Re: Orbit under Win32 - Plea for Help!

[Tor Lillqvist <tml iki fi>]

Michael Meeks writes:
 > Tor did some work porting ORBit2 to Win32, but hit some fun problems
 > with the WaitForMultipleFoo's implementation (I think). The root problem
 > is that glib's mainloop on Windows is profoundly different and somewhat
 > limiting (I think).

Yes. Basically (as I recall), the ORBit2 code itself wasn't that hard
to port (a cvs diff -u0 outputs 1164 lines). The problems were in the
assumptions the linc2 library makes on how GLib can be used.

For a writeup that hopefully clarifies the issues see
http://bugzilla.gnome.org/show_bug.cgi?id=120299

 > Tor - what was the final status of your work ? - compiling, but of
 > limited usefulness ? did we get most of your changes in to make it at
 > least closer to where it needs to go ?

Yes, it compiles, but without the giowin32 rewrite, is of limited
usefulness.

I didn't commit any of my changes to CVS, as I thought that I would
have time and/or inspiration to do the necessary rewrite of GLib's
giowin32.c relatively quickly (say, in a month or so), and thought I
should wait for that to make sure everything works first. Alas, that
didn't happen. I probably should present what changes I have for
comment. I'll send the diffs in a separate mail a bit later.

--tml