Re: how do you build this guile-gobject thing?

[Andreas Rottmann]

Matto Marjanovic <address@hidden> writes:

>  >That's a bit confusing: Are you using 0.5.3 or CVS? Both shouldn't
>  >yield this message.
>
> There don't seem to be any tarballs anywhere --- I looked, but even the
>  "Download" page of the guile-gtk site states:
>
>      "For the time being, release tarballs are not available.
>       Please use the development snapshots from CVS."
>
Ah, forgot that. I'd try CVS HEAD.

>  So, I checked out RELEASE_0_5_3 from CVS.  (The "Files" section of
>  the project page on Savannah is non-existent.)
>
>  >>  The README doesn't mention g-wrap at all.  The mailing lists show
>  >>  evidence of some development activity in an Arch archive somewhere,
>  >>  but I don't know how to access it or whether that is relevant.
>  >>
>  >See 
> http://mail.gnu.org/archive/html/guile-gtk-general/2004-01/msg00036.html.
>
> Awesome (and it looks like I am all set on the g-wrap side - v1.3.4).
> (Wow... what a coincidence --- I wonder if Greg Troxel's email was even
>  in the archive yet when I went looking through it yesterday.)
>
Indeed a funny coincidence. Let's hope interest in guile-gobject
increases *even* *more* ;-)

>  >> Any clues, suggestions?
>  >>
>  >Yes, compile with -O0. This is a known problem and already fixed my
>  >modified version of g-wrap; my changes will go upstream RSN.
>
> "Compile guile-gobject with -O0, due to bug in g-wrap" - right?
>  I'll give it a shot.
>
It's rather a feature of gcc (consuming load of memory when compiling
big source files with optimization). I've hacked g-wrap to spit out
~22K lines instead of ~140K lines (hint: do a 'wc -l
gnome/gtk/guile-gnome-gw-gtk.c' in the toplevel build directory) for
GTK+2, but this has yet to go upstream.

Regards, Andy
-- 
Andreas Rottmann         | address@hidden      | address@hidden | address@hidden
http://yi.org/rotty      | GnuPG Key: http://yi.org/rotty/gpg.asc
Fingerprint              | DFB4 4EB4 78A4 5EEE 6219  F228 F92F CFC5 01FD 5B62

Packages should build-depend on what they should build-depend.