Alright, this is turning into a flamewar and that's the least productive of all. Let me try to wrap up this discussion: The question: Will the gimp-1.3 developer releases depend on Gtk-1.3 HEAD CVS, or do we make certain every gimp-1.3.x release compiles with gtk-1.3.y? Arguments for depending on HEAD: - The Gtk-1.3 API is frozen, so using the latest won't break anything, it will only be better code - These releases are for developers only, normal users don't need to have anything to do with CVS. - Gtk will be tested well prior to release, avoiding the possible need of major changes after release of Gtk-2.0. Arguments against depending on HEAD, and just using the latest Gtk-1.3.y release to work with: - Gtk HEAD may not always compile, making it difficult for users to try out the development releases in the gimp-1.3 branch - If there are major advantages of CVS HEAD over the latest development release they will probably do a new release anyway, and besides, this is unlikely as Gtk-2.0 is late in its development cycle already. I might have missed one or two arguments, apologies in advance if that's the case. I think we need to ask ourselves why users would want to try the latest developer releases of Gimp. If they want to have the latest because of having the latest, I don't think they'll mind getting CVS HEAD branches and weeding out possible compile problems. But I think for gimp-1.1 there was a different reason. Gimp-1.1 had a whole lot of features that weren't in gimp-1.0. In fact, to me (as a user) Gimp-1.1 was a good graphics program, while Gimp-1.0 was hopelessly limited. So my question is, will Gimp-1.3/2.0, in the early stages of development, add much functionality? It seems to me it won't be an advantage, as for now it's basically the functionality of gimp-1.2 with a whole new implementation. But if there are no functional advantages the average user will be happy to keep using 1.2 for a while (I know I will at least). So in that case, it doesn't really matter, as long as the developers are happy. Once gimp-1.3 actually starts being a useable graphics package with more features than gimp-1.2 I think we need to worry about users being able to compile things easily, and I do believe simply depending on a fixed Gtk-version (which will then probably be at 2.0.x anyway) is a part of that. As for pango and atk, if I understand correctly they are simply part of Gtk-2.0, or at least standard companions to it. In that case why not use them? I'm sure there are gimp-users in Israel who'd like a Hebrew translation, and if that work is done already by the pango developers, why not make use of it? With Gtk-2.0, people will have it anyway. The same goes for atk. Please, try hitting the ball and not your opponent. It's not a nice thing to do, and given that your opponent is on your own team, pretty stupid as well. Lourens
