#gemrb@irc.freenode.net logs for 29 Feb 2012 (GMT)

Archive Today Yesterday Tomorrow
GemRB homepage


[00:05:53] <-- xrogaan has left IRC (Ping timeout: 245 seconds)
[00:45:31] --> brad_a has joined #gemrb
[01:15:06] <brad_a> i guess its a good thing gemrb is broken right now so i can do my homework instead of playing :-p
[01:38:05] <-- edheldil_ has left IRC (Ping timeout: 260 seconds)
[01:41:50] --> edheldil_ has joined #gemrb
[01:54:38] <-- edheldil_ has left IRC (Ping timeout: 240 seconds)
[02:18:51] <-- haad has left IRC (Quit: Leaving.)
[02:24:29] <brad_a> if we need to ifdef stuff to fix the blitter pixel format iOS is using SDL_PIXELFORMAT_ABGR8888
[02:25:21] <brad_a> id like to think we wont be ifdefing anything and instead maybe pass the pixel format around
[02:27:24] <brad_a> maybe use SDL_PixelFormatEnumToMasks to get the masks?
[04:02:53] --> joneirik has joined #gemrb
[05:13:59] <-- joneirik has left IRC (Remote host closed the connection)
[06:38:43] <-- brad_a has left IRC (Quit: brad_a)
[08:06:43] --> edheldil_ has joined #gemrb
[08:11:29] --> lynxlynxlynx has joined #gemrb
[08:11:29] --- ChanServ gives channel operator status to lynxlynxlynx
[08:15:26] <-- edheldil_ has left IRC (Ping timeout: 272 seconds)
[08:31:24] --> xrogaan has joined #gemrb
[08:31:24] <-- xrogaan has left IRC (Changing host)
[08:31:24] --> xrogaan has joined #gemrb
[09:05:41] --> edheldil_ has joined #gemrb
[09:17:01] <-- edheldil_ has left IRC (Quit: Really?)
[09:40:57] --> alx3apps has joined #gemrb
[10:46:23] <wjp> so iOS, OS X and Linux are all using different 32bpp pixel formats? Fun
[10:46:48] <fuzzie> but please do #ifdef it :-p
[10:51:27] <wjp> or more MSVCHack2 template arguments
[10:51:43] <fuzzie> Well, you seemed not too happy about the idea of combinatorial explosion.
[10:51:53] <wjp> no :-)
[10:52:04] <wjp> it's up to about 400kB stripped now
[10:54:29] --> SiENcE has joined #gemrb
[12:20:17] <lynxlynxlynx> http://forums.gibberlings3.net/index.php?showtopic=24076&hl=&fromsearch=1
[12:33:10] --> haad has joined #gemrb
[13:59:11] --> kida_laptop has joined #gemrb
[14:59:34] --> brad_a has joined #gemrb
[15:05:23] <brad_a> wjp: cant we just pass the sdl pixel format pointer around? it has members for the r/g/b/a masks and the r/g/b/a shift values
[15:06:12] <brad_a> also what was different about this before that this wasn't a problem?
[15:07:01] <fuzzie> brad_a: this is way slower if you can't hardcode the shifts/masks
[15:07:13] <brad_a> gotcha
[15:07:30] <brad_a> compiler optimization or something?
[15:08:43] <fuzzie> yes, you can generate much faster code if you can say "hey, I don't need to do any shifting or masking" :)
[15:08:50] <fuzzie> of course this is just my assumption :p
[15:10:16] <brad_a> even though there is shifting and masking with hardcoded values?
[15:10:53] <fuzzie> there isn't actually, in 32bpp
[15:11:14] <fuzzie> since it's all bytes
[15:15:06] <brad_a> so lines 668-673 in spriteRenderer.inl wouldn't get compiled the same if you were using a variable instead of hardcoded values then?
[15:15:46] <brad_a> because the values are byte sized?
[15:17:38] <brad_a> and that just reraises my question about what was being done before....
[15:18:17] <fuzzie> before it was shifting/masking using variables all the time
[15:18:19] <fuzzie> afaik
[15:19:06] <brad_a> well nobody complained about speed then :-p
[15:19:13] <fuzzie> really?
[15:19:22] <fuzzie> *I* complained about the speed there :P
[15:19:25] <brad_a> heh
[15:19:29] <brad_a> on android
[15:19:30] <brad_a> ?
[15:19:35] <fuzzie> no, on x86
[15:19:40] <brad_a> ah
[15:20:00] <fuzzie> on this 3ghz desktop machine it doesn't matter ofc
[15:20:10] <brad_a> naturally :)
[15:20:33] <brad_a> i just kinda figured if an iphone/ipad can play at a consistant 30fps most desktops could
[15:24:53] <brad_a> and now im curious why diffrent platforms use diffrent pixel formats by default :-/
[15:27:02] <fuzzie> am surprised if the ipad can keep up with 1024x768
[15:29:44] <brad_a> me too suppose considering the terribly inefficient surface to texture conversion
[15:30:58] <brad_a> if we could lay out a plan to make use of textures directly that would be swell :)
[15:31:26] <fuzzie> very surprised really
[15:32:21] <brad_a> well ios suspends most other execution when you are in an app
[15:33:39] <-- kida_laptop has left IRC (Ping timeout: 244 seconds)
[15:34:05] <fuzzie> yes, but this is actually in-game etc? :)
[15:34:21] <brad_a> yes from the people ive spoken with
[15:34:36] <fuzzie> nutty
[15:34:42] <brad_a> only slows down when there are many summons and effects going on
[15:38:07] <fuzzie> it manages <20fps in-game on my netbook
[15:38:16] <fuzzie> or, managed. haven't tried it.
[15:38:51] <fuzzie> netbook is a bit unloved for this kind of thing, since gemrb takes longer and longer to compile on it as y'all add complexity to the compiles.
[15:41:05] <-- brad_a has left IRC (Quit: brad_a)
[16:44:10] <-- SiENcE has left IRC (Quit: @all: cya)
[17:05:47] --> kida_laptop has joined #gemrb
[17:27:35] --> brad_a has joined #gemrb
[17:36:12] --> Yoshimo has joined #gemrb
[17:43:50] <wjp> brad_a: could you get me the exact pixelformat masks that you get in ios and os x?
[17:44:14] <brad_a> ios is abgr
[17:44:38] <wjp> memory order?
[17:44:50] <wjp> oh, it's BE I guess, so that doesn't matter for iOS?
[17:45:14] <brad_a> um… *shrugs*
[17:45:17] <fuzzie> this is all LE
[17:45:22] <brad_a> oh yes
[17:45:25] <brad_a> little
[17:45:31] <brad_a> ha i know what you mean now
[17:46:57] <wjp> in that case:
[17:47:01] <wjp> memory order? :-)
[17:49:13] <brad_a> as fuzzie said its all LE
[17:50:42] <wjp> no, I mean, is that abgr in memory order?
[17:50:55] <fuzzie> it never bodes well when I am in the google results for a search
[17:52:28] <brad_a> wjp: is this what you mean: r=ff g=ff00 b=ff0000 a=ff000000
[17:53:12] <wjp> yes
[17:54:46] <wjp> which is indeed the opposite of what we have now
[17:56:15] <brad_a> yes
[17:56:17] <brad_a> :)
[17:56:25] <brad_a> i was able to determine that
[17:56:57] <brad_a> and indeed "fix" it, but obviously i was unsure if we wanted to use ifdefs or what
[17:59:45] <fuzzie> why don't we mandate a surface format again?
[18:01:04] <fuzzie> oh because I said "you should find masks which work" and brad said "I just used 0" :p
[18:03:36] <fuzzie> so, that might well still be an option.
[18:06:04] <brad_a> yes i thought of that as well
[18:07:04] <brad_a> i didnt, however, imagine passing 0s would produce diffrent defaults
[18:07:46] <brad_a> I do know the masks that were preiously used instead of 0s were causing the same problem
[18:07:51] <fuzzie> wow, that was the day after you first came here :-)
[18:08:04] <brad_a> yeah that sounds about right
[18:08:19] <brad_a> that was the first step to getting a working ios port after all
[18:08:24] <fuzzie> yes, apparently we were requesting a 32bpp surface and not passing any alpha mask.
[18:08:35] <brad_a> was that it?
[18:08:45] <brad_a> I probably would have know now what was wrong
[18:08:52] <brad_a> i have learned so much since then :)
[18:08:54] <fuzzie> well, I don't know, I'm just going by what this suspicious-looking fuzzie-from-the-past said.
[18:09:16] <fuzzie> 23:56 <fuzzie> SDL 1.2 just ignores the masks, but SDL 1.3 returns an error because it checks the masks :)
[18:09:17] <brad_a> i think you are probably correct
[18:09:28] <fuzzie> 18:40 <fuzzie> i think it defaults to the wrong way around, with red/blue swpped, but not sure
[18:09:32] <fuzzie> :-)
[18:09:37] <brad_a> i remember actually going into the SDL source to figure that out
[18:10:03] <wjp> which SDL are you using?
[18:10:16] <brad_a> 1.2.15 on os x and 2.0 on ios
[18:10:27] <brad_a> but we can try what fuzzie is talking about
[18:10:35] <brad_a> and jsut pass all the expected masks
[18:10:38] <fuzzie> well, I don't know if it's stupid to do that
[18:10:46] <fuzzie> since maybe it ends up doing copies back/forth then
[18:10:52] <brad_a> hmm yes
[18:11:04] <brad_a> because it doesnt match te display?
[18:11:06] <fuzzie> yes
[18:11:12] <fuzzie> I really don't have a clue either way.
[18:12:24] <brad_a> wjp: backBuf->format->Rmask
[18:12:25] <brad_a> for os x
[18:12:29] <brad_a> woops
[18:12:36] <brad_a> didnt paste what i thought i copied :p
[18:12:43] <brad_a> r=ff00 g=ff0000 b=ff000000 a=0
[18:12:52] <fuzzie> ugh :)
[18:12:54] <wjp> oh, charming
[18:13:05] <brad_a> heh
[18:13:08] <wjp> x86 mac I assume?
[18:13:11] <brad_a> im sorry :-p
[18:13:12] <brad_a> yes
[18:13:20] <brad_a> but we used to work on ppc too ;-)
[18:13:35] <brad_a> the xcode binary is on x86/x8664
[18:13:40] <fuzzie> probably no-one tested on ppc for quite some time
[18:13:45] <brad_a> probably ot
[18:13:46] <fuzzie> i mean, on os x ppc
[18:13:57] <brad_a> tho i did compile and run it on ppc when i first started
[18:14:09] <brad_a> and it worked fine
[18:14:28] <brad_a> i think i have only seen 1 other person trying to run on ppc os x
[18:15:07] <wjp> so we have iOS=RGBA, OS X=XRGB, Linux=BGRX
[18:15:17] <brad_a> lol
[18:15:48] <brad_a> its called rbga even tho its bytes are backwards?
[18:16:45] <brad_a> and who knows if its diffrent on SDL 2.0 vs 1.2
[18:16:59] <brad_a> maybe i should build sdl 2 for mac and see
[18:17:18] <wjp> the orders I gave are the bytes in memory order
[18:17:47] <brad_a> ah
[18:17:55] <wjp> it looks like SDL internally uses the other convention. No idea which is standard
[20:56:24] --> Avenger has joined #gemrb
[20:56:24] --- ChanServ gives channel operator status to Avenger
[20:56:36] <Avenger> anyone knows more about this? --> http://www.joystiq.com/2012/02/29/report-baldurs-gate-site-teases-beamdog-developed-franchise-pr/
[20:58:17] <brad_a> lynx posted something earlier
[20:58:48] <lynxlynxlynx> there are two short threads on g3
[21:09:25] <Avenger> well, it really looks like trent oster is going to revive bg1 (or probably bg2) in hd
[21:09:59] <fuzzie> it seems spectacularly unlikely given Hasbro's position on licensing this stuff
[21:10:50] <Avenger> well, he registered baldursgate.com and he did some other revival of old bioware game (mdk)
[21:11:40] <Avenger> those are the facts, i just connected the points
[21:12:18] <fuzzie> yeah, I just wonder why everyone is leaping to the 'updated bg1/bg2' thing rather than 'shiny new baldur's gate game' :)
[21:13:32] <fuzzie> but I guess we will see
[21:13:44] <Avenger> it is more realistic to expect a dusting up than a brand new d&d game. for two reason: 1. having the old bioware engine attach widescreen onto it, and you have a hd version. That is about a month's work even for a company :)
[21:14:03] <fuzzie> yeah, but the lawyers are crazy here
[21:14:06] <Avenger> they will not start to develop a new game, it costs a lot, and you said: license
[21:14:38] <Avenger> i think dusting down an old game is probably easier
[21:14:52] <fuzzie> sure, but Hasbro are not cool with reusing licenses
[21:15:02] <Avenger> then how the gog version came out?
[21:15:10] <fuzzie> it's the same game, on the same platform
[21:16:00] <fuzzie> plus sublicensing to a third party dev is a huge no-no..
[21:16:05] <fuzzie> .. but, clearly there is something going on :)
[21:16:52] <fuzzie> so, I guess I will see where I am wrong :P
[21:17:34] <Avenger> why, do you predict a brand new game more than a revival?
[21:17:46] <fuzzie> oh, I definitely don't predict either
[21:17:51] <Avenger> i see
[21:17:55] <fuzzie> brand new game would be way too much work, as you say
[21:18:26] <fuzzie> I just can't imagine any huge improvements to bg1/bg2 being allowed, and .. then you compete with gog
[21:18:36] <Avenger> i just found it interesting that obsidian (chris avellone) also stirred recently, like they would do a pst sequel
[21:18:38] <fuzzie> but, obviously since something is allowed
[21:18:45] <fuzzie> then there must be something
[21:18:49] <fuzzie> so clearly I am wrong
[21:19:25] <Avenger> google: chris avellone kickstarter
[21:20:02] <fuzzie> sure, but chris avellone doesn't have to do anything related to d&d
[21:20:59] <Avenger> a pst sequel is related to d&d
[21:21:07] <fuzzie> well
[21:21:13] <fuzzie> where is something about a pst sequel?
[21:22:17] <Avenger> not torment, but planescape 2: http://beefjack.com/news/chris-avellone-would-like-to-make-planescape-2-but-not-torment-2/
[21:22:21] <fuzzie> just some single reply to a pst message and then a request for what people want :)
[21:22:38] <fuzzie> well sure
[21:22:50] <fuzzie> "I would like to do this" is easy to say :)
[21:23:28] <Avenger> well about fallout and icewind he didn't say that
[21:23:35] <fuzzie> if it was "I have some defense mechanism against Hasbro's lawyers killing me with legal paperwork", then ok, cool :)
[21:24:14] <Avenger> well, how many d&d games been around recently?
[21:24:21] <Avenger> a dead license is probably useless to hasbro
[21:24:42] <fuzzie> there has been a lot of fuss about it recently
[21:25:13] <fuzzie> Hasbro cancelled Atari's license in August, right?
[21:25:30] <fuzzie> It's just all xbox/facebook/etc games. :)
[21:27:46] <Avenger> though reading more, it seems wasteland is more likely. I almost don't remember that game. i know i played it, but i don't remember which one was fallout I and which one was wasteland :)
[21:28:15] <fuzzie> hehehe
[21:28:33] <fuzzie> well I am still curious about the bg thing
[21:28:43] <fuzzie> Lightkey pointed out earlier that there was clearly an iOS hint
[21:28:47] <fuzzie> it just .. makes no sense :)
[21:28:57] <fuzzie> maybe Hasbro finally got some lawyers who are interested in not being idiots, who knows.
[21:29:01] <fuzzie> very cool if so.
[21:29:40] <Avenger> well, lets hope they don't want to squish anything like us :D
[21:55:46] <lynxlynxlynx> i'm of a similar opinion
[21:56:07] <lynxlynxlynx> new games like these would be great, but i doubt it will happen
[21:56:16] <lynxlynxlynx> just some dusting off is needless competition to us
[21:58:11] <Avenger> well, that's what i fear of: the guy may not be aware of us, but he will find it out
[21:59:35] <lynxlynxlynx> nah, i think it's more likely he is aware
[22:04:20] <tomprince> Well, if they are doing a real dusting off, and the source code *doesn't* exist, we'd be a good starting off point, probably.
[22:04:59] <tomprince> at least for values of "starting off point" that don't involve lawyers, anyway. :)
[22:05:00] <Avenger> well, a week ago i asked about copyright holders
[22:05:35] <Avenger> at that time, i thought only about obsidian stirring
[22:05:54] <Avenger> but it seems all the developers feel nostalgic about the old stuff
[22:06:28] <Avenger> i guess they just realized that it is still alive after 10 years :D
[22:10:54] <Avenger> we still have the odd tint in sdlvideo?
[22:11:02] <fuzzie> oh, you have one too?
[22:11:13] <Avenger> no
[22:11:24] <Avenger> is it sdl 2.0 specific?
[22:11:27] <fuzzie> no
[22:11:34] <fuzzie> incorrect on mac too
[22:11:41] <fuzzie> fix pending :)
[22:11:53] <Avenger> it is good on msvc6
[22:12:24] <brad_a> win and linux seem to use the same defult pixel format
[22:13:30] <Avenger> bye!
[22:13:32] <-- Avenger has left IRC (Quit: bye!)
[22:15:22] <-- Yoshimo has left IRC (Quit: Yoshimo)
[23:34:04] <-- lynxlynxlynx has left IRC (Remote host closed the connection)
[23:48:57] <-- alx3apps has left IRC (Quit: Leaving.)