[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Build: another hack for translations (fix 1323). (issue2520041)
From: |
John Mandereau |
Subject: |
Re: Build: another hack for translations (fix 1323). (issue2520041) |
Date: |
Mon, 18 Oct 2010 12:58:47 +0200 |
AFAICT after a few hours of fiddling this issue is fixed, let's wait for
release to verify it (and not hopefully reopen it :-P). Let me know how
is current master. I've built and have checked both online and offline
targets, both split and big-page manuals, both docs in English and
Italian (so I've checked 8 output files). You may ignore details below
to save time.
Il giorno lun, 18/10/2010 alle 00.35 +0100, Graham Percival ha scritto:
> Ok. A few things spring to mind:
> - maybe the DEPTH environment variable isn't always set to the
> depth of the git source tree? I mean, that would be profoundly
> messed up, but this might explain why the offline docs
> (normally, before today at least) work? I mean, maybe the
> GNUmakefile in the translation pass in the depth-1 ?
It could be, but but it's not the case, see
find -name 'GNUmakefile*' -or -name '*.make' |xargs grep -B3 -n --color DEPTH
> - maybe there's something else in postprocess.py (or some related
> file) which automatically removes a ../ layer from the
> translations? I think this is much more likely. But if this
> is true, then fixing the "my $reldir" line in the texi2html
> init is going to make the later fix produce broken links.
AFAICS it's no longer the case in current master branch. I suspect that
there used to be a kind of something you decribe that used to caught the
URLs for "back to documentation index", but I can't find it in the
makefiles or *.py files in current master Git tree (I grepped for
'\.\.').
Cheers,
John
signature.asc
Description: This is a digitally signed message part