gtkmm: Berlin get-together

I spent the weekend in Berlin with some of the other gtkmm hackers. It was worth meeting people in person – it’s very difficult to guess peoples’ personalities otherwise. danielk seems to be back with us, but now finds it almost impossible to understand me face to face. It’s a good thing that we communicate so well online. I also met cactus again, who I’m really beginning to like because he laughs at my jokes. And Christof Pettig, who explained how glademm is actually a good thing. He seems to use gtkmm for lots of in-house stuff at work. For a settled family man, he also has a bizarre but harmless interest in those DnD-style boardgames. Andreas Holzmann, who is official maintainer of libgnomecanvasmm, has the confidence to say very little until necessary, or maybe he just has problems with my English too. He’s also lucky enough to be using gtkmm in a paid job. I didn’t know Sebastian Rittau very well before but he seems like a nice, energetic, guy, involved in regular GNOME as well as just gtkmm. He really saved us by suggesting the Old Emerald Isle pub as a meeting place.

cactus and I were also able to announce that the separate orbitcpp C++ CORBA mapping for ORBit2 now works. thanks to markmc‘s efforts. And I just got libbonobouimm working with it. The C++-in-ORBit-itself idea didn’t happen in the end but it’s working fine as a separate module. As soon as a release of ORBit2 HEAD is made then we can complete the gnomemm APIs using the libbonobouimm classes. And I’ll expect some help with that.

We did decide to binary freeze and release gtkmm 2.0 by Halloween at the latest, but I might try to do it before that. As an escape clause, we might break the API again for gtkmm 2.4 because it might be necessary for GTK+ API additions. But we’ll avoid it if we can. We just have to accept that we will miss some things.

In anticipation of the gtkmm freeze, I released a binary API-frozen libsigc++ 1.2.0. kenelson gave us his reworked 1.9 code and I hope that someone will get that new development phase underway and eventually take over maintainership.

Real life is also taking time away from development work, but in a good way. And with my dial-up, it’s also very difficult to be quite as tediously exact and omnipresent as normal so there’s a risk that it might look like I don’t care.