Config
Log for #openttd.dev on 4th June 2014:
Times are UTC Toggle Colours
05:29:51  *** Supercheese has joined #openttd.dev
05:29:51  *** ChanServ sets mode: +v Supercheese
09:04:20  *** Supercheese has quit IRC
16:24:01  *** frosch123 has joined #openttd.dev
16:24:01  *** ChanServ sets mode: +v frosch123
16:46:11  *** Alberth has joined #openttd.dev
16:46:12  *** ChanServ sets mode: +v Alberth
18:19:23  *** andythenorth has joined #openttd.dev
18:19:23  *** ChanServ sets mode: +v andythenorth
18:19:24  *** andythenorth has left #openttd.dev
20:05:53  <frosch123> fonsinchen: 8bpp/32bpp changed in trunk btw
20:06:06  <frosch123> ottd always uses 32bpp now, unless enabled via config file
20:08:04  *** Alberth has left #openttd.dev
20:19:16  <fonsinchen> what is that dropdown doing then?
20:20:21  <frosch123> what dropdown?
20:21:28  <fonsinchen> Oh, it changed in trunk, but not in the 1.4 branch. That might explain why FS#6012 doesn't happen in with r26620
20:31:30  <fonsinchen> I would be willing to shed a tear for the loss of openttd support for my funny mac zoo if willglynn actually goes and fixes the OSX video driver.
20:33:55  <planetmaker> http://bugs.openttd.org/task/6012 <-- you mean that?
20:33:59  <planetmaker> and the 'fix' ?
20:36:02  <frosch123> well, it also explains "fails to enter full screen"
20:36:14  <frosch123> so, there are two options
20:36:29  <fonsinchen> yes. Apparently the guy would like to have a stab at "modernizing the OSX port". That is definitely a good idea, independent of that specific bug.
20:36:30  <frosch123> teach the osx driver what the win drivers can do for 5 years or more
20:36:41  <frosch123> or ignore it, since trunk does not use 8bpp by default
20:36:56  <planetmaker> ah, he wrote something new today which I didn't yet see
20:37:01  <frosch123> we could just hardcode 32bpp into osx :)
20:37:19  <planetmaker> yes. It's nearly so anyway
20:37:54  <frosch123> anyhow, ottd should not crash on an empty list in any case, that should be fixed in the settings gui
20:38:07  <frosch123> that could also be backported, and the rest ignored
20:38:21  <planetmaker> We should give im a heads-up to modernize and throw-out old code for ancient OSXes
20:38:27  <planetmaker> maybe 10.6+ would do?
20:38:37  <planetmaker> (or I can't test anymore :P )
20:38:46  <frosch123> sure, if you update the compile farm for that :p
20:39:02  <planetmaker> uhm... I forgot that bear trap :(
20:39:12  <fonsinchen> argh
20:39:13  <frosch123> i have no idea about that stuff, can't discuss with you
20:40:40  <planetmaker> that's a definitive stumbling stone
20:42:39  <planetmaker> so what's the option for this dilemma?
21:02:46  * Rubidium wonders why OS X only increases the minor number while breaking APIs all around, whereas windows increases the major number while keeping even ancient APIs (reasonably) working
21:11:14  * fonsinchen should do an extra "oldmac" video driver that drops all the unnecesary newfangled stuff, and altivec-accelerated blitters and sprite sorters to go with it
21:11:36  <planetmaker> migth be a good idea to actually structure the stuff, fonsinchen
21:11:44  <planetmaker> then we could also maintain the old build
21:11:54  <planetmaker> you might comment that way in that FS task
21:12:04  <planetmaker> (I basically gave a somewhat summary of this talk here)
21:12:14  <planetmaker> or rather expansion of it
21:12:42  <planetmaker> I like that idea of separate drivers for old and new OSX
21:13:07  <planetmaker> like one 10.3 ... 10.5 for the current CF. And 10.6 ... 10.9 for a possible new one
21:13:25  <planetmaker> or whatever
21:13:58  <frosch123> we already have quartz, quickdraw, cocoa :p
21:14:26  <planetmaker> well...
21:14:43  <frosch123> "In Mac OS X v10.4, QuickDraw has been officially deprecated."
21:15:15  <planetmaker> that might go completely :)
21:16:21  <planetmaker> or whatever :)
21:16:27  <Rubidium> I do not think maintaining a plethora of drivers for OS X is going to do anyone a favour
21:17:06  <fonsinchen> The #ifdef mess seems worse than maintaining multiple drivers
21:17:14  <Rubidium> if there is a bug in the release binary, it is likely NOT the driver that the developer that tries to fix it uses (or even supports). As such it would become very tricky to fix such bugs
21:18:07  <Rubidium> so... if you want to do it, do it thoroughly; drop everything and only build a x86_64 driver
21:18:13  <Rubidium> s/driver/binary/
21:19:38  <fonsinchen> That will work for half a year and then apple will come out with some new API churn.
21:20:29  <Rubidium> and it will take about 3 years before *someone* starts working on modernizing the driver again, so a good time to skip another 3+ different APIs
21:22:07  <fonsinchen> Already now I can't really reproduce much of what is happening on a 10.9 mac, even with formally the same driver.
21:22:32  <planetmaker> you have a 10.9 install, fonsinchen ?
21:22:41  <fonsinchen> No, that's the point
21:24:14  <fonsinchen> gah, cmake ... enough
21:24:19  <fonsinchen> good night
21:24:25  <planetmaker> g'night :)
21:32:45  <Rubidium> also, what part of the video driver changes with each new release of OS X and what remains? In other words, would it make sense to have multiple drivers extending the same base?
21:39:43  *** frosch123 has quit IRC

Powered by YARRSTE version: svn-trunk