On Wed, Nov 09, 2005 at 01:29:58PM -0700, Joshua A. Andler wrote:
Bryce Harrington wrote:
On Wed, Nov 09, 2005 at 08:08:37PM +0100, Ralf Stephan wrote: No problem. It actually brings up a good point about whether it's time to think about if we'd gain from upgrading our dependencies. I have a hunch that we could switch to 2.6 with minimal pain at this point. 2.8 would be more of a stretch, but if the 0.44 release is a few months out we may even have enough time for that...
Since you brought that up... Given that .44 is a few months out, if we did require 2.8, how is this our problem if people don't upgrade? If people want to use the newest version of The GIMP when it hits (2.4), they'll be required to upgrade GTK+ for that. Newest Gnome? Already required.
Changing dependencies presents us with two tangible problems:
1. Some portion of our userbase is using older versions of the dependencies, and some portion of them will complain to us when they have to upgrade. Documenting and warning them will reduce this a little, but a lot of people won't pay attention to that and will complain. It won't be our problem to get them upgraded, but it WILL be our problem to have to wade through all their complaints and questions. ;-)
2. For packages for certain platforms, sometimes the dependency won't be available. This was a bit of a problem when we went to 2.4, and so we did still have a number of people who were stuck having to use older versions of inkscape.
I imagine the above two issues would probably be doable with gtk 2.6 since it's been out sufficiently long; for 2.8 I'm less sure.
It would probably be worthwhile to look at what version of Gtk is installed in the enterprise distros that were released around a year ago. That'd give a reasonably conservative estimate of what library versions we can count on being out there.
Bryce