< mclasen> time to start ? * mclasen somewhat distracted by waiting for a school bus < dom> no time like the present < ebassi> good for me < mitch> yea < mclasen> ok < mclasen> first topic is "what do we do about the public outcry over the adjustment changes" ? < bratsche> Hi < mclasen> the way I see it, there are 3 options here < mclasen> 1) revert it to the previous behaviour < mclasen> 2) stick to it and maybe explain it better, broken apps be damned < mclasen> 3) try to make it more palatable by changing spin buttons (what else? ) to ignore page_size < mclasen> anything else ? < mitch> is 3) feasible at all? < kalikiana> 4) restore the old behaviour and instead emit a critical warning < mitch> it's a model/view ting after all < ebassi> assuming that 3 did not break applications, 2+3 would be the perfect solution < mclasen> mitch: I haven't explored it really < Sonderblade> 3 is feasible, atleast for spin buttons < mitch> kalikiana: in spinbutton? < leio-dl> 1) with the fix reintroduced in a later bug fix release after better public notification < leio-dl> or something like 3) with a similar approach < mitch> i don't see how that can be done within GtkSpinButton? < kalikiana> mitch: it's not only spin buttons < mitch> kalikiana: i know < mitch> gimp was affected on all levels < kalikiana> and my volume is stuck at 80% < kalikiana> so behaviour should not change, still a warning would make sense to me < kalikiana> or am I overlooking something? < mitch> kalikiana: what should not change? < kalikiana> mitch: the behaviour that now breaks all of my desktop and previously worked < mitch> i still think that all these apps are broken and were broken before < mclasen> mitch: we can try again in 2.90... < mitch> mclasen: indeed < Sonderblade> what was the bug# and the commit that caused the new behaviour? < mitch> if we revert i would add a big fat warning to spinbutton and scale < mclasen> so, if we revert to accepting the full range, where would we warn ? < mclasen> when setting a problematic adjustment on the affected widgets ? < mitch> in all adjustment users apart from scrollbar < mitch> i would say there and in their "changed" callbacks < mclasen> that could be a lot of warning, though < mitch> not "value-changed" < mitch> "changed" < mclasen> ah, ok < kalikiana> mitch: That's what I just said as 4) :P < mclasen> do we have a volunteer to come up with a patch for that ? < mitch> until when? tomorrow? < mclasen> ideally < mitch> i can do that tomorrow, and revert my own original commit < mclasen> I may have a look myself tonight < mitch> mclasen: if you do, don't forget to also revert all the removed manual (upper - page_size) i removed < mitch> scratch one "removed" ;) < mclasen> we only revert on the gtk-2-14, and leave the strict code in trunk ? < ebassi> trunk should stay < ebassi> we have a year worth of warning < bratsche> That would go into 2.16 right? Is that intended? < mclasen> bratsche: maybe not < mclasen> but we can sort that out when 2.16 and 2.90 split ways < bratsche> Maybe save it for when 2.90 is branched separately from 2.16 < bratsche> Yeah. < mitch> i agree < ebassi> so < ebassi> ACTION: revert and add a warning in gtk-2-14 < ebassi> ACTION: add the strict code when we branch in 2.90 < mclasen> ok, thanks < mitch> it's three things to revert IIRC, starting 2008-08-05 < mitch> what? add strict code when we branch? < mitch> i thought we keep the strict code in trunk? < mclasen> mitch: for 2.90, yeah < mclasen> maybe not for 2.16 < bratsche> But trunk will become 2.16, right? < mclasen> but we can sort that out later < mitch> yea indeed, no hurry < mclasen> I just need the stable branch to be fixed right now < mclasen> was there another topic for today ? < mitch> i added one, it's all my fault today < mitch> i basically have patches ready to get rid of all H and V subclasses and have "orientation" properties in their parents < mclasen> we surely can't drop the subclasses ? < mitch> that gets rid of so much code duplication < mitch> mclasen: why not? < kalikiana> but we can deprecate the subclasses :) < mclasen> because apps are using them ? < mitch> sure, deprecate them, not drop :) < kris> I wonder if apps will ever move away from using them though ... < mitch> that would give us nice opportinities to fix default values, since e.g. GtkBox is abstract < mitch> so having GtkBox instances is essentially a new widget < mitch> where the packing defaults could be fixed < mitch> kris: sure, many people fix their code to not use deprecated, don't you? ;) < kalikiana> that would be a pretty neat side effect < bratsche> Is it possible to make like #define gtk_hbox_new helpers that alias to gtk_box_new and add the correct property, plus the remaining arguments to the #define? < ebassi> +1 for < mitch> bratsche: no, the function would have to stay < kris> mitch: okay okay < ebassi> #define gtk_vbox_new(t,s) < kris> mitch: no further input from me here as I don't maintain apps < ebassi> gtk_box_new (orientation, t, s) < mitch> we can't get rid of the symbol < mitch> ebassi: exactly < mitch> gtk_srparator_new() < mitch> gtk_ruler_new() < mitch> etc. * mclasen is all for making things flippable < bratsche> I know the symbol needs to stay during 2.16, but I meant during 2.90.. but anyway, that's for later. :) < mitch> ok, expect a ton of patches to be files in the coming days < kalikiana> mitch: Make a note to expressly document the properties, please, if you change them < mitch> filed < mitch> kalikiana: what change? < kalikiana> mitch: You said the defaults can be fixed < kalikiana> Or did I misunderstand? < Sonderblade> if vbox, hbox and friends become deprecated won't they still someday have to be removed? otherwise what is the point of deprecation? < mitch> kalikiana: that is unimplemented and needs to be discussed, i don't even have code for changed defaults < kalikiana> Sonderblade: Of course, but not within 2.16 < mitch> Sonderblade: exactly, but the actual point is a) reducing code duplication and b) making subclassing easier < mitch> right now you have to derive from GtkHRuler *and* GtkVRuler if you need to subclass < kalikiana> mitch: Sure, good enough if you add it to the bug orwhatever, just that it's considered right away :) < mitch> or look at GtkScaleButton, it has two subclasses for HScale and VScale < kalikiana> There are even 3rd party hvbox style widgets out there < mitch> i know, they can all die once 2.16 is out then < mitch> i will file separate bugs for each bunch of parent,v,h i hacked < mitch> ebassi: ACTION please ;) < Sonderblade> it would be very cool if they are runtime flippable :) < bratsche> heh < mclasen> one interesting thing is that we get runtime flippability < mitch> Sonderblade: they all are in my tree < ebassi> ACTION: mitch files bugs with the various patches < ebassi> :-) < mclasen> but if you have nested boxes, you probably want to flip them all at the same time < mitch> one small problem i have is e.g. GtkRange which has class members to control drawing details < mitch> that needs to be sorted somehow < Sonderblade> mitch: yay! :) < mitch> mclasen: yea i have code in testgtk for that too ;) < mitch> works nicely actually < bratsche> Oh, that's cool that it's runtime-flippable. < ebassi> neat < ebassi> this is definitely going in the minutes :-) < mclasen> are we ready to go to 'misc' ? < mitch> do we want set/get_orientation() API in addition to the properties? < mclasen> that would be the standard for widget properties, no ? < mitch> yea useless question < mitch> let's move along to misc < mclasen> the one thing I have is a follow up to the advisory board call on GTK3 < mclasen> thus mostly for kris, timj, mitch, I guess < mclasen> has anybody done any work on the action items we had coming out of that call ? < mclasen> in particular the www.gtk.org section on gtk3 < mitch> mclasen: everybody apart from me has been at the maemo summit, we will have to do lots of stuff this week < mclasen> ah, thats fine < mclasen> but we need to get going on those things soon < mitch> omg i promised work done by my coworkers ;) < mclasen> I'll keep the www.gtk.org update on my todo list < kris> we've been thinking about some more concrete features and such < ebassi> I could not make it for the conf call :-( < kris> but I dunno how that is going to proceed < kris> oh and there is timj < kris> timj: 22:31 < mclasen> has anybody done any work on the action items we had coming out of that call ? < ebassi> by the way: anybody has the actual list of action items < mclasen> wow, so late already in europe < ebassi> ? < mclasen> I think timj took notes < ebassi> (and possibly the log/minutes of the meeting of two weeks ago) < mitch> mclasen: we work our asses off at night here! < mclasen> at least he made it sound like he did... < timj> ebassi: did you get the slides? < ebassi> mitch: I had three pints down at the pub before the meeting, so I guess this qualifies as "working my ass off" as well ;-) < ebassi> timj: nope < timj> there was little point in minutes during a slide presentation < mitch> ebassi: exactly :) < ebassi> heh < mclasen> ebassi: I sent some impressions immediately after the call, I can forward that to you < mclasen> its a far cry from minutes, though < ebassi> mclasen: thanks, that would be cool < mclasen> ok, thats all I had on my list for today < mclasen> if nobody else has 'misc' items, I think we're done < bratsche> Out of curiosity, is anyone else here planning to be at Boston? < mclasen> I'm doing a stable release tomorrow, for the gnome guys < bratsche> I know ebassi and federico are going. Didn't know if anyone else is planning. < mclasen> I'll probably be there < bratsche> Cool < mclasen> Owen is going to be there, too < mitch> mclasen: i'll do the patches or review yours, whoever is faster (i'll not do anything tonight though) < kris> mclasen: we'll get back with roadmap stuff at some point, we did some concretizing and all < kris> bratsche: not me < kris> matlab fried my brain this evening < mclasen> mitch: the ass is already off... < bratsche> kris: heh < mitch> haha yes! < mclasen> good night then, everybody < bratsche> Bye. < ebassi> 'night everyone < bratsche> kris: Where is your uni anyway? < mitch> nite < kris> bratsche: Leiden < mitch> die Leiden des jungen Kris < timj> kris: could you fwd the advisory board meeting slides to ebassi? < bratsche> :) < kris> oh no I get lessons in German again < kris> timj: sure < bratsche> Okay, back to work. Bye everyone! < kris> later bratsche < ebassi> thanks kris < timj> mclasen: about AB meeting action items, it's mostly us creating a gtk-3 web presence (blog) and posting a roadmap. < mclasen> yeah < timj> i'll do my best to send out an initial draft of that in a few days (or have kris do it, since i'm on vacation next week) < mclasen> cool, thanks