Re: msgmerge and the handling of obsolete entries (Re: Updating pofiles on make dist)



Good point. And while we're at it, could we have msgmerge please NOT 
change the format of the po-entries, i.e. introducing new line-breaks? 
Sometimes the line-breaking I do has a particular meaning. And since we 
know it's supposed to be a single line anyway, msgmerge sometimes simply 
treats it as is and breaks the line where it wants, rather than where I 
want. Maybe we could have such option in msgmerge, but -- as default -- 
leave the line-breaks where they are? Would that be possible? Comments? 
Yes? No? If not, why not? Has it been done already? Anything?

Cheers,
Bernd


Karl Eichwalder schrieb:

>Sven Neumann <sven@gimp.org> writes:
>
>  
>
>>I was expecting that the fuzzy merger would do a better job when being
>>run frequently. I am not sure since I didn't do any tests but I would
>>expect more helpful fuzzy matches if the po file is synced with the
>>changes from time to time.
>>    
>>
>
>The problem is: sync'ing will produce many a lot fuzzy messages and
>msgmerge will never try to assign a better matching translation to such
>a fuzzy marked message.  This makes it difficult for the translator to
>make use of an own catalog file (AKA translation memory).
>
>Also morphing of obsolete, but valid translations can happen; once a
>obsolete message is used for a (fuzzy) match it will go away from the
>.po file silently:
>
>ke@tux:/tmp> cat test.de.po
>#~ msgid "the message 'foo'"
>#~ msgstr "die Meldung 'foo'"
>
>ke@tux:/tmp> cat test.pot
>msgid "the message"
>msgstr ""
>
>ke@tux:/tmp> msgmerge -o - test.de.po test.pot 
>.. done.
>#, fuzzy
>msgid "the message"
>msgstr "die Meldung 'foo'"
>
>Bug or feature, I don't know ;)  I'm incline to believe 'msgmerge' must
>not throw away any obsolete entry.  I guess this is more an academic
>problem than something that hurts in real life.
>
>  
>


-- 
Disclaimer: http://apac.redhat.com/disclaimer





[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]