http://www.gmx.net/info
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Using bind_textdomain_codeset() after all?
Posted by Branko Čibej <br...@xbc.nu>.
Erik Huelsmann wrote:
>(although this is starting to be a discussion with myself)
>
>There is one more option: Rip out the 'charset=' at build time. Can that be
>done in the Windows build?
>
>
Anything "can be done" in the Windows build if we assume Python is
available. Which we do. :-)
-- Brane
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Using bind_textdomain_codeset() after all?
Posted by Erik Huelsmann <e....@gmx.net>.
>
> I forget to add for those falling in the middle that Nicolas argues that
> the
> requirement for the 'charset=' string also stems from po editors depending
> on it to determine which charset to save the file in.
>
> Emacs po mode does seem to do just that...
>
> > > >If the encoding of the po file (say UTF-8) does not match the active
> > > >locale (say iso-8859-1) then at least GNU gettext will translate the
> > > >UTF-8 from the po to iso-8859-1 in the presence of a 'charset='
> string.
> > > >When this happens the output of *gettext() won't be
> > > >the required UTF-8...
> >
> > > >That's why I don't see any other way than preventing gettext to know
> > > >what to translate from. You are ofcourse free to temporarily add
> them,
> > > >but they can't be in the Subversion repository or the 'compiled'
> mo...
> > > >
> > >
> > > I'm saying this:
> > >
> > > Implementations which convert charset and support
> > > bind_textdomain_codeset():
> > > so we call bind_textdomain_codeset() and everything is OK.
> > >
> > > Implementations which do not convert charsets and do not suppoer
> > > bind_textdomain_codeset():
> > > so we make all the .po files UTF-8 encoded.
> > >
> > > Is there an implementation which convert between charsets and do not
> > > support bind_textdomain_codeset()? Only if such an implementation
> exists
> > > it should be required to strip the encoding specification from the .po
> > > files, IMO.
> >
> > I don't know any such system, but only have GNU gettext available and
> know
> > about the existence of the Solaris one. Although I have tried to find
> out
> > I
> > have no idea if this completely describes the problem space. (Usually
> that
> > means it doesn't :-)
(although this is starting to be a discussion with myself)
There is one more option: Rip out the 'charset=' at build time. Can that be
done in the Windows build?
bye,
Erik.
--
"Sie haben neue Mails!" - Die GMX Toolbar informiert Sie beim Surfen!
Jetzt aktivieren unter http://www.gmx.net/info
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Using bind_textdomain_codeset() after all?
Posted by Erik Huelsmann <e....@gmx.net>.
I forget to add for those falling in the middle that Nicolas argues that the
requirement for the 'charset=' string also stems from po editors depending
on it to determine which charset to save the file in.
Emacs po mode does seem to do just that...
> > >If the encoding of the po file (say UTF-8) does not match the active
> > >locale (say iso-8859-1) then at least GNU gettext will translate the
> > >UTF-8 from the po to iso-8859-1 in the presence of a 'charset=' string.
> > >When this happens the output of *gettext() won't be
> > >the required UTF-8...
>
> > >That's why I don't see any other way than preventing gettext to know
> > >what to translate from. You are ofcourse free to temporarily add them,
> > >but they can't be in the Subversion repository or the 'compiled' mo...
> > >
> >
> > I'm saying this:
> >
> > Implementations which convert charset and support
> > bind_textdomain_codeset():
> > so we call bind_textdomain_codeset() and everything is OK.
> >
> > Implementations which do not convert charsets and do not suppoer
> > bind_textdomain_codeset():
> > so we make all the .po files UTF-8 encoded.
> >
> > Is there an implementation which convert between charsets and do not
> > support bind_textdomain_codeset()? Only if such an implementation exists
> > it should be required to strip the encoding specification from the .po
> > files, IMO.
>
> I don't know any such system, but only have GNU gettext available and know
> about the existence of the Solaris one. Although I have tried to find out
> I
> have no idea if this completely describes the problem space. (Usually that
> means it doesn't :-)
>
> bye,
>
>
> Erik.
>
> --
> "Sie haben neue Mails!" - Die GMX Toolbar informiert Sie beim Surfen!
> Jetzt aktivieren unter http://www.gmx.net/info
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
> For additional commands, e-mail: dev-help@subversion.tigris.org
>
--
NEU : GMX Internet.FreeDSL
Ab sofort DSL-Tarif ohne Grundgeb�hr: http://www.gmx.net/dsl
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org