Am Dienstag, 14. Oktober 2014, 09:03:16 schrieb Burkhard L=FCck:
> Am Montag, 15. September 2014, 20:10:58 schrieb Albert Astals Cid:
> > El Dilluns, 15 de setembre de 2014, a les 19:52:08, Luk=C3=A1=C5=A1=
 Tinkl va
>=20
> escriure:
> > > Dne 15.9.2014 v 19:26 Albert Astals Cid napsal(a):
> > > > El Dilluns, 15 de setembre de 2014, a les 13:06:31, Chusslove I=
llich
> > > > va
> > > >=20
> > > > escriure:
> > > >>> -> lupdate mangles/breaks the context.
> > > >>>=20
> > > >>> WONTFIX on Qt side (fixing lupdate)
> > > >>>=20
> > > >>> Solution: directly extract po from Qt frameworks, so the cont=
ext are
> > > >>> correct; requires few changes on the infrastructure
> > > >>=20
> > > >> This sounds strange. Is it related to
> > > >> https://bugreports.qt-project.org/browse/QTBUG-40444?
> > > >=20
> > > > Don't know, tbh i never understood; the problems of such things=
 being
> > > > explained with words instead of with examples on the command li=
ne :D
> > > >=20
> > > > Lukas can you give us the expamples of what is really wrong?
> > > >=20
> > > > Cheers,
> > > >=20
> > > >    Albert
> > > >>=20
> > > >> If so, I see Oswald's
> > > >> comments as correct, including the need-more-info resolution.
> > > >>=20
> > > >> For cases where PO is just an intermediate format, Gettext cal=
l
> > > >> conventions
> > > >> do not apply (e.g. that context in PO must match the context i=
n
> > > >> code).
> > > >> Any
> > > >> conventions are instead established by the conversion tool. So=
 the
> > > >> only
> > > >> relevant thing is whether installing the final TS file (after
> > > >> TS->PO->TS
> > > >> process) results in seeing all translations in the program. If=
 it
> > > >> doesn't,
> > > >> then something is broken in the PO->TS part of lconvert, and n=
ot in
> > > >> lupdate
> > > >> or the TS->PO part of lconvert.
> > >=20
> > > I explained the problem and its symptoms (directly with examples)=
 in the
> > > above bug report.
> >=20
> > Ok, i didn't know if that was the bug or not, so reading the bug i =
see two
> > things you mention.
> >=20
> > * The msgctxt has to match the source code context exactly and the =
comment
> > has to be put before the entry as a plain comment.
> >=20
> > =09I do not agree with this, anywya seems more a stylistic thing th=
an a real
> >=20
> > problem.
> >=20
> >=20
> > * When converting such PO file back to TS with lconvert, the contex=
t is
> > lost. That doesn't happen here, i'm attaching my kcompletion5_qt.po=
 and
> > the
> > ts genereated with lconvert. The ca.ts is fine, no?
>=20
> But if I convert l10n-kf5/ca/messages/frameworks/kcompletion5_qt.po t=
o the=20
> a ts file I get an entry like:
>     <message>
>         <location filename=3D"klineedit.cpp" line=3D"1125"/>
>         <source>Text Completion</source>
>         <comment>KLineEdit|@title:menu</comment>
>         <translation>Compleci=F3 del text</translation>
>     </message>
>=20
> Whereas the your ca.ts has:
>     <message>
>         <location filename=3D"klineedit.cpp" line=3D"1125"/>
>         <source>Text Completion</source>
>         <comment>@title:menu</comment>
>         <translation type=3D"unfinished"></translation>
>     </message>
>=20
> There are different strings in <comment></comment>:
> "KLineEdit|@title:menu" !=3D "@title:menu"
>=20
> Attached two translation catalogs for kitemviews5_qt.
>=20
> 1) kitemviews5_qt.ts
> extracted with lupdate from sources, filled up with the german transl=
ations
> and installed with lconvert to a qm file
> ->  "Search:" in kdebugdialog5 is translated
>=20
> 2. kitemviews5_qt.po-de2ts.ts
> that is the translated german po file converted with lconvert to a ts=
 file
> and then installed with lconvert to a qm file
> ->  "Search:" in kdebugdialog5 is *NOT* translated
>=20
> Both ts files have obviously different translation entries.

Locale x-test really rocks :-)

I installed x-test in the folder kf5/share/locale/zu/ and found the sim=
ple fix=20
for this issue, a missing line in our po file headers:

$ svn diff l10n-kf5/de/messages/frameworks/sonnet5_qt.po=20
Index: l10n-kf5/de/messages/frameworks/sonnet5_qt.po
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
--- l10n-kf5/de/messages/frameworks/sonnet5_qt.po       (Revision 14029=
67)
+++ l10n-kf5/de/messages/frameworks/sonnet5_qt.po       (Arbeitskopie)
@@ -22,6 +22,7 @@
 "Content-Type: text/plain; charset=3DUTF-8\n"
 "Content-Transfer-Encoding: 8bit\n"
 "X-Generator: Lokalize 1.5\n"
+"X-Qt-Contexts: true\n"
 "Plural-Forms: nplurals=3D2; plural=3Dn !=3D 1;\n"
=20
 #: core/loader.cpp:149

--=20
Burkhard L=FCck