On Wed, Aug 8, 2012 at 12:40 PM, Andrea Del Signore <seje...@gmail.com> wrote: > Hi all, > > recently with these two commits [1], [2] support for the posix and > dova profiles was removed from the vala compiler. >
Hi again, I've made some progress on this issue and I published my code here: http://gitorious.org/ideas/vala-codegen-posix Now that I reached a somewhat working implementation, I didn't do any more tests apart from the included examples, I've a mixed feeling on the patches required to implement this codegen plugin On one hand I like the simplification that the removal made to the compiler, but also I fear how now vala is bound with gobject in significant part of the code that instead should be general enough. In a previous mail Juerg asked me why implement a plugin system instead of just forking the compiler. Now I think that every fork should maintain a set of patches not only to implement the codegen, but also to remove all the gobject-ism than inevitably valac will include (see [1],[2]), so may be a dual profile, GObject vs all the others, can be healthy for the whole compiler. What's your opinion? Regards, Andrea P.S. That code was tested only on windows, since I don't have access to any linux box ATM. [1] http://gitorious.org/ideas/vala-codegen-posix/blobs/master/valac-patches/0002-Don-t-use-GLib-types-when-the-current-profile-is-not.patch [2] http://gitorious.org/ideas/vala-codegen-posix/blobs/master/valac-patches/0003-Add-tests-for-unavailable-standard-types.patch _______________________________________________ vala-list mailing list vala-list@gnome.org https://mail.gnome.org/mailman/listinfo/vala-list
