On Mon, 20 Aug 2001, Thomas Diehl wrote:
I have to leave that to the people responsible for CVS and kcmlocale matters (ie to Coolo and to Hans Petter). Maybe they can comment on this.
We will do this after KDE 2.2.1. Even if you changed to UTF-8 in KDE 2.2.1, it will not be smart to rename the translation because it means that the user would have to change the language in kcmlocale again.
And changing the .po file encoding isn't really smart unless it's a major release. The reason is that some third part applications may be shipped with the translations, and they should still work after you upgrade.
Personally, I suspect that we won't probably do that for 2.2.1 since this should only be a bug fix release with minimal code changes. (I'm also no sure how this naming scheme would break the utf-8 conversion?)
It wouldn't break it. And the UTF-8 change is for HEAD only. It should still be supported in KDE 2.2.x.
After 2.2.1 we will probably have a lot of renaming since we are aiming at a conversion to 3-letter codes. This would be a better occasion to find a solution for the naming of traditonal Chinese as well.
Right.
Norwegian University of Science and Technology, Engineering Cybernetics bieker@stud.ntnu.no / bieker@kde.org / +47 97651633 Dagaliveien 25 A | Elgesetergate 23 B N-0776 Oslo | N-7030 Trondheim Siste nytt fra industrien finner du på: WWW.Industri-Informasjon.NO
Am Dienstag, 21. August 2001 10:23 schrieb Hans Petter Bieker:
On Mon, 20 Aug 2001, Thomas Diehl wrote:
I have to leave that to the people responsible for CVS and kcmlocale matters (ie to Coolo and to Hans Petter). Maybe they can comment on this.
We will do this after KDE 2.2.1. Even if you changed to UTF-8 in KDE 2.2.1, it will not be smart to rename the translation because it means that the user would have to change the language in kcmlocale again.
And changing the .po file encoding isn't really smart unless it's a major release. The reason is that some third part applications may be shipped with the translations, and they should still work after you upgrade.
I wouldn't propose switching to utf-8 for KDE 2.2.1 either. But I got worried about leaving the PO files in big5 since the traditional Chinese translation of KOffice was dropped during packaging. And apparently due to encoding problems. See the attached message.
Regards,
Thomas
Re: Errors when building koffice-i18n Datum: Mon, 23 Jul 2001 10:32:36 -0500 Von: David Faure david@mandrakesoft.com An: Stephan Kulow coolo@kde.org Kopie: thd@kde.org
On Mon, Jul 23, 2001 at 10:46:50AM +0200, Stephan Kulow wrote:
On Thursday, 19. July 2001 14:17, you wrote:
Hello, packaging RC1 I'm hitting the same errors/warnings :) More below....
On Wednesday 16 May 2001 23:27, Stephan Kulow wrote:
On Wednesday 16 May 2001 20:21, David Faure wrote:
*** Building in ru *** Building in sk *** Building in sl *** Building in sv *** Building in tr *** Building in zh_TW.Big5 ./messages/koffice/koshell.po: warning: Charset "utf8" is not a portable encoding name. Message conversion to user's charset
might not
work. ./messages/koffice/kspreadcalc_calc.po: warning: Charset
"utf8"
is not a portable encoding name. Message conversion to user's
charset
might not work. ./messages/koffice/kword.po:1986: invalid
multibyte
sequence ./messages/koffice/kword.po:1987: end-of-line within
string
found 2 fatal errors ./messages/koffice/kword.po:1966: invalid multibyte sequence ./messages/koffice/kword.po:1967: end-of-line within string found 2 fatal errors ./messages/koffice/kchart.po:107: invalid control sequence ./messages/koffice/kchart.po:173: invalid control sequence found 2 fatal errors ./messages/koffice/koffice.po:1119: invalid control sequence found 1 fatal error ./messages/koffice/kspread.po:602: invalid control sequence ./messages/koffice/kspread.po:603: end-of-line within string ./messages/koffice/kspread.po:1425: end-of-line within string ./messages/koffice/kspread.po:1481: invalid control sequence ./messages/koffice/kspread.po:4697: invalid control sequence ./messages/koffice/kspread.po:4698: end-of-line within string ./messages/koffice/kspread.po:4701: invalid control sequence ./messages/koffice/kspread.po:4707: invalid control sequence [...] too many errors, aborting
Any idea why this happens ? Since you know the sequence... is that related to zh_TW.Big5,
or could
that be any other one ? Is this a big problem ?
You're using gettext 0.10.36 - asking for trouble.
gettext 0.10.38 now. Better or still too old ?
Too young! We're only supporting gettext 0.10.35 :( They added an extra check to their msgfmt, which heavily breaks
compatibility
to gettext 0.10.35. The message about utf8 is just a warning without
any
consequence, but all that Big5 stuff is broken - I told the
translator to
switch to utf8, but he refused so far :(
I see. Hmm, there's not much point in me releasing zh_TW.Big5 then. I can't really revert the gettext version on that machine, it's a Mandrake server. One day I'll get ADSL (hopefully), but since then I need a fast-connected server. Well, maybe this is one more very good reason why he should switch to utf8. Up until then, his (at least koffice) translations are useless.....
-
Hans Petter Bieker -
Thomas Diehl