Re: guile-gnome bits

[Andreas Rottmann]

Kevin Ryde <address@hidden> writes:

> Andreas Rottmann <address@hidden> writes:
>>
>> See glib/examples/main-loop.scm, I think this should answer your
>> questions.
>
> I couldn't figure out g-io-add-watch-full.  What should the GIOFunc
> be?  (I'm wanting the equivalent of gtk-input-add, an i/o fd in the
> main loop.)
>
g-io-add-watch-full is superfluous, it will be removed. Use
g-io-add-watch which is a custom wrapper around
g_io_add_watch_full(). There is even example use of it in
glib/examples/fifo.scm, which seems to be exactly what you want,
AFAICT. I polished the example a little bit, changes are in
address@hidden/glib--dev--0--patch-4.

Trying it out, however, revealed a bug in the way G-Wrap handled
default arguments when invisible arguments are used also (such as
GError, as is the case in g_io_channel_new_file). Either call
g-io-error-new-file with two arguments, or use a fixed G-Wrap (fix is
in address@hidden/g-wrap--dev--0--patch-11, a fixed Debian
package will be prepared soonish).

> Also, how do get-size args come back?  Eg. gdk-drawable-get-size.
> That sort of thing was multiple-values in the old g-wrap, has that
> changed?
>
I now added the approriate type-rules to gdk-spec.scm
(address@hidden/gtk--dev--0--patch-4) to make gint*
and guint* output arguments. And yes, output arguments are returned as
multiple-values.

HTH, Rotty
-- 
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

Make free software, not war!