Hi, I am not sure that I can help with libwpg problem but I can with the /po files one. Maybe only a quick question, did some of previous revisions compile well with libwpg? (rev. 12993 has only updated gl.po file)
pbhj píše v Čt 31. 08. 2006 v 04:11 +0100:
[trimmed libwpg error part ]
Ultimately I still can't get past the po files:
file=./`echo ja | sed 's,.*/,,'`.gmo \
&& rm -f $file && /usr/bin/msgfmt -c -o $file ja.po
ja.po:15: nplurals = 1...
ja.po:899: ...but some messages have 2 plural forms
/usr/bin/msgfmt: found 1 fatal error
make[2]: *** [ja.gmo] Error 1
make[2]: Leaving directory `/usr/local/svn/inkscape/po'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/usr/local/svn/inkscape'
make: *** [all] Error 2
I've found one other mail ( bug 1343164 @ https://sourceforge.net/tracker/index.php?func=detail&aid=1343164&gr... ) on the mailing list that seems to refer to this type of situation but there appears to be no resolution. I too can't see why I have to compile all these languages that I wont need.
That is the cost of having multi-lingual software. Sure you can remove them from makefile, but in result, the compile time you save is only 0.1% of total compile time.
I've removed problematic plural forms from the ja.po file and going to contact last translator of this file to inform him about it. Now the file should work for you too. But I've got a backup for any other cases. Please do "svn up" and report me back whether the translation's compilation works or not.
Moreover as grammar doesn't change on a per project basis I don't see how these parts (the plural forms) can be wrong?
This problem occurs when the plural forms definition does not matches the number of msgstr[x] string in translations. In case of Japanese translation there were two msgstr[] definitions on some places but only one was supposed to be there(only one according to pl. forms definition). This can usually happen when translator does not watch his steps and does not check possible plural forms' errors. If the string that caused this error has still been marked as "fuzzy" then you probably could not even notice that something like this is there. Weird is that it did not cause any build problems on several machines I've tried to compile it.
Thanks for the report, sincerelly, Josef 'cornelius' Vybíral (cwebb)