I think it would be a good idea to have a hardware flag for tablet users. For example, to compile qt3 with tablet support I added use tablet && myconf="${myconf} -tablet" to qt-3.0.4.20020606-r1.ebuild, line 61 Does this make sense?
That depends. If "tablet" is a USE flag that several packages can take advantage of, and would reduce bloat, then it makes sense.
It would make sense as well to disable this flag, if you don't have a tablet. For example, emerging xfree there's no need to patch xf86Wacom.c (resolving bug #1632) in this case, which is an extra download and done by default at the moment.
I don't know of any ebuild other than qt that has optional tablet support, but then I don't have tablet support, so I wouldn't know. As far as bloat goes, I never saw or noticed any real bloat from the inclusion of tablet support in qt, but again I haven't checked and wouldn't know The X11 example, as I understand it, isn't valid since it doesn't add support for tablets, jsut fix it (it's there without the patch, but is broken). Also in X11 the support is a driver module and isn't loaded if you don't have a tablet, so there's no bloat except a little diskspace waste. So - I'm not entirely convinced. For this to become a use flag I'd need examples of 1) other packages with optional tablet support and/or 2) noticeable bloat in qt because of tablet support. I don't know of any examples for (1). I'll check (2) when I have the time (i.e. if you do it first that's ok :-).
If I'm not mistaken this is another example: In gtk+-1.2.10 there is a default of --with-xinput=xfree , which adds support for the xfree tablet drivers. From what I read I guess this is not necessary without a tablet. Rasmol is not in portage yet, I think this would be another example for an application with optional tablet support. Anyway I'll keep my eyes open. I must add, I don't know much about tablets in general, just got a Wacom Intuos2 a couple of days ago, now I'm fooling around.
GTK+ and QT are enough for this i think. I'll measure bloat for both (built with and without and compare sizes) and request confirmation on gentoo-core (which will be granted unless the bloat is trivial and someone comes up with a reason not to).
Thanks. Another ebuild that by default uses --with-xinput=xfree is gtkmm. I'll try to measure the difference as well, but I'm not sure how to do so.
I forgot to close this. When I posted on -core with the use flag proposition, Daniel Robbins replied with: "I don't think we should have a "tablet" USE flag. Let's just make sure that our ebuilds work with tablets, enabling any options as necessary." I also compiled qt without tablet support and thre was no difference in size. SO there'll be no new use flag. We're struggling to reduce the number of use flags anyway.
