Date: prev next · Thread: first prev next last
2011 Archives by date, by thread · List index



On Tue, 2011-06-21 at 13:13 +0200, Andras Timar wrote:
       Instead of calling textdomain - since (I suspect) we can probably call
lots of these resource mgrs in quick succession - eg. initializing an
sw/ widget derived from sfx2/ from vcl/ - I suggest we drop the
textdomain() call - and store a const char *pDomain; that is a UTF8 copy
of aPrefix (on InternalResMgr if I see right).

As far I can tell from the OSL_TRACE messages, textdomain() is called
only once for each resource domain. I think resources are read when
needed and they are stored in memory until application closes.

        textdomain sets some sort of global variable (last I looked) which
subsequent gettext calls then work from. That shouldn't work well in
LibreOffice's case where we need to multiplex lots of catalogs at speed;
I strongly recommend dgettext :-)

Oh, yes, let's discuss this now because it affects design. I think
Bjoern was thinking about Linux only so they could release translation
updates for LibreOffice in Ubuntu without rebuilding the source.

        Which makes lots of sense for translators.

I thought it was a multi platform feature because all platforms could
benefit from separation of localization and building.

        Well - we still need to build the english resource files, and then load
and lookup the strings (as english) and then do the gettext-iness on
them, so this is going to be rather slow I suspect. The .po files are
also rather bigger than the .res files I think - since they need both
english and translated strings in them. So this is not clear to me; lets
discuss it at the next TSC meeting.

       Clearly we want to special case it only for the case
that there is no native translation.

What do you mean by that? Fallback to gettext when no native resources
are installed? Yes, this way we can add a language to an existing
installation. Is that worth the effort?

        Right.

Everywhere means 2 libs only: tl and the other one which
processes .xcd files.

        But you forget how evil cut+paste coding is; it is the moral equivalent
of eating your children ;-)

        ATB,

                Michael.

-- 
 michael.meeks@novell.com  <><, Pseudo Engineer, itinerant idiot



Context


Privacy Policy | Impressum (Legal Info) | Copyright information: Unless otherwise specified, all text and images on this website are licensed under the Creative Commons Attribution-Share Alike 3.0 License. This does not include the source code of LibreOffice, which is licensed under the Mozilla Public License (MPLv2). "LibreOffice" and "The Document Foundation" are registered trademarks of their corresponding registered owners or are in actual use as trademarks in one or more countries. Their respective logos and icons are also subject to international copyright laws. Use thereof is explained in our trademark policy.