Re: Patch for cpplib.pot generation

[Zack Weinberg <zack at codesourcery dot com>]

"Joseph S. Myers" <joseph@codesourcery.com> writes:

> The instructions in translation.html for regenerating cpplib.pot, and
> the corresponding rules in libcpp/Makefile.in, look rather untested:
>
> * There is no cpplib.pot target, only a po/cpplib.pot one.
>
> * The cpplib.pot in CVS uses file names from before cpplib was moved
> to its own directory.  Also, they don't include a srcdir path but the
> rule does nothing to exclude one from the file.
>
> * The Makefile.in rule does not generate an output file because the
> --keyword options to xgettext use "," instead of the ":" which at
> least xgettext 0.14.1 expects.
>
> * With that fixed, the messages from the _, N_, SYNTAX_ERROR and
> SYNTAX_ERROR2 macros - which existed before the move - still aren't
> extracted; those keywords need adding as well.
>
> I tried regenerating cpplib.pot and found those issues.  Although I'm
> happy to deal with the .pot file regeneration and submission to the TP
> (item 11 of the branching checklist) when 4.0 branches, having been
> dealing with TP matters lately (unless the i18n maintainer wishes to
> take over these matters again), it seems better for them to be fixed
> now rather than waiting for discovery when the branching checklist is
> being followed.
>
> This patch fixes those issues.  The fix to stop srcdir paths getting
> in follows that in my unreviewed exgettext patch
> <http://gcc.gnu.org/ml/gcc-patches/2004-11/msg00451.html>.
>
> Bootstrapped with no regressions on i686-pc-linux-gnu, and
> successfully regenerates cpplib.pot.  OK to commit?  (I'd regenerate
> cpplib.pot after committing.)

This is OK, and I also approve the exgettext patch which you
reference.

zw