Re: en@quot.po [Re: en_US.po]

[Malcolm Tredinnick <malcolm commsecure com au>]

On Fri, Mar 07, 2003 at 04:35:57PM -0800, Evan Martin wrote:
> On Fri, Mar 07, 2003 at 06:30:59AM +0100, Karl Eichwalder wrote:
> > Noah Levitt <nlevitt@columbia.edu> writes:
> > > I've written an en_US.po for gtk+. The main differences from
> > > the untranslated strings are curly quotes and hyphens and
> > > stuff like that.
> > 
> > Check out en@quot.po that comes with gettext; it's machine generated.
> 
> I notice that gettextize pulls in a bunch of sed scripts and headers to
> generate en@quot.po [1], but glib-gettextize doesn't.  Is this a
> limitation of glib-gettextize or is there some other good reason to not
> do this?  (For example: it works in a different way / Unicode quotes are
> bad / etc.)

Some of the extras that you get from today's gettextize were not around
at glib-gettextize creation time. Other portions were deemed unnecessary
and written out of the process. Realise that glib-gettextize is a small
script (it was actively trimmed) that was designed to emulate portions
of the process of gettext-0.10.x (for some value of 'x' that I cannot
find right now). It's not a perfect solution, but it is a very good
solution and almost everything uses it at the moment because of some
pain that was introduced by gettext versions over the past couple of
years. Recently, gettext has stabilised and some sanity has been
restored to the point that packages can consider moving back, but, not
surprisingly, everybody is pretty skittish.

Malcolm

-- 
How many of you believe in telekinesis? Raise my hand...