#gemrb@irc.freenode.net logs for 1 May 2013 (GMT)

Archive Today Yesterday Tomorrow
GemRB homepage

[01:39:21] --> brada has joined #gemrb
[02:27:03] <-- brada has left IRC (Quit: brada)
[02:48:03] --> brada has joined #gemrb
[03:06:33] <-- brada has left IRC (Quit: brada)
[03:07:50] --> brada has joined #gemrb
[03:21:50] <-- brada has left IRC (Quit: brada)
[03:33:09] <-- Maighstir has left IRC (Quit: .)
[04:15:02] --> brada has joined #gemrb
[04:46:21] <-- brada has left IRC (Quit: brada)
[04:50:00] --> brada has joined #gemrb
[04:56:10] <-- brada has left IRC (Quit: brada)
[06:20:50] --> edheldil_ has joined #gemrb
[06:28:59] <-- edheldil_ has left IRC (Ping timeout: 246 seconds)
[06:30:36] --> vampi-the-frog has joined #gemrb
[06:31:13] --> edheldil_ has joined #gemrb
[06:38:46] <-- vampi-the-frog has left IRC (Quit: Leaving)
[06:39:27] --> vampi-the-frog has joined #gemrb
[07:18:09] --> edheldil__ has joined #gemrb
[07:48:34] <-- edheldil_ has left IRC (Read error: Operation timed out)
[07:57:32] <-- edheldil__ has left IRC (Ping timeout: 268 seconds)
[07:58:45] --> Yoshimo has joined #gemrb
[08:05:22] --> fizzle has joined #gemrb
[08:17:30] --> edheldil_ has joined #gemrb
[09:31:57] --> cj_schnell has joined #gemrb
[09:37:32] <cj_schnell> brada: I'm not sure if your monologue was directed at me and what I reported. If so I just want to clear things up. The most reliable way for me to ensure that a button action is triggered is to tap twice in rapid succession. A quick tap only works half of the time. Framerate issue etc, who knows?
[09:40:29] <fuzzie> also we totally highlight buttons on RC
[09:40:33] <cj_schnell> If I do a long press I can see the button getting "pushed" but when I let go of the button the button action seldom triggers. I do not expect any right-clicking mechanics, I just want to find the best way to press the buttons that triggers the event.
[09:43:05] <cj_schnell> And I am definitely not dragging my finger around like a mouse. What would be the point of that? :-) I just want to push buttons :-)
[10:00:38] <-- Yoshimo has left IRC (Quit: Yoshimo)
[10:14:21] <-- cj_schnell has left IRC (Quit: KVIrc 4.0.4 Insomnia http://www.kvirc.net/)
[10:26:48] <-- edheldil_ has left IRC (Ping timeout: 252 seconds)
[10:44:11] --> rocket_hamster has joined #gemrb
[10:59:07] <-- vampi-the-frog has left IRC (Quit: Leaving)
[11:12:26] --> edheldil__ has joined #gemrb
[11:57:36] --> edheldil_ has joined #gemrb
[12:12:21] --> Yoshimo has joined #gemrb
[12:22:09] <-- rocket_hamster has left IRC (Remote host closed the connection)
[12:41:23] <-- Yoshimo has left IRC (Ping timeout: 260 seconds)
[13:30:01] <-- fizzle has left #gemrb
[14:09:56] <-- edheldil_ has left IRC (Ping timeout: 255 seconds)
[15:01:48] --> lynxlynxlynx has joined #gemrb
[15:01:48] --- ChanServ gives channel operator status to lynxlynxlynx
[15:10:24] --> brada has joined #gemrb
[15:15:33] <-- |Cable| has left IRC (Ping timeout: 276 seconds)
[15:22:18] <brada> fuzzie: in what game do we highlight on RC?
[15:22:27] <brada> i tried bg2...
[15:22:43] <fuzzie> i tried bg2
[15:22:53] <fuzzie> right-click on the discussed button, it highlights
[15:22:56] <fuzzie> does it not for you?
[15:23:02] <fuzzie> the code seems pretty clear honestly
[15:23:17] <fuzzie> since it doesn't check right-click until after the highlight toggle..
[15:23:29] <brada> where in code?
[15:23:34] <fuzzie> in Button mouseup
[15:23:49] <brada> and no it didnt for me, but i was testing touch input
[15:24:16] <brada> ok thats why
[15:24:39] <fuzzie> also do you understand cj_schnell's general issue now?
[15:25:05] <brada> not at all
[15:25:09] <brada> do you?
[15:25:26] <fuzzie> well I mean what the issue is, not why it happens :P
[15:25:38] <brada> still not sure
[15:25:45] <fuzzie> it's just: touches don't work.
[15:26:32] <brada> i know thats what he says
[15:26:45] <brada> but he seems to still think the button needs to do its push animation
[15:26:49] <brada> which it doesnt
[15:26:50] <fuzzie> no
[15:26:53] <fuzzie> it needs to highlight, which it does
[15:26:58] <brada> no it doesnt
[15:27:02] <brada> not on BG2 anyway
[15:27:15] <fuzzie> well, "not sending mouseup event" is broken
[15:27:28] <fuzzie> so I don't know the consequences quite there
[15:27:37] <fuzzie> if that's what you were implying
[15:28:32] --> |Cable| has joined #gemrb
[15:28:55] <brada> im implying that since the mouse down event is delayed till the touch context is known then on finger up both are sent together so its too quick for any of the button highlights to occur
[15:29:16] <fuzzie> yes
[15:29:18] <fuzzie> you don't get it
[15:29:21] <brada> no i dont
[15:29:23] <fuzzie> it only highlights once you mouse-up.
[15:29:38] <fuzzie> it's a selection highlight, not a feedback highlight.
[15:29:40] <brada> it would help if i knew what button
[15:29:43] <fuzzie> oh
[15:29:46] <fuzzie> you should've said ;p
[15:29:52] <fuzzie> the male/female buttons at the start of char gen
[15:29:57] <brada> on BG1?
[15:29:59] <fuzzie> bg2
[15:30:14] <brada> ok ones that dont disappear after being pressed
[15:30:24] <fuzzie> yeah, it's the selection ones
[15:30:40] <fuzzie> well, maybe cj_schnell was playing with bg1, I don't know, I reproduced with mouse with bg2
[15:30:57] <brada> the highlight on RC?
[15:31:01] <fuzzie> yeah
[15:31:19] <brada> yeah we probably shouldnt do that :p
[15:31:21] <fuzzie> it seems like a bug on our part, but I don't really know, so I didn't fix it..
[15:31:29] <fuzzie> that's the only thing I looked at though
[15:32:25] <fuzzie> although I notice your SDL_FINGERDOWN has a 'if (!finger0) numFingers++;'
[15:33:31] <fuzzie> so you'd think it'd work without any frame problems?
[15:34:01] <brada> yeah and he reported FPS similar to ipad1 which doesnt suffer that ailment...
[15:34:20] <brada> so im still very confused
[15:34:28] <brada> psch cant replicate?
[15:34:29] <fuzzie> oh
[15:35:19] <fuzzie> int mouseButton = (firstFingerDown.fingerId >= 0) ? GEM_MB_ACTION : GEM_MB_MENU;
[15:35:47] <fuzzie> is that good?
[15:36:06] <brada> yes
[15:36:16] <fuzzie> the SDL code is very careful to not assume anything about fingerId
[15:36:18] <brada> we set it to -1 when we clear it
[15:36:29] <brada> yeah it used to be > 0
[15:36:34] <brada> now its >= 0 :p
[15:37:12] <fuzzie> ok
[15:37:20] <brada> it would help if cj were ever here when i am
[15:37:35] <fuzzie> yeah
[15:37:40] <fuzzie> it's pretty weird though, the code looks ok
[15:38:01] <brada> definately from iOS perspective anyway
[15:38:10] <brada> and psch didnt complain
[15:38:16] <brada> tho maybe he just doesnt like to :p
[15:40:31] <fuzzie> the SDL touch code is a bit unexpectedly nutty about finger indexes
[15:41:10] <fuzzie> when a finger is released, it will swap that finger with the last one in the array
[15:42:53] <fuzzie> i thought maybe the 'rare case where the touch has been removed before processing' bit, but the event.tfinger.x/y is the same as the one in the finger
[15:50:00] <brada> i really dont know what to think
[15:50:20] <brada> its clear from what he describes he is getting a right click
[15:50:21] <psch> i could replicate his highlight on long touch problem, yeah
[15:50:26] <brada> well yeah
[15:50:31] <fuzzie> well, the right click isn't his problem
[15:50:32] <brada> thats not a problem with touch input tho
[15:50:35] <fuzzie> his problem is that a quick tap doesn't work
[15:50:42] <brada> or now always
[15:50:45] <fuzzie> obviously the long-touch thing is predictable and we know that
[15:50:45] <brada> not
[15:50:52] <psch> that happens to me on lowish fps mostly
[15:50:58] <psch> with the ~30 i get now i hardly happens
[15:51:04] <fuzzie> right, so also reproducible..
[15:51:04] <psch> but gemrb also trained me to touch really briefly :P
[15:51:10] <brada> which i could see, but 20+ shouldnt be a problem
[15:51:20] <brada> we can up the delay...
[15:51:22] <fuzzie> why would it happen at low framerate though?
[15:51:23] <brada> if that would help
[15:51:56] <brada> well something so low that either the underlying system is destroying the events or the 500ms delay triggers
[15:52:28] <brada> you could do a build with a 1 sec delay and see if thats better
[15:53:00] <psch> i can try that, yeah
[15:55:28] <brada> like fuzzie said we should just make that configureable
[15:55:39] <brada> tho still need to figure out what a good default is
[15:58:46] <psch> well, TOUCH_RC_NUM_TICKS 1000 seems to give me more ghost touches, not less
[15:58:54] <fuzzie> interesting
[15:59:13] <psch> i don't think that makes sense, and im also pretty sure i can't quantifiably touch for a certain amount of time
[15:59:35] --> cj_schnell has joined #gemrb
[15:59:56] <psch> but for a correct left click to register i have to pretty speedy on the release
[16:00:07] <psch> more so than before
[16:00:46] <brada> o_O
[16:01:10] <psch> ive just build 250, curious what's gonna happen haha
[16:01:37] <cj_schnell> brada: I was playing bg1 when I had the male/female bug, but maybe you got it sorted by now?
[16:01:54] <fuzzie> well there are two problems there
[16:02:05] <fuzzie> one of which is that a long-press gives you a right-click and that gives wierd results
[16:02:06] <brada> male / female bug is unrelated to touch
[16:02:13] <fuzzie> and the other of which is that apparently you have problems with short clicks
[16:02:17] <fuzzie> so we're discussing the latter now
[16:03:32] <cj_schnell> Indeed it is
[16:03:40] <psch> 250 is hardly workable, but that's pretty much to be expected i guess; touches have to be really short, but 250 ticks probably aren't much time
[16:03:48] <cj_schnell> Ok, don't let me interfere.
[16:04:22] <brada> well i think i do have an idea for improvement
[16:04:47] <brada> click drags in buttons are problematic
[16:04:57] <brada> and i suspect that happens much of the time on touch screen
[16:05:43] <psch> i've build with TOUCH_RC_NUM_TICKS 2500 now
[16:05:52] <psch> and ghost touches still happen
[16:06:05] <fuzzie> ghost touches?
[16:06:07] <psch> yes
[16:06:15] <psch> i touch and the interface reacts but not really
[16:06:21] <psch> like, i get click sound but button isnt depressed
[16:06:29] <psch> or male is highlighted and done not enabled
[16:06:42] <fuzzie> hmph
[16:06:48] <fuzzie> i wonder if those really are right clicks
[16:06:53] <psch> i don't think so
[16:06:59] <psch> i think delta pixels is more likely
[16:07:10] <psch> or rather, some kind of gesture detection
[16:07:15] <fuzzie> right
[16:07:15] <brada> if they werent right clicks how else would you get that behavior?
[16:07:19] <brada> the drag?
[16:07:22] <brada> like i said?
[16:07:24] <psch> yeah
[16:07:31] <fuzzie> that makes sense to me
[16:07:33] <psch> as in, when i double touch on the same spot, i get a left click on a long second touch
[16:07:47] <psch> double touch with about a second delay that is
[16:09:40] --> rocket_hamster has joined #gemrb
[16:09:46] <brada> so
[16:10:03] <-- cj_schnell has left IRC (Quit: KVIrc 4.0.4 Insomnia http://www.kvirc.net/)
[16:10:07] <brada> maybe in finger motion we sought to check the gui control type
[16:10:20] <brada> and not always set ingnorenextfingerup
[16:10:21] <fuzzie> maybe add a 'break;' right after the 'case SDL_FINGERMOTION:' to make sure it's actually the problem if psch can?
[16:10:38] <brada> well i know its a problem
[16:10:45] <brada> dont know if its *the* problem
[16:11:39] <brada> bah
[16:11:40] <brada> well
[16:11:42] <brada> so much for that
[16:11:51] <brada> we set ignore to false in the single finger case
[16:12:30] <brada> so why isnt mouse up not getting ran after a drag?
[16:12:38] <brada> does that happen on PC too?
[16:14:27] <brada> hmmm
[16:14:34] <psch> i never noticed anything like that, no
[16:14:42] <brada> maybe the way the mouse gets "moved" in finger up is problematic
[16:16:30] <brada> goes to show i never retested this on actual touch hardware after the rewrite lol
[16:18:51] <brada> can somebody test on PC?
[16:18:54] <brada> gemrb on pc
[16:19:08] <brada> for dragging buttons canceling the action?
[16:19:11] <fuzzie> it's fine
[16:19:13] <fuzzie> sorry :P
[16:19:18] <brada> :/
[16:19:29] <brada> i see mouse up getting called after the drag
[16:19:31] <brada> on the button
[16:19:34] <fuzzie> moving *off* the button makes it fail
[16:19:36] <brada> but nothing happens
[16:19:39] <brada> yeah
[16:19:42] <fuzzie> but that's it
[16:19:43] <brada> im not talking about off
[16:19:45] <brada> tho
[16:19:51] <brada> maybe the coordinates are
[16:20:07] <brada> easy to check i suppose
[16:20:33] <fuzzie> mouseup does check the coords
[16:20:50] <brada> yeah
[16:20:59] <brada> i meant the touch input passing wrong ones
[16:21:03] <brada> but it looks fine
[16:22:04] <brada> fuzzie: is it the event handlers that do the action?
[16:22:07] <brada> im guessing
[16:22:24] <fuzzie> i assume so
[16:22:51] <brada> so that test doesnt pass for touch input
[16:22:53] <fuzzie> but that's why the highlightable gender buttons are nice
[16:22:53] <brada> for some reason
[16:23:52] <fuzzie> since they happen early, before even the coords check
[16:25:06] <brada> it is that mousebutton assignment you referenced earlier
[16:25:42] <brada> by that i mean its because i recycle mousebutton
[16:25:45] <brada> which i shouldnt
[16:26:00] <brada> er
[16:26:04] <fuzzie> ..?
[16:26:05] <brada> im confusing myself
[16:26:25] <brada> but im definately seeing that set to a RC when it shouldnt be
[16:26:54] <brada> line 489 fyi
[16:26:59] <brada> 498
[16:28:00] <fuzzie> weird
[16:30:51] <brada> psch: ok try this patch
[16:33:16] <psch> which patch?
[16:33:22] <psch> HEAD?
[16:35:28] <brada> http://paste.debian.net/1504/
[16:35:30] <brada> no
[16:35:31] <brada> sorry
[16:35:37] <brada> i got distracted
[16:35:50] <Pepelka> debian Pastezone
[16:36:07] <brada> that makes it so i can drag a bit inside a button and still get a left click up
[16:36:11] <brada> so yay
[16:36:22] <brada> fixes the thing i was complaining about anyway :p
[16:36:32] <brada> tho dont know about other side effects of that
[16:36:36] <brada> i couldnt think of any
[16:36:47] <brada> but we could test the focus control type if we have to
[16:40:59] <psch> input feels great after the patch
[16:41:16] <psch> as in, slowish touches work perfectly
[16:42:03] <brada> yay
[16:42:12] <brada> pushing now
[16:42:34] <brada> psch: will you post a new build?
[16:42:41] <brada> since that is a pretty serious bug
[16:42:52] <psch> i guess i can do that
[16:42:58] <Pepelka> [commit] bradallred: Touch Input: do an actual movement on fingerup event instead of simply updating cursor position https://github.com/gemrb/gemrb/commit/86c552f45e95d54e0f5af7ae0281b714ad2ee886
[16:43:00] <Pepelka> [commit] bradallred: Touch Input: drag events should always be left click events https://github.com/gemrb/gemrb/commit/7370efa0f0e989cbba96c020ee2b9833ace6c828
[16:43:10] <psch> though i obviously can't put it up on sourceforge
[16:43:33] <brada> ah yes, lynx will be happy to help im sure :D
[16:43:50] <brada> i need some actual touch hardware :p
[16:44:12] <brada> kinda impressed touch input works as well as it does considering :p
[16:48:34] <brada> fuzzie: is there anytime you do a right click drag?
[16:48:42] <brada> i can think of formation rotation
[16:48:58] <brada> so maybe i broke that for touch input
[16:49:11] <brada> but i think it should be fine
[16:52:48] --> Maighstir has joined #gemrb
[17:12:35] <lynxlynxlynx> sure, just link me
[17:13:41] <lynxlynxlynx> i'm leaving for almost a month of iceland on saturday, so you'll need to ask other ops to upload in the meanwhile (if there will be any significant changes)
[17:14:04] <brada> sounds fun :)
[17:14:09] <lynxlynxlynx> mhm
[17:14:13] <brada> dont fall into a volcano
[17:14:33] <lynxlynxlynx> first three days in uk to get a cheaper flight, then who knows, it's an adhoc trip :)
[17:15:08] <lynxlynxlynx> swimming in a volcano is on the agenda though, but it's a mostly dormant one
[17:15:14] <psch> so you haven't decided yet if you want to fall into a volcano
[17:15:16] <psch> oh
[17:15:17] <psch> nevermind that
[17:15:23] <lynxlynxlynx> hehe
[17:15:54] <brada> :D
[17:16:09] <lynxlynxlynx> if nothing interesting will be going on touch/mobile-wise, please try to make a build slave for the android build
[17:16:58] <psch> who's behind those again?
[17:18:39] <psch> oh btw, i also have "copy GemRB.cfg from the sdcard-root, if it exists there" in the build currently
[17:19:00] <psch> i put it in because i wasn't sure what the problems with cj were
[17:19:15] <psch> im not sure if that should stay in; it's not that useful
[17:19:56] <psch> on the other hand, "copy a working config onto the sdcard" vs "edit /sdcard/Android/data/net.sourceforge.gemrb/files/GemRB.cfg" seems a bit more user-friendly
[17:20:00] <-- rocket_hamster has left IRC (Remote host closed the connection)
[17:20:14] <lynxlynxlynx> tomprince is the guy to talk to about the details
[17:20:15] <psch> on the gripping hand, i hope i can find the time to get the gui done around the weekend
[17:20:38] <psch> http://filebin.ca/fceJ5ty5DYd/GemRB-0.8.0-7370efa.apk md5 b01e11b71f5717c598a2473162e9b3f1
[17:21:12] <lynxlynxlynx> i agree that the copying is not that useful (most people won't have such a config), but it definitely shouldn't be on by default, since then you can destroy valid configs
[17:22:37] <brada> if we can get it to build in a mac environment i can run it
[17:22:51] <brada> but it requires CS file system
[17:24:17] <lynxlynxlynx> ok, the apk is on sourceforge
[17:24:45] <lynxlynxlynx> one of you to please post it to the forum too, i only know it fixes some clicking problem
[17:24:59] <fuzzie> "fixes some clicking problem" sounds like good description to me
[17:25:08] <psch> lynxlynxlynx: i'm only copying if there's no config in the app-specific path
[17:25:34] <psch> and, of course, there is a config on the sdcard
[17:25:39] <psch> otherwise im extracting the template
[17:26:35] <psch> brada: didn't someone build succesfully on a CS filesystem on mac a few weeks back?
[17:26:43] <brada> no
[17:26:51] <lynxlynxlynx> ok
[17:26:59] <psch> oh right there was some header thing i didnt understand
[17:27:02] <brada> i mean yes we can build on mac
[17:27:07] <lynxlynxlynx> i guess it could be useful to testers
[17:27:14] <brada> but currently we need a virtual file system
[17:27:35] <brada> that is CS
[17:27:46] <psch> also, i remembered something else wrt SDL2; their template project sets apilevel 10 as target, but the activity doesn't actually compile for apileve 10, but need level 12
[17:27:55] <psch> i think they might want to know about that eventually?
[17:28:13] <brada> that sounds like a mistake to me
[17:28:24] <brada> 10 should be all that is required unless im missing something
[17:28:56] <psch> it was some class in the activity that isn't recognized for level 10
[17:28:58] <psch> let me check
[17:29:33] <psch> maybe it's fixed already, anyway
[17:30:35] <psch> well, i built against android-10 and it works, so i guess it's fine
[17:31:25] <brada> ok im off for a bit bbl
[17:31:27] <-- brada has left IRC (Quit: brada)
[18:57:12] --- ChanServ gives channel operator status to wjp
[19:17:35] --> vampi-the-frog has joined #gemrb
[19:21:13] --> rocket_hamster has joined #gemrb
[19:21:18] --> brada has joined #gemrb
[19:37:07] --> Yoshimo has joined #gemrb
[20:10:32] <-- vampi-the-frog has left IRC (Quit: Leaving)
[20:13:43] <-- Yoshimo has left IRC (Quit: Yoshimo)
[20:51:04] <-- Maighstir has left IRC (Quit: .)
[20:51:29] --> Maighstir has joined #gemrb
[21:21:19] <-- rocket_hamster has left IRC (Remote host closed the connection)
[22:09:48] <-- lynxlynxlynx has left IRC (Remote host closed the connection)
[22:15:35] --> cragganmore has joined #gemrb
[22:17:16] <cragganmore> Hi there. I have a question about a problem I'm encountering with the most recent Android build (GemRB-0.8.0-7370efa.apk).
[23:07:58] <brada> go on
[23:11:24] <-- Maighstir has left IRC (Quit: .)
[23:14:53] <cragganmore> Hi brada. My setup is as follows: Nexus 7, GemRB-0.8.0-7370efa.apk, BG2 gog.com version patched to BGT with the most recent fix pack and tweak pack. This BG2 installation runs fine on the 0.8 iOS build. When I run it on Android, the opening movies play, then the title menu screen appears with music. However, the screen is - how best to explain - yellow, dithered like crazy, and barely readable.
[23:15:16] <brada> set fullscreen=1
[23:15:19] <brada> in your cfg
[23:15:29] <brada> dont know why that is but that will fix it
[23:16:09] <cragganmore> OK I've tried setting fullscreen = 1, but with the same results. What should my height and width be set to? And should I be at 16bpp or 32?
[23:16:18] <brada> 32 but it doesnt matter
[23:16:28] <brada> that setting is a noop on android/ios
[23:16:35] <brada> for the SDL2 build anyway
[23:17:12] <cragganmore> Gotcha. I'll try it again with fullscreen = 1, height = 480, width = 640, at 32bpp. One sec.
[23:18:48] <brada> cragganmore: is this what you see? http://postimg.org/image/s3dsniqx9/
[23:18:48] <Pepelka> View image: 2013 04 30 14 26 15
[23:19:34] <cragganmore> Exactly what I see!
[23:19:49] <brada> Fullscreen=1 should fix that
[23:19:55] <brada> as was reported anyway
[23:20:03] <brada> no devs have reproduced yet
[23:20:49] <-- brada has left IRC (Quit: brada)