Re: Problems building 2.0 API; gtkmmproc error after update



Murray Cumming wrote:
On Wed, 2008-02-27 at 15:46 +0100, Toralf Lund wrote:
Murray Cumming wrote:
On Tue, 2008-02-26 at 23:43 +0100, Toralf Lund wrote:
[ ... ]


Yeah, it may seem like the signal handler return types changed at some point, although I though they shouldn't based on the fact that GTK+ is supposed to have a completely stable API...

There have been very tiny mistakes like this in the past, and small
differences of opinion about what is public API and what is a break.
gtkmm only notices them because we use almost the whole API.
Yeah, I suppose the functions and pointers we're talking about here may at least be considered more "internal" or "private" than a large proportion of the gtk_... calls.
I'm not sure what GTK+ version I tried to build with, as I'm on a different machine now, but I would assume it was more or less the latest and greatest. What version I had last time it worked is probably impossible to tell now...
Can't you just use a newer version of gtkmm?
That's not really the question here. We're talking about a special build system on MacOS X, which is supposed to support a "package" for the Gtkmm 2.0 API, only now this support is broken. There's still a question of whether anybody will actually use it, of course, but gtk and gtkmm 1.2 are also supported, if you know what I'm saying...

In any case, my question is really what I generally need to change (i.e. which source files etc.) in order to fix problems like this. Is there some sort of guideline etc?

In this case you should probably just patch the generated .h/cc files.
Yes. Perhaps that's the best idea here. I'm always trying to do things the "correct" way with respect to auto-generated sources etc, you know, but maybe it's just silly to be particular about this for something that was just supposed to be a quick-fix that I wasn't necessarily going to use...

- Toralf



[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]