pretty-ipa merge 3 (well lto merge ;): inline_failed reason as enum

Joseph S. Myers joseph@codesourcery.com
Sun Mar 29 01:20:00 GMT 2009


On Sat, 28 Mar 2009, Steven Bosscher wrote:

> On Sat, Mar 28, 2009 at 12:16 PM, Jan Hubicka <hubicka@ucw.cz> wrote:
> >> [ xf. http://gcc.gnu.org/ml/gcc-patches/2009-03/msg01444.html ]
> >>
> >> The new file cif-code.def is missing from the posted patch.
> >>
> >> How does this interact with i18n, i.e. how are the messages from the
> >> .def file supposed to end up in the .pot/.po?
> >>
> >> Ciao!
> >> Steven
> >
> > I am attaching the file.  Sorry for that.
> > I didn't really think of .pot generation, is there way to say that
> > even .def should be scanned?
> 
> The .def files are always scanned too (see e.g. diagnostics.def and
> params.def).  But I'm not sure where the N_() marker is supposed to
> go. Not in the .def file, it seems.

The normal approach would be to make sure the relevant macro has the 
relevant parameter named "msgid".  (Of course you also need to make sure 
the string is translated when output, otherwise the translations in .po 
files are useless.  If it's passed directly as the format string to a 
diagnostic function this is automatic - such functions call _() on their 
format arguments - while otherwise it needs to go through _() exactly 
once.  This only applies to uses in diagnostics; dump files for example 
are not translated.)

-- 
Joseph S. Myers
joseph@codesourcery.com


More information about the Gcc-patches mailing list