This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Broken translations
- From: "Joseph S. Myers" <joseph at codesourcery dot com>
- To: Translation Project Robot <translation at iro dot umontreal dot ca>
- Cc: gcc-patches at gcc dot gnu dot org, murf at e-tools dot com, ndandali at yahoo dot fr, muvia1 at yahoo dot fr, s24211045 at tuks dot co dot za, karemacarole at hotmail dot com, ngenda_denis at yahoo dot co dot uk, akiberwa at yahoo dot co dot uk, ndonatienuk at yahoo dot co dot uk, antoine at e-tools dot com, translation-team-rw at lists dot sourceforge dot net
- Date: Wed, 6 Apr 2005 00:53:52 +0000 (UTC)
- Subject: Broken translations
- References: <20050405135434.CBE2415E06A@bor.iro.umontreal.ca>
[subject changed to avoid spam block which these messages got because of
their volume]
I have committed the 4.0-b20050226 versions of the Kinyarwanda
translations for gcc and cpplib to mainline and 4.0 branch. I am puzzled
at the large number of messages sent lately with translations for
different GCC versions; I could understand translators who started several
years ago translating 3.0 and only updated to a more recent version when
it was finished, but not the utility of producing translations for every
different gcc.po version, 3.0 onwards, *none of them having more than 7
non-fuzzy translations out of thousands of messages*. (And, similarly,
the current cpplib translation has zero non-fuzzy translations.) When
there are so many messages with translations of different versions there's
liable to be confusion about which version if any should be committed;
whereas if translators are sending several updates a day of translations
for the most recent version, at least I know I can ignore all but the most
recent message when committing to CVS.
I have not committed anything to 3.4 branch or 3.3 branch and given the
tiny numbers of non-fuzzy translations I doubt the utility of doing so.
The messages include many which are completely garbled, contain none of
the % format specifiers in the original messages, map many diagnostics to
the same text, and have other problems. I don't really believe that "ku"
is the Kinyarwanda translation of "bb %d on wrong place", "Align some
doubles on dword boundary", " by %qD", "Disable optimizations observable
by IEEE signaling NaNs" and many other messages, for example. Could the
translation team please explain what is going on with this translation?
If there was a four-day delay between submission of these messages to the
TP and them being sent out on 5 April, could the TP messages include the
date of submission to the TP in future to avoid such confusion with
genuine translations?
--
Joseph S. Myers http://www.srcf.ucam.org/~jsm28/gcc/
jsm@polyomino.org.uk (personal mail)
joseph@codesourcery.com (CodeSourcery mail)
jsm28@gcc.gnu.org (Bugzilla assignments and CCs)