Re: [Monotone-devel] broken i18n : xgettext options

[Nathaniel Smith]

On Fri, Dec 02, 2005 at 04:20:04PM +0100, Benoît Dejean wrote:
> Hi,
>       i don't know if my system is broken, but xgettext no longer gets
> configured to scan for F()/FP().
> 
> LC_ALL=C intltool-update -x fr
> Working, please wait...Building monotone.pot...
> Running /usr/bin/xgettext --add-comments --directory=.
> --output=monotone.pot --files-from=./POTFILES.in.temp --keyword=_
> --keyword=N_ --keyword=U_ --keyword=Q_ --from-code=ASCII
> Removing generated header (.h) files...done.
> Wrote monotone.pot
> Merging ./fr.po with monotone.pot...................... done.
> 112 translated messages, 9 fuzzy translations, 67 untranslated messages.
> 
> Someone please fix this, because this unmark many many i18n strings :/

This seems to have been broken in jmmv's commit
   d89f5514bb5446463549ba42ced34964de15317

intltool-update appears to open po/Makevars (or else falling back on
po/Makefile.in.in, but we have a po/Makevars so that doesn't matter),
and apply some regexpy gunk to figure out what the value of the
XGETTEXT_OPTIONS variable is.  In that commit, we change things so
that at configure time, we write out the XGETTEXT_OPTIONS variable to
po/Makefile.in, which is pretty useless from intltool-update's point
of view.

The reason for this change was to support running xgettext on systems
where xgettext does not support --flag.  (We use --flag to enable
gettext's C format string checking code for our boost formatters, to
make sure translators don't mix up their %-placeholder types and
such.)  I'm not sure what such systems exist, or why we need to update
our template files on them... Julio?

-- Nathaniel

-- 
"But in Middle-earth, the distinct accusative case disappeared from
the speech of the Noldor (such things happen when you are busy
fighting Orcs, Balrogs, and Dragons)."