#gemrb@irc.freenode.net logs for 4 Mar 2012 (GMT)

Archive Today Yesterday Tomorrow
GemRB homepage

[00:10:08] <gembot> build #565 of mingw32 is complete: Success [3build successful] Build details are at http://buildbot.gemrb.org/builders/mingw32/builds/565
[01:05:58] <-- Yoshimo has left IRC (Ping timeout: 248 seconds)
[02:26:02] --> joneirik has joined #gemrb
[02:38:22] <-- joneirik has left IRC (Ping timeout: 252 seconds)
[02:38:56] --> joneirik has joined #gemrb
[02:52:21] --> mrugiero has joined #gemrb
[02:53:13] <-- mrugiero has left #gemrb
[03:36:18] --> joneirikb has joined #gemrb
[03:39:33] <-- joneirik has left IRC (Ping timeout: 248 seconds)
[04:08:19] --> brad_a has joined #gemrb
[07:35:40] <-- joneirikb has left IRC (Remote host closed the connection)
[07:40:21] --> Yoshimo has joined #gemrb
[08:16:37] <CIA-28> GemRB: 03avenger_teambg * r3b62d36e12f3 10gemrb/gemrb/plugins/CREImporter/CREImporter.cpp: cleaned up sloppy code
[08:21:04] --> lynxlynxlynx has joined #gemrb
[08:21:04] <-- lynxlynxlynx has left IRC (Changing host)
[08:21:04] --> lynxlynxlynx has joined #gemrb
[08:21:05] --- ChanServ gives channel operator status to lynxlynxlynx
[08:45:00] --> kida_laptop has joined #gemrb
[09:15:02] <-- brad_a has left IRC (Quit: brad_a)
[09:44:19] --> alx3apps has joined #gemrb
[10:09:51] --> wrotek has joined #gemrb
[11:53:05] --> SiENcE has joined #gemrb
[12:00:09] <lynxlynxlynx> [SDLVideo/ERROR]: BlitGameSprite not supported for this sprite
[12:00:19] <lynxlynxlynx> i get a lot of these just going into the inventory
[12:00:48] <lynxlynxlynx> the real jackpot is guirec though
[12:01:01] <lynxlynxlynx> scrolling the character description gives one per line
[12:10:24] <fuzzie> button pictures?
[12:12:21] <fuzzie> no, I guess not
[12:13:09] <fuzzie> but nothing shows up wrong?
[12:13:32] <fuzzie> i assume the button code is for the blood overlay on portraits
[12:18:16] <fuzzie> ok, new question: which game is this?
[12:20:38] <fuzzie> oh, bg1? hm
[12:33:52] <lynxlynxlynx> bg2
[12:34:23] <lynxlynxlynx> i thought it was maybe the effect icons, but it is there even when they're already hidden
[12:34:44] <-- Yoshimo has left IRC (Ping timeout: 245 seconds)
[12:34:48] <fuzzie> i can't make that error trigger at all in bg2
[12:34:56] <fuzzie> but is it modded?
[12:35:14] <lynxlynxlynx> yes
[12:35:27] <lynxlynxlynx> don't see anything wrong though
[12:35:46] <fuzzie> and the grey overlay on death works fine?
[12:36:32] <lynxlynxlynx> only one char :)
[12:36:40] <lynxlynxlynx> i can confirm there's no message in vanilla
[12:36:54] <fuzzie> custom portrait?
[12:37:23] <fuzzie> it looks like it will fail if you have a 256-colour portrait, at a glance
[12:38:22] <lynxlynxlynx> no, one of the bg1 ones
[12:38:36] <lynxlynxlynx> the death overlay works in vanilla, but it's not really gray
[12:39:53] <fuzzie> i can't reproduce at all
[12:40:35] <lynxlynxlynx> well, it's just a warning if i can't tell the difference
[12:40:51] <fuzzie> but 'one of the bg1' ones = not shipped with bg2?
[12:41:47] <lynxlynxlynx> shipped
[12:42:18] <fuzzie> i don't know which is which i guess :) but they all seem fine for me
[12:42:20] <lynxlynxlynx> safana it seems
[12:42:35] <lynxlynxlynx> they differ in style a bit
[12:43:09] <lynxlynxlynx> i could catch it with gdb if you're interested in pursuing it
[12:44:56] <fuzzie> well I thought I could just catch it in gdb myself.. :)
[12:45:10] <fuzzie> but if you could make sure it's coming from the button code (or if not, where) then that would be helpful I imagine
[12:47:10] <lynxlynxlynx> it is coming from button code
[12:47:14] <lynxlynxlynx> bpp is 1 at that point
[12:47:46] <fuzzie> my safana bmps (from inside bifs) are not 8bpp
[12:47:53] <fuzzie> but ok, confirmed, sigh :)
[12:49:03] <fuzzie> it looks like it wouldn't be too painful to just make it use the palette in the 1bpp case though
[12:50:02] <fuzzie> unless I am missing some reason why we can't just tweak the BlitSpriteBAM_dispatch logic a bit
[12:50:28] <lynxlynxlynx> don't look at me :)
[12:52:00] <fuzzie> well it is probably best asked of wjp rather then me blindly poking
[12:52:37] <fuzzie> but as far as I can tell, the only thing in BlitSpriteBAM_dispatch that requires a BAM is the RLE check, which can just be false for a non-BAM image
[13:11:26] <-- SiENcE has left IRC (Quit: byebye Folks)
[15:19:16] <-- wrotek has left IRC (Ping timeout: 276 seconds)
[15:24:13] --> Yoshimo has joined #gemrb
[15:24:45] <-- Yoshimo has left IRC (Client Quit)
[15:38:56] --> wrotek has joined #gemrb
[16:04:46] --> Herrman has joined #gemrb
[16:44:48] --> Yoshimo has joined #gemrb
[16:54:45] <-- Yoshimo has left IRC (Quit: Yoshimo)
[18:19:26] --> brad_a has joined #gemrb
[18:19:30] --> culot has joined #gemrb
[18:29:31] <CIA-28> GemRB: 03avenger_teambg * r53fa6df876d7 10ielister/ielister.cpp: ielister update
[18:30:57] <CIA-28> GemRB: 03avenger_teambg * rba38925e9ded 10dltcep/ (5 files): dltcep update
[18:34:38] * tomprince notices that there are 5 days left for gsoc applications
[18:40:49] <lynxlynxlynx> ups
[18:45:03] --> SiENcE has joined #gemrb
[20:21:48] <wjp> some bitmap format that used to work no longer works?
[20:22:06] <fuzzie> wjp: 8bpp
[20:22:21] <wjp> I could've sworn I didn't change that logic
[20:22:31] <fuzzie> you removed support from it in BlitGameSprite :)
[20:23:16] <fuzzie> or did it just silently fail before?
[20:23:18] <fuzzie> either way, not so good
[20:23:18] --> barra_home has joined #gemrb
[20:23:35] <wjp> in the !BAM, !32bpp case, it used to print an error and call BlitSprite, it seems
[20:24:04] <fuzzie> but is there anything to stop it from just using BlitSpriteBAM_dispatch?
[20:24:21] <wjp> if you have raw 8bpp data plus a palette? Just the RLE check I suppose
[20:24:47] <fuzzie> i mean, to be clear: it only happens becuase you added a BlitGameSprite call for the tinting :)
[20:24:55] <wjp> aha
[20:26:54] <wjp> so I guess that means either another dispatch, or a tweak to BAM_dispatch
[20:27:21] <fuzzie> The API seemed sort of odd to me.
[20:27:59] <wjp> well, the function pre-rewrite handled BAM and 32bpp non-BAM
[20:28:31] <wjp> maybe the RGB_dispatch should become an SDL_dispatch, and check the pixel format
[20:29:14] <fuzzie> well
[20:29:35] <fuzzie> I mostly meant e.g. the way it gets passed srcdata which is always spr->pixels.
[20:29:54] <wjp> ah
[20:30:40] <wjp> too many arguments to have full overview over them :-)
[20:30:43] <fuzzie> and it gets passed spr->Width and spr->Height and those are promptly ignored by the actual blit functions in favour of spr->Width and spr->Height themselves in *some* places.
[20:31:57] <wjp> right, so that clearly needs some cleaning up
[20:32:06] <fuzzie> so yes, I kind of stepped back and didn't want to fiddle with the dispatch function... :)
[20:32:33] <brad_a> maybe fix the pixel format stuff while youre at it :) pretty please?
[20:32:54] <fuzzie> well that is annoying :P
[20:33:08] <fuzzie> esp. since we can't reproduce easily
[20:33:24] <fuzzie> *my* OS X boxes give me a nice sane pixel format :p
[20:33:52] <brad_a> are you on 16 bpp?
[20:34:56] <brad_a> if its inconsistant we definately need ot go back to actually dynamically reading the shift values
[20:35:45] <wjp> I was thinking having a variable one and a platform-specific default. Maybe chosen at run-time
[20:36:11] <brad_a> anyway 16 bpp works fine for me
[20:36:43] <wjp> could you give me the masks you get in 16 bpp on both platforms, for reference?
[20:36:56] <fuzzie> brad_a: no, I am on ppc :)
[20:37:01] <brad_a> oh i see
[20:38:14] <brad_a> wjp: ios = r=ff g=ff00 b=ff0000 a=ff000000
[20:38:30] <wjp> that's very 32 bpp
[20:38:32] <fuzzie> that isn't 16bpp :p
[20:38:39] <brad_a> oh yeah i know
[20:39:01] <brad_a> i didnt realize you asked for 16. esp since 16 works on os x anyway
[20:39:04] <fuzzie> a platform-specific default and a variable one makes sense
[20:39:21] <fuzzie> "WARNING: your hardware is weird, now you will suffer these performance issues!"
[20:39:42] <fuzzie> I looked at some ipad benchmarks and I still can't believe people get 30fps in-game on it. :P
[20:40:17] <brad_a> i dont know what the actual FPS are
[20:41:24] <fuzzie> well you can see it's pretty good in the iPad 2 videos on youtube
[20:41:52] <brad_a> so what about it makes it hard to believe?
[20:42:06] <fuzzie> because the iPad 2 is waaay faster
[20:42:06] <brad_a> i thought it had really good hardware specs for what it is
[20:42:18] <brad_a> I thought thats what we were talking about :-p
[20:42:53] <fuzzie> oh, well, if you say performance is good on ipad then i don't expect a "except if you don't have a fancy new one" disclaimer :P
[20:43:08] <brad_a> new one? the 3 is about to come out!
[20:44:02] <fuzzie> I'm sorry, I am one of those weird old people who considers hardware that was only released 12 months ago ago to be new.. :-p
[20:44:11] <brad_a> ah :)
[20:44:16] <-- alx3apps has left #gemrb
[20:44:54] <fuzzie> the ipad 2 has a Cortex-A9 with all these fancy early-90s processor architecture innovations
[20:45:56] <fuzzie> so it's more believable that it can just throw cpu at it
[20:46:45] <fuzzie> (if that's what is happening)
[20:46:56] <fuzzie> on the other hand, agh, my battery life :P
[20:47:04] <brad_a> must be cuz the SDL 2 driver is a joke
[20:50:32] <brad_a> um bpp doesnt seem to have an effect under ios/sdl2
[20:51:49] <wjp> ok, well that's easy then :-)
[20:53:16] <brad_a> i mean it will create the back buff with SDL_PIXELFORMAT_RGB565 but it still ends up looking the same with the blue and all
[20:53:36] <brad_a> i mean it does apparently convert to 32bpp when you render
[20:54:08] <wjp> ah. Ok, then, what are the masks you get? :-)
[20:54:40] <wjp> although that strongly suggests 565 with R and B swapped compared to the current code
[20:55:50] <CIA-28> GemRB: 03avenger_teambg * re3e42bff34b4 10gemrb/gemrb/ (3 files in 2 dirs): one step ahead in iwd2 action bar
[20:57:43] <brad_a> wjp: rmask=f800 gmask=7e0 bmask=1f amask=0
[21:01:55] <CIA-28> GemRB: 03avenger_teambg * rf2e1b99313dd 10gemrb/gemrb/core/Store.cpp: the fence flag is in Flags, not in Type
[21:03:20] <fuzzie> hm, isn't that standard 565?
[21:03:37] <brad_a> looks like to me
[21:04:07] <brad_a> if it werent i think sdl would report SDL_PIXELFORMAT_BGR565
[21:04:59] <wjp> that should be what we're outputting
[21:10:29] <wjp> the half-alpha blender for 16bpp looks a bit broken, but that doesn't even seem to be used
[21:10:54] <wjp> which parts of the screen are broken?
[22:05:44] <brad_a> same as before. the buttons and npcs etc
[22:05:53] <brad_a> inventory items
[22:06:03] <brad_a> controls of any kind really
[22:29:22] <-- lynxlynxlynx has left IRC (Remote host closed the connection)
[22:43:56] --> kettuz has joined #gemrb
[23:12:04] <-- SiENcE has left IRC (Quit: byebye Folks)
[23:54:40] <-- barra_home has left IRC (Quit: Verlassend)