Re: orbit-config.in patch

[Michael Meeks <michael ximian com>]

Hi Ross,

        Ok, I don't support this whatsoever - I think we should just bite
the bullet and go with pkgconfig; but we're trying to stay faithfull to
Elliot's last will and testament: avoid pkgconfig :-)

        So ...

On 13 Oct 2001, Ross Golder wrote:
> I've got a patch for orbit-config.in to make it use pkg-config.
>
> Apparently, orbit-config.in is ./configured to orbit-config, and then
> copied to orbit2-config by the Makefile. What's all that about? :)

        That sucks; I just did the CVS surgery, can you remove the old
orbit-config as you commit your stuff ?

> So, as my program uses /usr/share/aclocal/ORBit2.m4, which rather than
> use pkg-config directly (shouldn't it?)

        I would use pkg-config directly for everything; [ preferably via.
the nice autoconf macros it provides, see eg. libbonobo's use of it ].

> I've patched it to use pkg-config --cflags ORBit-2.0 (which includes
> the glib-2.0 --cflags).

        Sounds fine - if it works, and the old was broken; we can only
hide pkgconfig nowadays, glib-config-2.0 etc. is dead.

> But, I'm thinking that the correct approach would be to have ORBit2.m4
> use pkg-config directly, bypassing the need for an orbit2-config. Am I
> off the mark here?

        I'd prefer to leave the concealment in a shell script than in the
m4 scripts which get turned into a myriad configure scripts :-)

> And another query - I guess I'm doing the right thing posting all
> these patches here, but what happens next (or... which of the
> following potential outcomes are good, bad, likely or unlikely)?

        Sorry - I don't read this list often enough - mailing just me to
poke me is prolly best.

> - I get bored waiting for a response, commit it,

        Not a good strategy as yet.

        Anyhow :-) thanks for all your great work, it's looking nice.

                Michael.

-- 
 mmeeks@gnu.org  <><, Pseudo Engineer, itinerant idiot