#gemrb@irc.freenode.net logs for 12 Feb 2013 (GMT)

Archive Today Yesterday Tomorrow
GemRB homepage

[00:16:34] --> brada has joined #gemrb
[00:16:45] <brada> added that debug output
[00:17:20] <brada> now to see if there are 2 different audio formats being buffered
[00:17:29] <brada> to the same source i mean
[00:23:32] <brada> there is also a commented out print in AmbientMgrAL::AmbientSource::enqueue that may prove useful
[00:24:08] <brada> also this explains why it seems to happen randomly :p
[00:24:20] <brada> because the ambient sounds are in fact randomly selected
[00:24:35] <brada> perhapps some of them are multi channel and others arent?
[00:27:01] <brada> same problem could be related to music, but that may have been incidentally fixed by those recent cahnges
[00:27:17] <brada> do let me know if you run into any more issues with music
[00:40:03] <-- brada has left IRC (Ping timeout: 245 seconds)
[01:03:00] --- ermo is now known as ermo^
[01:56:17] --> brada has joined #gemrb
[02:19:50] <-- edheldil has left IRC (Ping timeout: 255 seconds)
[03:39:48] <-- tomprince has left IRC (*.net *.split)
[03:47:57] --> tomprince has joined #gemrb
[06:40:28] <-- brada has left IRC (Quit: brada)
[09:02:39] --> lynxlynxlynx has joined #gemrb
[09:02:39] <-- lynxlynxlynx has left IRC (Changing host)
[09:02:39] --> lynxlynxlynx has joined #gemrb
[09:02:39] --- ChanServ gives channel operator status to lynxlynxlynx
[09:55:35] --> WingedHussar has joined #gemrb
[12:57:33] <-- traveler has left IRC (Ping timeout: 245 seconds)
[13:09:59] --> traveler has joined #gemrb
[13:47:08] <-- chiv_ has left IRC (Ping timeout: 245 seconds)
[14:47:48] --- ermo^ is now known as ermo
[15:41:26] --> brada has joined #gemrb
[15:54:42] --> psch has joined #gemrb
[15:55:46] --> mikol79 has joined #gemrb
[15:58:23] <mikol79> in bg2, the ambient sounds and music play in game, but not in bg1. Why do you think it is a bg1 issue specifically? (latest git version)
[15:59:50] <mikol79> or rather, why it seems to happen in bg1, but not bg2?
[16:01:18] <edheldil_> mikol79: any interesting errors in the console?
[16:01:45] <mikol79> I'll have a look-just a minute...
[16:04:21] <mikol79> I can't find any meaningful openal info, but will paste the log in a minute
[16:10:22] <mikol79> log is here: http://paste.ubuntu.com/1639787/ Ignore the resourcemanager errors, as everything is setup correctly, and in game ambient noise and music was playing until recently
[16:10:23] <Seniorita> Ubuntu Pastebin
[16:18:17] <mikol79> got something that might help
[16:19:34] <mikol79> I just loaded the 'missionpack' save that comes with the game, and in game ambient and music worked. Strange...., I'll paste that one as well.
[16:20:09] <mikol79> If we can compare the two pastes it might show what is happening
[16:21:34] <mikol79> here's the second paste: http://paste.ubuntu.com/1639829/
[16:21:39] <Seniorita> Ubuntu Pastebin
[16:26:39] <mikol79> amb_e02a.wav etc... loaded in second pastebin, whereas no amb* files loaded in first.
[16:26:43] <fuzzie> well, the first paste does have it trying to play music
[16:27:34] <fuzzie> so hm
[16:31:48] <wjp> mikol79: do you also not get the bg1 theme music at startup? or just a problem in-game?
[16:32:20] <wjp> (also, is audio enabled and are volumes not turned down?)
[16:35:38] <mikol79> Yes, the title music is still there and the introductory music when I start a new game, but then the actual in game music and ambient sound don't play
[16:36:19] <mikol79> Yes, audio is enabled and at the normal levels
[16:37:29] <mikol79> It definitely happened after I updated my git version, as fizzle also noticed yesterday
[16:37:42] <fuzzie> um
[17:23:04] <mikol79> however, in bg1, the combat music still plays (when triggered) , so it is the in game background music and ambient sounds that seem to be problematic
[17:33:19] --> geekt has joined #gemrb
[17:35:32] <-- mikol79 has left #gemrb
[17:36:30] <brada> compile with _DEBUG
[17:36:40] <brada> damn too late
[17:37:09] <brada> also what is with all the nulls for the openal info :/
[17:38:16] <geekt> I've finally got GemRB to start up on my Mac, after much searching for the config file…but now it crashes as soon as I try to buy something at the Inn, in the VERY beginning of Baldur's Gate…anyone in here running it on a Mac?
[17:38:30] <brada> yes
[17:40:23] <brada> BG1 i assume
[17:40:28] <geekt> yes
[17:42:19] --> fizzle has joined #gemrb
[17:43:33] <geekt> is there anything special I need to do for the config on a Mac? I changed the Cache to ~/Library/Caches/GemRB/, because it was complaining about the cache, set the GameType, GameName, GamePath, but commented out the CD's
[17:44:01] <brada> depends on where your data came from i guess
[17:44:13] <brada> gog games seem to need one of the cd paths set
[17:44:43] <geekt> I would have thought that it would complain rather than just crash...
[17:45:25] <geekt> I used recommended install, because a Full install was not available…could that be the problem?
[17:45:39] <brada> it probably does complain
[17:45:48] <brada> yes you need a full install
[17:46:07] <brada> i can buy items just fine
[17:46:16] <brada> did you build from git?
[17:46:21] <brada> or download?
[17:48:36] <geekt> I downloaded…building from git gave me problems, so I wanted to try the program out before I tried compiling again
[17:49:05] <brada> tomprince: are you aware buildbot is down?
[17:49:23] <brada> geekt: what problems?
[17:50:25] <psch> the git cmake has -Werror in it, and there's 3 or so warnings during the process, for me at least
[17:50:31] <psch> i'm guessing that might have been geekt's problem
[17:50:54] <brada> well that is easily disabled
[17:51:09] <geekt> I think it was just related to the requirements…a few weren't being found
[17:51:24] <brada> did you download SDL?
[17:51:34] <geekt> rather than go through and type the paths, I tried the download first
[17:51:35] <brada> because everything else comes standard on mac
[17:51:42] <geekt> I already have SDL
[17:51:51] <geekt> I use homebrew
[17:51:57] <brada> thats your problem
[17:52:06] <brada> download the SDL framework
[17:52:08] <brada> instead
[17:52:44] <geekt> I will…I wanted to get the program working first though
[17:52:55] <geekt> I'll reinstall it on XP, custom install, copy it over again, and see if it works, if not, I'll put an image of the CD's…and come back here...
[17:52:59] <brada> then make sure you have a full install
[17:53:07] --> gembot has joined #gemrb
[17:53:18] <tomprince> gembot: dance
[17:53:19] <gembot> <(^.^<)
[17:53:20] <gembot> <(^.^)>
[17:53:21] <gembot> (>^.^)>
[17:53:22] <gembot> (7^.^)7
[17:53:22] <brada> yay
[17:53:24] <gembot> (>^.^<)
[17:53:40] <brada> geekt: a new mac build will be up shortly
[17:54:30] <brada> probably wont fix your problem but you wont have to build yourself to get latest fixes
[17:56:03] <lynxlynxlynx> yeah, sounds like missing data
[17:57:31] <brada> geekt: the console log will tell you what is missing
[18:01:28] --> mikol79 has joined #gemrb
[18:01:47] <brada> mikol79: build with _DEBUG flag
[18:01:57] <mikol79> brada: I'll try the debug build
[18:02:01] <brada> that should log the buffer formats when queuing
[18:03:02] <mikol79> I don't know why 'null' info is shown on the openal info: I have libopenal1 installed etc.
[18:04:55] <brada> it probably means nothing
[18:05:00] <brada> fizzle didnt have it either
[18:06:09] <brada> since the problem seems to only happen in bg1 it is most likely some ambient noises have a different format than others
[18:20:56] <fizzle> traveler: how can I test shapeshifting with bears in bg1?
[18:26:46] <fizzle> it seems the problem with background music disappearing is because it doesn't loop and stops after the first playthrough
[18:29:56] <brada> i dont think its supposed to loop
[18:30:07] <brada> isnt there a play list?
[18:30:31] <fizzle> no idea
[18:30:47] <fizzle> but I often see "No Other Music to play"
[18:31:05] <brada> well there should be more than one piece of music...
[18:31:55] <brada> hmm
[18:32:05] <brada> that would be due to driver->MusicPlaying being false
[18:33:16] <brada> do you see an error before that?
[18:33:26] <fizzle> no
[18:33:54] <fizzle> it happens at the regular end of the previous piece of music
[18:34:35] <brada> the only way i see at a glance for musicplaying to be false without error is if something pauses it
[18:36:00] <mikol79> I'll attach the full log, but I get "No Other Music to play" as well
[18:37:12] <brada> im more concerned about ambients right now
[18:37:27] <brada> the music may not be a bug in the audio driver
[18:37:38] <fizzle> concerning that: is source 27 special in any way?
[18:38:39] <brada> afik no
[18:38:51] <brada> we just grab them as needed
[18:39:08] <fizzle> it seems to be the only one to get stereo buffers
[18:39:40] <brada> is it getting mono too?
[18:39:54] <fizzle> not that I've seen yet
[18:40:05] <fizzle> but I've not seen any errors, yet, either
[18:40:08] <brada> ok
[18:41:27] <-- brada has left IRC (Quit: brada)
[18:42:54] <mikol79> Here's the bg1 log from the debug build: http://paste.ubuntu.com/1640248/
[18:42:55] <Seniorita> Ubuntu Pastebin
[18:48:00] <fizzle> got one, and yes, it looks like format changes are the problem
[18:48:11] <fizzle> Attempting to buffer audio source:15 Frequency:22050 Bits:16 Channels:2
[18:48:18] <fizzle> Attempting to buffer audio source:15 Frequency:22050 Bits:16 Channels:1
[18:48:27] <fizzle> Unable to queue buffer: 0xa004 - Invalid Operation
[18:48:50] <fizzle> (nb: all those newlines in the output aren't helping...)
[19:02:10] --> brada has joined #gemrb
[19:04:55] <brada> well i plan on removing that code when the problem is fixed
[19:05:01] <brada> but nice to have confirmation
[19:05:23] <brada> looks like the easiest way to fix it is for the ambient manager to obtain 2 sources
[19:05:30] <brada> one for mono and the other sterio
[19:05:33] <brada> stereo
[19:06:20] <fizzle> woudln't it make more sense to allocate sources dynamically?
[19:06:27] <brada> maybe
[19:06:35] <brada> i would like to know why we do it this way first
[19:06:50] <fizzle> you could also get 22k vs 44k or other format differences
[19:07:01] <brada> have you seen that?
[19:07:13] <fizzle> no
[19:07:14] <brada> or are you just saying if somebody were to make their own dataset?
[19:07:28] <fizzle> i have only seen bg1
[19:07:36] <fizzle> no idea what the other games do
[19:07:49] <fizzle> and modders are certainly prime suspects here
[19:07:57] <brada> sure
[19:08:06] <brada> but ive never seen this problem in bg2
[19:08:18] <fizzle> have you seen it in bg1?
[19:08:29] <brada> i dont typically test/play in bg1
[19:08:30] <fizzle> maybe it just affects localized versions
[19:08:38] <brada> i prefer bg1 in bg2 engine
[19:09:27] <fizzle> all I'm saying is: the sound system should be robust against such changes in sample format if there's no format guarantee
[19:10:28] <brada> i think if i am to dynamically allocate them i should keep them around instead of allocating and destroying frequently
[19:12:03] <brada> fuzzie: any input?
[19:12:45] <brada> i think maybe do a quick fix now
[19:12:50] <brada> and rewrite after release
[19:13:42] <brada> or if we are still a ways away from release i can do it before
[19:16:40] <fizzle> it might also be possible to pool sources and make sure to only queue buffers to sources which are either idle or using the same format already
[19:17:45] <fizzle> but it's not entirely clear to me from the documentation whether queuing a different format when the old buffer has already been processed is valid
[19:18:00] <brada> yeah i thought about that
[19:18:08] <brada> it seems to indicate it is valid
[19:18:48] <brada> tho iirc you may have to "reset" the source by queuing a null buffer
[19:20:07] <brada> lynx: how close to release?
[19:20:32] <lynxlynxlynx> not before end of month
[19:22:57] --> Yoshimo has joined #gemrb
[19:23:09] <lynxlynxlynx> btw, the music lists are in the MUS playlists
[19:23:21] <lynxlynxlynx> no idea if they should loop by default
[19:23:26] <brada> we could probably just extent AudioStream to have fields for bits and channels and frequency etc
[19:24:17] <brada> im not 100% sure how musicplaying is getting set to false
[19:29:47] <fizzle> MUSImporter:PlayNext is calling Stop if it hits the end of its list
[19:33:01] <brada> ok
[19:33:19] <brada> so it should probably generate a new list instead?
[19:33:39] <fizzle> more likely restart the one it has
[19:34:06] <brada> well that would be easy enough, but how does original work
[19:34:19] <brada> not that i particularly care about replicating original behavior here
[19:38:12] <-- mikol79 has left #gemrb
[19:41:45] <brada> anyway thanks for taking the time to track that down and confirm the issue
[19:43:36] <brada> looking at the code of PlayNext() it does (in theory) start a new playlist
[19:45:16] --> edheldil has joined #gemrb
[19:45:16] --- ChanServ gives channel operator status to edheldil
[19:46:11] <brada> is it jsut me or does PlayMusic() not actually play music?
[19:50:26] <-- |Cable| has left IRC (Ping timeout: 245 seconds)
[19:57:40] <-- edheldil has left IRC (Ping timeout: 272 seconds)
[19:58:27] <fizzle> iesdp indicates the mus playlists should loop by default
[20:03:14] --> brada_ has joined #gemrb
[20:03:15] <-- brada has left IRC (Read error: Connection reset by peer)
[20:03:15] --- brada_ is now known as brada
[20:03:31] --> |Cable| has joined #gemrb
[20:13:43] --> brada_ has joined #gemrb
[20:13:43] <-- brada has left IRC (Read error: Connection reset by peer)
[20:13:44] --- brada_ is now known as brada
[20:17:02] --> edheldil has joined #gemrb
[20:17:02] --- ChanServ gives channel operator status to edheldil
[20:23:54] --- ermo is now known as ermo^
[20:24:17] <-- Yoshimo has left IRC (Quit: Yoshimo)
[20:32:54] <-- edheldil has left IRC (Ping timeout: 244 seconds)
[20:43:21] --> brada_ has joined #gemrb
[20:43:23] <-- brada has left IRC (Read error: Connection reset by peer)
[20:43:23] --- brada_ is now known as brada
[20:44:07] --> edheldil has joined #gemrb
[20:44:07] --- ChanServ gives channel operator status to edheldil
[20:48:45] <traveler> fizzle: try included save from tosc, there jaheira has innate wolf, bear and cave bear.
[20:48:56] <traveler> sorry for late reply, i was away
[20:49:11] <fizzle> when does she get that?
[20:49:22] <traveler> you know this ulgoth's bear bundled one
[20:49:24] <traveler> i have no idea
[20:49:28] <fizzle> because she doesn't have it in my game...
[20:49:41] <traveler> that's not possible...
[20:49:53] <traveler> i mean, she gets that with druid lvls
[20:50:16] <traveler> probably, and i'm speaking of default bundled save?
[20:50:54] <traveler> if you are sure she doesn;t have this as innate, i can tar you this save and put it on dropbox, but i'm sure ulgoth's beard bundled save should be the same
[20:51:34] <traveler> it's called 'MIssion-Pack-Save'
[20:51:36] <fizzle> thanks, I just gave Imoen the innate for testing :)
[20:52:06] <traveler> should be the same
[20:52:08] <fizzle> still, would be interesting to know whether Jaheira gets it when you run totsc with an old party, too
[20:52:31] <traveler> you mean on what level?
[20:52:45] <fizzle> if it's tied to level, yes
[20:53:19] <traveler> i'm pretty sure she gets it with levels, first wolf, then bear and and lastly cave bear. iirc from my original run years ago.
[20:54:07] <traveler> https://encrypted.google.com/search?hl=en&q=tosc%20%22mission%2DPack%2DSave%22 pretty common concept i would say
[20:54:09] <Seniorita> tosc "mission-Pack-Save" - Google Search
[21:01:45] <psch> fwiw, i've successfully built 0.7.2-git for android, with the instructions from the FR on the sf.net tracker
[21:02:58] <psch> i only have pst for testing here, character creation works, the games loads but during the intro dialogue with mort i have an issue with text lines breaking early, up to the point that the text scrolls below the screen and the replies aren't visible in the box anymore
[21:03:20] <psch> the pst has widescreen mod installed, nothing else
[21:08:30] <psch> i think i'll have a look without widescreen mod, see what happens then
[21:09:33] <traveler> pst is considerably less polished than bg1/2
[21:10:32] <traveler> i thought android package has it's own screen mode, so maybe widescreened pst data are indeed interfering
[21:10:43] <brada> no
[21:10:47] <brada> that is ios
[21:10:58] <traveler> ah, ok
[21:11:28] <psch> well, tablet android has kind of unconventional screen size, because of on-screen back, home and switch (or whatever the third is) buttons
[21:11:36] <psch> but i don't know if that should interfere
[21:12:04] <psch> ill try and get one of the baldurs gates, anyway
[21:12:27] <psch> should i put the apk online somewhere? the official package is rather out of date afair
[21:12:47] <traveler> there was some wip lately
[21:13:07] <traveler> but more packages to test probably wouldn't hurt if it's indeed working
[21:13:15] <psch> trasd opened an issue on the tracker, that's where i got the build instructions from
[21:13:21] <psch> issue 3599172
[21:13:34] <brada> yeah but that build was done the same way
[21:13:51] <psch> which build? the latest downloadable on sf.net?
[21:13:57] <brada> yes
[21:14:13] <psch> well, yes, but it's 0.7.0-something, isn't it
[21:15:34] <brada> no
[21:15:42] <psch> fair enough then
[21:15:44] <brada> im talking about the recent 0.7.2
[21:15:55] <brada> dont know where you download it guess someplace else
[21:16:25] <psch> i've downloaded the apk from http://sourceforge.net/projects/gemrb/files/Other%20Binaries/android/
[21:16:31] <Seniorita> GemRB Game Engine - Browse /Other Binaries/android at SourceForge.net
[21:17:12] <psch> i didn't find any other location for android builds, which prompted me to look around for building it
[21:17:22] <traveler> uhm
[21:17:31] <traveler> there were some 0.7.2 test builds
[21:17:42] <traveler> but maybe they didn't got onto sf
[21:18:08] <traveler> lynx: was miha doing them?
[21:18:48] <brada> yes
[21:19:25] <brada> anyway we have heard of that problem before. dont remember if it was on android tho
[21:19:39] <brada> are you using ttf fonts?
[21:20:00] <lynxlynxlynx> yeah, we need to streamline the building
[21:20:21] <brada> yeah i was hoping for this release :po
[21:20:32] <psch> i'm not actually sure im using ttf
[21:20:38] <psch> as in, i don't know
[21:20:46] <psch> i did see that i have ghostdog installed though
[21:20:56] <psch> maybe that's breaking something, i dont remember if i resized the font
[21:21:07] <psch> and it's kind of hard to tell for me just from looking at the ingame screen
[21:23:05] <brada> maybe it is mod releated then
[21:23:42] <brada> certainly nothing ive ever been able to replicate
[21:23:54] <brada> heard about it once before but dont remember anything else
[21:24:06] <psch> any of the baldur's gates prefered?
[21:24:45] <brada> BG2 is the best supported afik
[21:24:57] <brada> bg1 may be more completeable but has other bugs
[21:25:07] <brada> such as the audio stuff we were talking about earlier :p
[21:25:24] <brada> the main story should be completeable in several games
[21:25:50] <brada> nothing is completable in IWD2 save the prolouge
[21:26:03] <brada> i think chapter 1 except for the fact that you cant level up
[21:27:15] <psch> there's a few hiccups in the build process as it's written on the tracker, i think they're related to out of date makefiles, e.g. i had to outcomment Win32Console.{cpp,h} cause they wouldn't compile
[21:27:33] <psch> i'd guess whoever is working on this is probably aware of those issues
[21:27:54] <psch> anyway, i'll try bg2 then
[21:28:54] --- ermo^ is now known as ermo
[21:29:16] <brada> yeah they know and they will be irrelevant when we make our own buildscript
[21:30:27] <brada> i keep hoping that all the people wanting android builds will just submit one to us :p
[21:31:00] <psch> a buildscript?
[21:31:01] <-- brada has left IRC (Read error: Connection reset by peer)
[21:31:14] --> brada has joined #gemrb
[21:31:52] <psch> i was thinking about trying to get from what is in that one git repo towards a buildscript specifically for gemrb and nothing else
[21:32:24] <psch> but from what i can see there's still porting work in libsdl inside that repo that isn't in the official libsdl
[21:32:34] <psch> and that's not really stuff i want to try and write myself
[21:32:45] <brada> dont then
[21:32:49] <brada> use official sdl
[21:33:10] <brada> its the only way (currently) to use multitouch anyway
[21:35:31] <psch> i see
[21:35:33] <brada> use the directions and java stuff from sdl
[21:44:00] --> traveler__ has joined #gemrb
[21:44:27] --- ermo is now known as ermo^
[21:45:28] <-- traveler has left IRC (Ping timeout: 245 seconds)
[21:47:34] <traveler__> brada: as far as i can tell, bg1 is devoid of issues affecting gameplay in major way, and yes, should be really 100% completable as far as original data permits, unless I missed something or regression happened.
[21:52:29] <lynxlynxlynx> psch: help would be appreciated
[21:52:37] <lynxlynxlynx> currently we're stuck with two choices
[21:53:00] <lynxlynxlynx> the market builds are from those instructions and use that custom sdl 1.3 port
[21:53:24] <lynxlynxlynx> supposedly work on more devices and have the nice settings and demo download stuff with them
[21:53:54] <lynxlynxlynx> we've had only some test sdl2 builds in the past, but i think the problems were resolved
[21:57:24] <lynxlynxlynx> so that's a viable starting point, since the build system is much cleaner
[21:57:28] <psch> all the experience i have with the ndk comes from trying to follow those instruction, i'm not sure i can really be of much help
[21:57:46] <psch> but i'd definitely like to take a look at the sdl2 build system that exists
[21:58:01] <lynxlynxlynx> the goal is to make a build slave that the buildbot can then autorun
[21:59:24] <lynxlynxlynx> i think it uses vms, so the easier the initial setup, the better
[22:00:30] <brada> following the directions with SDL should produce a useable result
[22:00:44] <brada> the only downside would be the loss of the extra goodies
[22:00:53] <brada> but we could add those back in our own way
[22:01:19] <brada> but you also gain better touch input so overall a win imo
[22:01:33] <brada> the working on more devices is of little benifit i think
[22:01:52] <brada> since i think devices pre gles 2.0 are likely to not run very well
[22:04:40] <lynxlynxlynx> that's not something we can judge easily
[22:05:00] <fuzzie> i don't think that's particularly the problem
[22:05:03] <lynxlynxlynx> for the release i'd like to give people an option
[22:05:11] <fuzzie> it would be cool to know what it does/doesn't work on.
[22:05:39] <lynxlynxlynx> the play version is the play version and i there i think the extras matter more than touch improvements
[22:05:49] <lynxlynxlynx> getting data is a hassle even with that
[22:08:46] <fuzzie> the play version only helps if you take the demo though?
[22:09:38] <brada> what improvement are there that are better than touch input?
[22:09:50] <fuzzie> brada: well, working at all :)
[22:10:14] <lynxlynxlynx> demo downloader, keybinding support
[22:10:27] <lynxlynxlynx> some other settings that i don't know how relevant they are
[22:10:31] <fuzzie> not very.
[22:10:34] <brada> id rather be able to use touch gestures then key binding
[22:10:50] <brada> demo download is also not very useful for me :p
[22:10:53] <lynxlynxlynx> it's a waste to not use keys when they are available
[22:11:07] <fuzzie> i mean, if someone had the time+ability to do the work, it wouldn't be hard to get it to the same state
[22:11:13] <brada> sure, but you are saying that like we cant have both
[22:11:17] <fuzzie> but you're not volunteering, right? :P
[22:11:46] <lynxlynxlynx> it doesn't magically work if you switch to sdl2, no
[22:11:57] <brada> sure it did
[22:12:01] <fuzzie> but you can't really put something on the market which plain doesn't work.
[22:12:03] <brada> last time beholder was around
[22:12:11] <brada> there was one minor fix needed
[22:12:14] <fuzzie> and if i install an sdl2 gemrb and hit the icon it doesn't work.
[22:12:35] <brada> that doesnt mean anything to me
[22:12:41] <fuzzie> there's no cfg by default
[22:12:47] <brada> where did you get an sdl2 gemrb?
[22:12:49] <lynxlynxlynx> pelya doesn't have sdl2 and i doubt beholder ported that code over
[22:12:50] <fuzzie> that is lynx's point, i think
[22:13:03] <fuzzie> the automated demo download gives you a way to get a usable gemrb
[22:13:06] <brada> but beholder did make an sdl2 build!
[22:13:10] <fuzzie> and once it works you can install a game if you want
[22:13:16] <lynxlynxlynx> a plain sdl2 build
[22:13:19] <brada> he never released it for whatever reason
[22:13:20] <fuzzie> beholder's sdl2 build was pelya's sdl1.3, surely
[22:13:23] <brada> no
[22:13:29] <fuzzie> well, that's the only thing he talked about on irc
[22:13:39] <brada> it was a private chat with me
[22:13:43] <fuzzie> aha
[22:14:05] <lynxlynxlynx> then it was plain
[22:14:18] <fuzzie> well I can't theorise on stuff in private :-P
[22:14:23] <brada> heh
[22:14:29] <brada> well im trying to tell you
[22:14:33] <fuzzie> but you can't put a build on the play store that doesn't work without manually making a cfg and etc etc
[22:14:38] <fuzzie> and that is the point here, not anything technical
[22:14:49] <brada> he was able to build with sdl 2 and we were able to use touch input after a small fix
[22:14:57] <brada> unfortunately that fix broke it for ios
[22:14:58] <fuzzie> and afaik we don't have any volunteer for fixing that, no matter how easy it might be
[22:15:13] <brada> there has been some development to sdl since so who knows now
[22:15:17] <fuzzie> the hardware thing is a different issue, if we get builds working then it's easy enough to test that
[22:15:38] <fuzzie> just ask on the forum and ask people to try it with the test data or whatever
[22:15:49] <brada> i dont think we necessarily need to put it on the store
[22:16:05] <brada> and we can always have 2 builds
[22:16:16] <fuzzie> so everyone is agreed then :P
[22:16:30] <lynxlynxlynx> *always* is not a good idea
[22:16:51] <brada> fuzzie: do you have any feedback for that openal thing?
[22:17:07] <fuzzie> brada: i haven't had time to look at what you changed
[22:17:09] <brada> best would be reimplement our own wrapper interface
[22:17:24] <brada> i mean for fixing the bug
[22:17:29] <brada> not what i already changed
[22:18:00] <brada> we know for certainty now that the problem is trying to queue mono audio on stereo source and vice versa
[22:18:07] <fuzzie> i can glare at you for using bg1 in bg2 engine if it helps though
[22:18:13] <brada> heh
[22:18:30] <fuzzie> but, well, summary of code change, it worked before, you broke it? :)
[22:18:35] <brada> no
[22:18:39] <fuzzie> what did you do, make it use the same stream stuff as everything else?
[22:19:04] <fuzzie> because the stream classes were totally a hack not intended for general use, so it would make sense to extend/wrap stuff if you need
[22:19:21] <brada> talking stricktly about ambients here
[22:20:21] --> brada_ has joined #gemrb
[22:20:21] <-- brada has left IRC (Read error: Connection reset by peer)
[22:20:21] --- brada_ is now known as brada
[22:20:40] <fuzzie> it's still using the old code?
[22:20:50] <brada> what old code?
[22:21:04] <fuzzie> well, ok.
[22:21:08] <brada> afik this problem has been around for ever
[22:21:11] <fuzzie> i mean, i seriously have't had time to look what changed
[22:21:12] <fuzzie> oh
[22:21:24] <fuzzie> that's a bit weirder
[22:21:25] <brada> my recent cahnges were so we could track it down
[22:21:48] <brada> and apparently some ambient noises are stereo and others are mono.
[22:21:54] <brada> in bg1
[22:22:24] <fuzzie> that has worked fine in the past
[22:22:30] <fuzzie> maybe an awful long time ago
[22:23:15] <brada> you think i broke it?
[22:24:10] <fuzzie> if you say you didn't then i am fine assuming that
[22:24:26] <fuzzie> i feel the message of me not having looked at changes isn't getting through :)
[22:24:36] <fuzzie> just the convo above seemed to start out with that assumption, sorry
[22:24:43] <brada> well it depends on what set of cahnges :p
[22:25:02] <brada> recently no ididnt break it becaue those changes were for finding the problem
[22:25:09] <brada> error reporting / checking mostly
[22:25:29] <brada> before that i did a changeset for queuing speech
[22:26:24] <brada> but i made that into a function i didnt actually change logic afik
[22:27:25] <brada> anyway i wanted your opinion on how to fix it
[22:28:07] <brada> maybe specifically wanted to know why we allocate all the streams up front
[22:29:19] <wjp> sources?
[22:29:38] <fuzzie> no, we allocate new sources
[22:29:48] <wjp> ah, right, we only pre-count them
[22:30:12] <brada> stream
[22:30:13] <fuzzie> well, apparently we don't
[22:30:18] <brada> sorry
[22:30:23] <fuzzie> but this is confusing
[22:30:41] <brada> sorry for that too :p
[22:31:19] <brada> AmbientMgrAL has a stream it queues to
[22:31:34] <brada> and then it proceeds to queue multiple audio formats
[22:31:38] <brada> and boom
[22:31:46] <brada> that is my understanding
[22:32:03] --- ermo^ is now known as ermo
[22:32:05] <brada> perhapps fizzle can chime in with facts since he is the one who was replicating the problem for me
[22:32:39] <fuzzie> ah
[22:32:53] <fuzzie> looks to me like you broke it
[22:33:12] <brada> explain?
[22:33:14] <fuzzie> - alGenSources( 1, &Source );
[22:33:18] <fuzzie> + if(!Source || !alIsSource(Source)) {
[22:33:18] <fuzzie> + alGenSources( 1, &Source );
[22:33:25] <fuzzie> (17d7e3eea29b69916524666885e4678fc5ac71db)
[22:34:19] <fuzzie> oh, no, the ambients code is separate?
[22:34:28] <fuzzie> maybe this should just all be rewritten
[22:34:33] <brada> sure
[22:34:39] <brada> thats what i wanted your opinion on
[22:35:12] <fuzzie> the ambient stuff is certainly really insane
[22:35:59] <fizzle> lynxlynxlynx: I think I know what the problem with shape-changing is: multiple polymorph and/or anim_id effects fighting for supremacy
[22:36:03] <brada> and i didnt do that :D
[22:36:29] <fizzle> and since RefreshEffects is called on every single update I don't have any good ideas how to fix that
[22:37:30] <lynxlynxlynx> that was my suspicion too
[22:37:47] <lynxlynxlynx> does the original allow directly polymorphing twice?
[22:38:12] <fuzzie> and fizzle isn't on iOS?
[22:38:18] <lynxlynxlynx> if it does, the removal is most cleanly done in the effect itself, i guess (on first apply)
[22:38:22] <brada> fuzzie: no why?
[22:38:51] <brada> travler too has those problems. iu think we all would if we played bg1 long enough
[22:39:06] <fuzzie> yeah, i'm just trying to follow the code
[22:39:13] <fuzzie> thinking 'what idiot wrote this commit?' and then it says fuzzie
[22:39:26] <brada> heh nobody is immune!
[22:39:45] <fizzle> lynx: we also need to check for anim_id_modifier in polymorph and vice versa
[22:40:01] <fizzle> not sure what combinations the original allows or not
[22:40:12] <fuzzie> i don't think anything relevant changed in a long time
[22:40:41] <lynxlynxlynx> fizzle: i wouldn't worry about that one, polymorph probably just overrides it
[22:40:55] <fizzle> right now it doesn't
[22:41:24] <lynxlynxlynx> which spells can do that though? i can think only of the girdle of sex change
[22:41:35] <fizzle> cloak of the wolf, too
[22:41:37] <fuzzie> brada: but do you have any idea on a fix?
[22:41:43] <brada> i have a few
[22:41:47] <lynxlynxlynx> doesn't that polymorph?
[22:41:52] <fuzzie> because it seems that clearly you can't fix it by e.g. making a new stream, unfortunately
[22:41:54] <fizzle> no
[22:42:11] <brada> we could dynamically make the streams instead of precreating the 30 we do now
[22:42:20] <brada> but maybe there is a reason we do it that way
[22:42:24] <brada> such as performance
[22:42:26] <fuzzie> does that help here though?
[22:42:40] <lynxlynxlynx> it's easy to solve though, we could store an extra state or check for the presence of the other effect directly
[22:42:42] <fuzzie> the ambient code is still going to be buffering onto the end of an existing stream..
[22:42:57] <brada> yes but it could be able to get more than one
[22:43:04] <fuzzie> but it doesn't want more than one, right?
[22:43:16] <brada> currently it can even if it wanted to
[22:43:29] <fuzzie> i thought the whole point of the (crazy) ambient tick/etc code is that it can get perfect control of the stream by buffering as it wishes
[22:43:48] <brada> i did not study the code enough it seems
[22:44:14] <fuzzie> or I didn't
[22:44:31] <brada> well i didnt look much at all at the ambiental class
[22:44:34] <fizzle> lynx: go for it then :) I'll be back for the weekend
[22:44:39] <wjp> it basically just keeps a single (AL) source around per ambient source at the moment, releasing it when goes out of range
[22:44:47] <wjp> at least in theory
[22:45:20] <lynxlynxlynx> you think this affects more than just the cloak and polymorph combination?
[22:45:20] <-- brada has left IRC (Read error: Connection reset by peer)
[22:45:22] <wjp> (or ignoring it if there are too many ambient streams already)
[22:45:32] --> brada has joined #gemrb
[22:45:38] <fizzle> no idea, really
[22:46:09] <brada> well for now a simple quick fix might be to have 2 streams one for streo and mono each
[22:46:29] <brada> but the better fix would probably involve more of a rewrite
[22:47:08] <wjp> ew
[22:47:12] <brada> especially since that "fix" only solves the problem for the data at hand
[22:47:14] <brada> yes i know
[22:47:46] <brada> thats why im asking the previous authors for their take :)
[22:48:16] <fuzzie> i don't see how quite a bit of this is meant to work
[22:48:39] <wjp> anything in particular?
[22:48:44] <brada> it does seem a bit spaghetti like
[22:49:43] <fuzzie> wjp: where's the guarantee the ambient stream won't vanish from under it?
[22:50:02] <brada> i dont think the streams ever disappear
[22:50:21] <brada> they are static in the audio driver
[22:50:42] <fuzzie> they can be reused though.
[22:50:47] <brada> yes
[22:50:54] <-- fizzle has left #gemrb
[22:50:59] <fuzzie> and there's nothing stopping them from being reused while the ambient driver is, in fact, using them.
[22:51:05] <fuzzie> afaict.
[22:51:08] <wjp> they should be locked
[22:51:27] <fuzzie> ah
[22:51:49] <fuzzie> this lack of member variable naming makes this really difficult to read
[22:52:18] <fuzzie> thanks
[22:55:19] <fuzzie> brada: why the check for AL_STATIC?
[22:56:03] <fuzzie> it seems kind of bad if that ever happens
[22:56:08] <fuzzie> or if that ever can happen
[22:57:30] <brada> we shouldnt need it true. i added it to eliminate possible causes for that ambiguous error
[22:58:10] <brada> we know the problem now. definitely nothing to do with queuing to static sources
[22:58:27] <brada> also since i refactored Play() we dont have any static sources afict
[22:58:34] <brada> so yes it is useless
[22:59:18] <fuzzie> so um
[22:59:25] <fuzzie> all the changes look fine
[22:59:44] <fuzzie> AmbientSource::tick is not doing 'stream = -1;' when it calls ReleaseStream though
[22:59:49] <fuzzie> that seems like a disaster lurking
[22:59:59] <fuzzie> if i understand that code right
[23:04:45] <brada> time to leave. ill bug you about it tomorrow
[23:05:05] <fuzzie> your theory sounds as good as any :)
[23:05:12] <fuzzie> but should be easy enough to try..
[23:05:19] <fuzzie> if you didn't already confirm
[23:05:19] <brada> yes
[23:05:36] <brada> well fizzle confirmed the issu. not necissarily the root cause
[23:05:53] <brada> the issue is definately attempting to queue 2 diffrent formats to the same source
[23:06:41] <fuzzie> well, the snippet of a diff I pasted above meant that previously you'd get a new source for non-ambient plays
[23:06:52] <wjp> hm, yes, not setting it to -1 looks like a serious bug
[23:07:02] <brada> nobody has said anything about other sounds breaking
[23:07:05] <fuzzie> and you still get a new source for ambient plays
[23:07:26] <brada> well i was thinking that the ambient sounds themselves come in 2 formats in bg1
[23:07:39] <brada> i wasnt able to replicate in bg2
[23:07:41] <fuzzie> but i am not convinced about that being a possible cause, it's the wrong way around
[23:08:13] <fuzzie> brada: well I'm thinking it might've worked by coincidence due to these bugs/weirdnesses. or we might all be stupid.
[23:08:20] <fuzzie> the -1 thing seems like it would badly break ambients though
[23:08:23] <wjp> yes
[23:08:27] <wjp> fix that and see what happens :-0
[23:08:30] <wjp> s/0/)/
[23:08:39] <brada> sure good starting place :D
[23:09:05] <brada> anyway leaving for real this time
[23:09:08] <-- brada has left IRC (Quit: brada)
[23:09:49] <wjp> oh, hm, guess I'll have to search through the logs for the reasoning behind that statement about formats then
[23:11:33] <fuzzie> enjoy :P
[23:11:52] <wjp> nothing conclusive in here
[23:12:31] <wjp> at least, nothing that means ambients would have different formats rather than ambients using the wrong stream. But not all log info was pasted
[23:55:53] <-- traveler__ has left IRC (Ping timeout: 245 seconds)