#gemrb@irc.freenode.net logs for 8 Jan 2014 (GMT)

Archive Today Yesterday Tomorrow
GemRB homepage


[00:06:07] <-- raevol has left IRC (Ping timeout: 272 seconds)
[00:21:12] --> raevol has joined #gemrb
[00:47:43] <Pepelka> [wiki] pst_spells - it is more difficult than it should be to make a save like this http://www.gemrb.org/wiki/doku.php?id=pst_spells&rev=1389141887&do=diff
[01:00:10] <-- raevol has left IRC (Quit: Leaving.)
[01:11:53] <Pepelka> [wiki] pst_spells - [Mage 1] level 1 spells tested http://www.gemrb.org/wiki/doku.php?id=pst_spells&rev=1389143410&do=diff
[01:48:12] <Pepelka> [wiki] pst_spells - [Mage 2] level 2 spells tested. ouch. http://www.gemrb.org/wiki/doku.php?id=pst_spells&rev=1389145527&do=diff
[02:12:24] <Pepelka> [wiki] pst_spells - [Mage 3] level 3 spells tested http://www.gemrb.org/wiki/doku.php?id=pst_spells&rev=1389147137&do=diff
[02:13:37] <chiv> probably best I just stop there and try and fix some of this...
[02:17:31] <dolio> Looks like there's plenty. :)
[02:20:18] <chiv> a lot of the problems seem related and should go away together
[06:21:29] --> edheldil__ has joined #gemrb
[06:44:07] <-- edheldil__ has left IRC (Ping timeout: 272 seconds)
[06:53:34] <-- psch has left IRC (Remote host closed the connection)
[07:56:13] <fuzzie> optimistic
[08:04:49] --> lynxlynxlynx has joined #gemrb
[08:04:50] <-- lynxlynxlynx has left IRC (Changing host)
[08:04:50] --> lynxlynxlynx has joined #gemrb
[08:04:50] --- ChanServ gives channel operator status to lynxlynxlynx
[08:51:43] <-- edheldil has left IRC (Quit: Really?)
[08:51:50] <-- edheldil_ has left IRC (Quit: Really?)
[08:52:09] --> edheldil has joined #gemrb
[08:52:09] --- ChanServ gives channel operator status to edheldil
[08:59:05] <edheldil> lynxlynxlynx: "Cannot remove eye"
[09:00:08] <edheldil> my guess is that you can only replace him, similar to Annah's or FFG's clothes ;-)
[09:00:20] <edheldil> s/him/it/
[09:00:31] <lynxlynxlynx> where and which eye?
[09:01:21] <edheldil> Nameless One's
[09:01:46] <edheldil> I think you were asking me about it yesterday
[09:03:11] <lynxlynxlynx> where = original/pst, which = item resref
[09:03:37] <lynxlynxlynx> err s,pst,gemrb,
[09:05:16] <edheldil> in the original
[09:05:22] <edheldil> resref ...
[09:06:40] <edheldil> can't find it now, it's the (for example) caleidoscopic eye
[09:13:35] <lynxlynxlynx> ok
[09:13:42] <lynxlynxlynx> so you can just swap it with another eye
[09:14:03] <lynxlynxlynx> so chiv was wrong and i made another faulty commit ><
[09:17:51] <fuzzie> ..?
[09:19:06] <fuzzie> now I can remove the eyeball entirely
[09:20:00] <fuzzie> annah's armor is working ok though
[09:30:28] <edheldil> lynxlynxlynx: found EYEGLAS1, EYEGLAS2, EYEKAL, EYEKALEM, EYEVECNA
[09:42:11] <lynxlynxlynx> yeah, we have them
[10:10:24] --> Eli2 has joined #gemrb
[10:11:22] <-- Eli2_ has left IRC (Ping timeout: 245 seconds)
[10:24:14] <-- tomprince has left IRC (Ping timeout: 264 seconds)
[10:37:37] --> tomprince has joined #gemrb
[12:36:50] <-- WingedHussar has left IRC (Ping timeout: 264 seconds)
[12:37:36] --> WingedHussar has joined #gemrb
[13:39:35] <chiv> i dont get it. i said you can only replace an eye with another eye?
[13:47:55] <chiv> fuzzie: I am optimistic about the spells. apart from the silly final fantasy style ones most of them are pretty simple
[13:49:08] <chiv> now I can take out my eye. I should not be able to take out my eye.
[13:52:18] <chiv> http://picpaste.de/upload.php
[13:52:18] <Pepelka> PicPaste - public beta v5!
[13:52:20] <Pepelka> »PicPaste ist ein Loginfreier Service zum Bilder hochladen«
[13:52:40] <chiv> ach stupid keyboard http://picpaste.de/windows_user_here-Lk7TmIP0.jpg
[13:52:41] <Pepelka> PicPaste - windows_user_here-Lk7TmIP0.jpg
[13:52:42] <Pepelka> »PicPaste ist ein Loginfreier Service zum Bilder hochladen«
[13:53:07] <-- lynxlynxlynx has left IRC (Ping timeout: 246 seconds)
[13:57:08] <fuzzie> yes lynx broke it :p
[13:57:49] <chiv> but somehow it's my fault :) ?
[14:00:02] <fuzzie> 20:42 <chiv> important: grace and dakkon should never let you have their special items
[14:00:11] <fuzzie> ^- this is the only confusing thing I see in backscroll from you
[14:08:39] <chiv> meh it'll get fixed anyway, I was just reading the log
[14:13:48] <chiv> I'm not exactly the master of torment knowledge, when you guys ask me stuff it's just easier for me to look because I have a winxp netbook
[14:14:26] <chiv> and a million savegames
[14:14:48] <fuzzie> yes it's very much appreciated
[14:19:54] <chiv> and I appreciate all your patience with me, I know it's hard to see, but the game really is making huge progress
[14:23:39] --> brada has joined #gemrb
[14:34:00] --> lynxlynxlynx has joined #gemrb
[14:34:00] <-- lynxlynxlynx has left IRC (Changing host)
[14:34:00] --> lynxlynxlynx has joined #gemrb
[14:34:00] --- ChanServ gives channel operator status to lynxlynxlynx
[14:40:32] <-- edheldil has left IRC (*.net *.split)
[14:43:09] --> edheldil has joined #gemrb
[14:55:15] --> Beholder has joined #gemrb
[14:55:32] <Beholder> hi
[14:57:08] <fuzzie> hey Beholder
[14:57:35] <wjp> hi
[15:02:58] <edheldil> Hi, Beholder. Do you need a help with submitting your code to Github? :)
[15:03:32] <edheldil> first!
[15:03:40] <Beholder> i need to create my own repository, right?
[15:04:06] <edheldil> create an account and fork gemrb repository
[15:05:24] <edheldil> i.e. our gemrb repository, https://github.com/gemrb/gemrb
[15:05:26] <Pepelka> gemrb/gemrb · GitHub
[15:05:27] <Pepelka> »gemrb - Engine Made with preRendered Background«
[15:10:14] <brada> there is a handy little fork button in the top right if you arent already aware
[15:14:24] <edheldil> yes, sorry, that's what I meant
[15:28:59] <Beholder> registered as BehoIder
[15:29:14] <-- brada has left IRC (Quit: brada)
[15:29:38] * wjp predicts regular typos there :-)
[15:35:04] --> brada has joined #gemrb
[15:39:52] <-- brada has left IRC (Client Quit)
[15:43:19] <Beholder> what can i do for send my changes?
[15:44:37] <wjp> do you already have them committed locally?
[15:45:35] <Beholder> yes
[15:48:26] <wjp> what is your local repository setup at the moment? Do you have a clone of the gemrb/gemrb repo, or of the behoIder/gemrb repo?
[15:51:23] <Beholder> a master clone
[15:51:57] <Beholder> i need to push it to my repo
[15:52:01] <Beholder> right?
[15:52:04] <fuzzie> yes
[15:52:46] <Beholder> but what url?
[15:53:03] <fuzzie> it's the one on the front page of your repo
[15:53:24] <fuzzie> where it says 'clone url'
[15:53:33] <fuzzie> so, https://github.com/BehoIder/gemrb.git
[15:53:35] <Pepelka> BehoIder/gemrb · GitHub
[15:53:36] <Pepelka> »gemrb - Engine Made with preRendered Background«
[15:54:21] <-- Drakkar has left IRC (Ping timeout: 276 seconds)
[15:56:32] <fuzzie> I don't know if you can easily just push to that with the visual studio plugin
[15:56:40] <Beholder> pushed
[15:56:43] <fuzzie> but I guess you can :P
[15:56:46] <wjp> looks like it :-)
[15:56:53] <fuzzie> thanks very much, Beholder
[15:57:00] <Beholder> :)
[15:57:38] <Beholder> Matrix & Shader my own helper classes
[15:57:50] <Beholder> from another project
[15:58:05] <Beholder> and sorry for code formatting
[15:58:17] <fuzzie> heh, we can always fix that stuff later
[15:58:22] <fuzzie> most important is to make it *work*
[15:58:56] <Beholder> i can push Visual Studio project files to my repo
[16:00:39] <fuzzie> we've been trying to make it so that they're auto-generated from cmake I think
[16:00:51] <Beholder> i cant work with gemrb code formatting (braces I mean)
[16:00:56] <fuzzie> there are VS project files in another repo somewhere but they rot..
[16:03:18] <Beholder> my renderer class inherits SDL20Video
[16:04:39] <Beholder> can anyone try to build it?
[16:05:20] <Beholder> opengl es is a default, to use opengl needs to define USE_GL macro
[16:05:33] <Beholder> glew needed
[16:08:18] --> Drakkar has joined #gemrb
[16:11:56] <Beholder> but what about covers?
[16:12:02] <Beholder> :)
[16:13:20] <wjp> covers are good! :-)
[16:14:30] <wjp> can I clarify anything?
[16:16:33] <Beholder> what is cover? I see its a class with dimension fields, coords and pixel data
[16:17:18] <Beholder> but how it work?
[16:17:49] <wjp> cover is a mask bitmap that has a checkerboard pattern where sprites are hidden behind walls
[16:18:05] <wjp> they're generated dynamically while the map is rendered
[16:18:33] <fuzzie> do we have a screenshot?
[16:19:45] <fuzzie> hm I guess you can't render the mask
[16:20:35] <wjp> hm, I've had a cover debug mode at some point where the covered bits where drawn white
[16:20:48] <fuzzie> yes, I have that
[16:21:18] <fuzzie> http://fuzzie.org/covers_1.png has the cover polygons in yellow, and the covered bits in white
[16:21:52] <fuzzie> I was just wondering if you could render the mask too somehow, but I guess not trivially
[16:22:38] <wjp> the cover mask is not the same size as the sprite, but it's never smaller
[16:22:43] <Beholder> what stores in pixels?
[16:23:00] <fuzzie> pixels is bytes, 0 for not-covered (draw normally) and 1 for covered (do not draw)
[16:23:37] <fuzzie> wjp: the bit I had trouble explaining is how the sprite PosX/PosY interacts with the cover PosX/PosY
[16:24:57] <wjp> their respective hotspots correspond to the same point on-screen
[16:24:58] <fuzzie> pixels is cover->Width*cover->Height pixels, and basically you start at the pixel 'cover->XPos - sprite->XPos' in the x direction and 'cover->YPos - sprite->YPos' in the y direction
[16:25:14] <wjp> right
[16:25:50] <fuzzie> so if you did this by making a texture from pixels, you'd have to pass that offset to the shader
[16:25:56] <wjp> the sprite's rectangle is a sub-rectangle of the cover's rectangle starting at ( cover->XPos - sprite->XPos, cover->YPos - sprite->YPos )
[16:27:26] <-- WingedHussar has left IRC (Quit: WingedHussar)
[16:27:38] <Beholder> cover its a mask for sprite or for wall?
[16:27:42] <fuzzie> for sprite
[16:27:58] <fuzzie> each item on-screen has its own cover
[16:27:59] <wjp> the walls are drawn as part of the background tiles
[16:28:28] <Beholder> in original game i see actors semi transparent if they covered by walls
[16:28:35] <fuzzie> yes, that's stippled
[16:28:44] <fuzzie> look at http://fuzzie.org/covers_1.png to see the mask there
[16:29:12] <fuzzie> the mask used for the PC is semi-transparent, and the mask used for the guard is fully-transparent
[16:29:56] <fuzzie> (the white bits are the parts of the sprite which shouldn't be drawn, this is just a debug mode to help see)
[16:30:03] <Beholder> is a range 0-255 in cover
[16:30:11] <fuzzie> no
[16:30:32] <fuzzie> it's just 0 for visible, or 1 for non-visible
[16:31:06] <Beholder> i need to use halftrans?
[16:31:11] <fuzzie> no
[16:31:15] <fuzzie> it's not really trans
[16:31:28] <fuzzie> it's just a stippled mask (every other pixel is masked)
[16:34:01] <Beholder> actor looks semi-transparent and dotted in BG
[16:34:14] <fuzzie> yes, that's just a stippled mask, no..?
[16:34:20] <fuzzie> maybe it's also semi-transparent
[16:34:28] <fuzzie> I guess that depends on the render flags?
[16:34:33] <fuzzie> the tints can also be applied
[16:35:13] <fuzzie> I hadn't quite followed your question there
[16:35:20] <fuzzie> but you're trying to work out what the shader would have to do?
[16:35:26] <fuzzie> and the answer is, I think you have to support everything
[16:36:23] <Beholder> i added tinting support for all sprites
[16:41:11] <Beholder> may be a good idea to do it not stippled (not true but will looks nicer)
[16:41:54] <fuzzie> yes but that is a more difficult thing :)
[16:43:00] <fuzzie> and you have to handle the masks anyway, so it has to come first
[16:44:52] <Beholder> bg using a 30mb of ram instead a 150+ of gemrb :)
[16:45:44] <fuzzie> shouldn't be 150+ for gemrb
[16:45:55] <fuzzie> but we do some silly things, like converting MOS to 32bpp
[16:46:23] <Beholder> and i creating many 32bpp textures)
[16:46:37] <fuzzie> also doesn't help
[16:46:43] <fuzzie> they should almost all be 8bpp though
[16:46:51] <fuzzie> not sure if you want that or not
[16:47:52] <Beholder> not a problem for modern devices
[16:48:51] <fuzzie> the 32bpp ones?
[16:49:03] <Beholder> and memory
[16:49:29] <fuzzie> ok
[16:49:33] <fuzzie> then stay with that :)
[16:49:41] <fuzzie> the MOS files are actually 8bpp 'tiles'
[16:49:49] <fuzzie> but if we left them as 8bpp, then you'd have more sprites (one per tile)
[16:54:14] <Beholder> pixels is a full cover pixels, i need to skip some data to get mask covers sprite, right?
[16:54:31] <fuzzie> yes
[16:54:44] <fuzzie> as I said, you start at the pixel 'cover->XPos - sprite->XPos' in the x direction and 'cover->YPos - sprite->YPos' in the y direction
[16:55:06] <fuzzie> but that's different for every sprite and every frame, often
[17:13:06] --> kpederse1 has joined #gemrb
[17:16:33] <-- kpedersen has left IRC (Ping timeout: 272 seconds)
[17:30:10] <Beholder> can i add a new blit flag?
[17:30:50] <lynxlynxlynx> sure
[17:33:29] --> raevol has joined #gemrb
[17:34:19] <fuzzie> can we ask which one? :P
[17:35:24] <fuzzie> but if it makes things more convenient then shouldn't be a problem
[17:36:55] <Beholder> i need a flag for using external mask
[17:38:04] <Beholder> i see 0x100 is not used
[17:39:40] <Beholder> i need to define flag in video.h or may be internally in my renderer?
[17:55:22] <-- Beholder has left #gemrb
[17:58:08] <fuzzie> oh
[17:58:17] <fuzzie> well I would suggest video.h so we don't accidentally conflict with it
[17:58:23] <fuzzie> don't know if Beholder will see that
[17:58:41] <fuzzie> I assume he's constructing shaders using the flags, and a flag for the mask makes a lot of sense there
[18:02:22] --> Beholder has joined #gemrb
[18:02:47] <Beholder> why i get cover!=NULL if my sprite(actor) is not covered
[18:03:05] <Beholder> it causes actor fully invisible
[18:03:13] <fuzzie> how do you mean?
[18:03:26] <fuzzie> the cover you get will always cover the sprite
[18:04:15] <Beholder> i'll paste code first
[18:05:30] <fuzzie> it'll be all-zeros in the open world, I guess that is an obvious optimisation to make
[18:05:40] <fuzzie> it's just an optimisation though
[18:05:50] <Beholder> a
[18:06:12] <Beholder> but how i check sprite covers or not?
[18:06:34] <fuzzie> (and we should do it on our end, obviously)
[18:07:46] <fuzzie> but if you get a cover, it will *always* be ok for the sprite: 0 for visible, 1 for invisible
[18:08:19] <Beholder> http://pastebin.com/qSpE8wSL
[18:08:20] <Pepelka> [C++] GLuint coverTexture = 0; if (cover) { int trueX = cover->XPos - glSprite- - Pastebin.com
[18:08:40] <Beholder> this code creates a mask for cover
[18:09:12] <fuzzie> that doesn't work
[18:09:14] <Beholder> and if pixels are all zeros it cause actor totally invisible
[18:09:15] <fuzzie> because your input is bad
[18:09:22] <Beholder> hm
[18:09:26] <fuzzie> and zeros means visible.. :)
[18:09:30] <edheldil> hmm, could this 'stippled' flag be somehow reused to get pst's darken effect?
[18:09:38] <fuzzie> edheldil: no
[18:09:45] <fuzzie> stippled is stippled
[18:09:45] <edheldil> thought so :)
[18:09:51] <Beholder> )
[18:10:01] <fuzzie> Beholder: you can't just index into cover->pixels like that
[18:10:21] <fuzzie> Beholder: and if you do this like this, you will get a new texture *every frame* for most covers, is that a good idea?
[18:11:51] <Beholder> i do a more optimization future
[18:12:04] <fuzzie> if you want to do it this way, you have to copy line-by-line
[18:12:17] <Beholder> you are right
[18:12:24] <Beholder> damn
[18:12:32] <Beholder> memcpy needed
[18:12:35] <fuzzie> or set GL_UNPACK_ROW_LENGTH if you can do that in ES
[18:13:08] <Beholder> as cover width?
[18:13:10] <fuzzie> but no, it's optional
[18:13:16] <Beholder> heh
[18:13:25] <fuzzie> ES 2.0 only has it with GL_EXT_unpack_subimage
[18:19:58] <Beholder> no examples in google....
[18:20:32] <lynxlynxlynx> so what's the better solution?
[18:21:05] <fuzzie> you have to memcpy if you want to work without the extension
[18:21:17] <fuzzie> with the extension, yes, glPixelStorei with GL_UNPACK_ROW_LENGTH and cover->Width or something
[18:21:47] <fuzzie> and the better solution is passing the offset as data like a tex coord
[18:22:04] <fuzzie> but Beholder is right that it's an optimisation
[18:22:12] --> brada has joined #gemrb
[18:27:45] <Beholder> hm
[18:27:56] <Beholder> i added GL_UNPACK_ROW_LENGTH
[18:28:27] <Beholder> and actor is invisible if is not covered
[18:28:42] <Beholder> and normally draws if covered
[18:28:45] --> fizzle has joined #gemrb
[18:30:39] <fuzzie> so you're not treating zero as visible :)
[18:31:16] <fuzzie> I guess we can probably invert that logic if it's really annoying for you
[18:32:31] <Beholder> i think so
[18:36:12] <-- brada has left IRC (Read error: Connection reset by peer)
[18:36:34] --> brada has joined #gemrb
[18:47:32] <-- fizzle has left IRC (Read error: Operation timed out)
[18:48:08] --> fizzle has joined #gemrb
[18:49:54] --> Yoshimo has joined #gemrb
[18:51:53] <EVBB-Eugene>
[18:53:07] <Beholder> "cover" may be not a fully cover the sprite, right?
[18:53:15] <wjp> it always fully covers the sprite
[18:55:00] <wjp> it has an assert() for that in the software rendering code
[18:55:35] <Beholder> i adds error handling/asserting later
[18:56:31] <wjp> I meant that it really always covers it
[18:56:44] <wjp> and we double-check that with an assert
[19:01:08] <Beholder> i committed changes
[19:02:37] <Beholder> may be anybody try to build my source?
[19:05:16] <Beholder> http://i-fotki.info/16/b435697e361bd43b0a8a8f938359931abc8616171867957.jpg.html
[19:05:18] <Pepelka> Ifotki.info - бесплатный хостинг фотографий для форумов, дневников, гостевых книг
[19:05:20] <Pepelka> »Ресурс для публикации фотографий на форумах, дневника, личных страницах«
[19:06:15] <Beholder> i think a real semi-transparency will look better
[19:07:11] <wjp> tricky
[19:07:21] <chiv> ya tozhe tak dumayu ;)
[19:07:51] <wjp> problem is that an actor consists of multiple sprites (main sprite, helmet, shield, weapon, sometimes effects)
[19:08:25] <wjp> so you'd end up stacking transparency
[19:08:54] <wjp> maybe that won't matter
[19:10:19] <lynxlynxlynx> awesome
[19:10:30] <lynxlynxlynx> i'll have to try it another day though
[19:10:53] <wjp> I'm sadly not behind a development machine at the moment so can't try to build it either
[19:11:38] <chiv> i can't help with any technical stuff, but i can build it and see what happens
[19:23:19] <fuzzie> Beholder: looks good
[19:23:46] <fuzzie> and yes I'm not sure either way about the covering
[19:23:55] <fuzzie> you could render the whole actor to an intermediate texture
[19:24:36] <fuzzie> then render the intermediate texture to the screen and render the masked tile on top
[19:24:52] <fuzzie> (or just use a shader which can do both)
[19:25:12] <fuzzie> but it seems overly-complex at some point
[19:26:10] <wjp> will be a fun texture for a dragon :-)
[19:26:26] <fuzzie> video memory is cheap :)
[19:26:48] <wjp> might be worth trying to see how it looks if you "just" render them at 50% alpha and see what happens.
[19:27:15] <fuzzie> yes
[19:27:31] <-- brada has left IRC (Quit: brada)
[19:27:37] <wjp> but if e.g. three parts (actor, shield, vvc) overlap there'll only be 12.5% of the wall itself left
[19:28:01] <wjp> I don't know how it will look
[19:28:18] <fuzzie> what would we need, an extra blit flag, plus a video flag so we create the cover mask without stipple?
[19:28:44] <wjp> blit flag?
[19:29:13] <fuzzie> for 'render this stippled'
[19:29:37] <fuzzie> I mean to differentiate the 'transparent cover' vs 'full cover' cases
[19:29:51] <fuzzie> which I don't think we have another way of telling at video driver level atm?
[19:30:03] <fuzzie> oh we could just attach it to the sprite cover as a member variable
[19:30:32] <fuzzie> or do we do that already? what's cover->flags?
[19:30:41] <fuzzie> oh damn
[19:30:51] <fuzzie> it's polygon-specific
[19:31:44] <fuzzie> writing different values into the mask for different modes does not sound like the kind of thing which would be efficient at shader level?
[19:32:24] <wjp> wouldn't be so bad with 0/128/255?
[19:32:32] <chiv> call me dumb, but how do you switch to the opengl video?
[19:32:42] <wjp> (i.e., cover mask == alpha)
[19:32:50] <fuzzie> chiv: build with it enabled, which I guess involves messing with cmake
[19:32:56] <fuzzie> wjp: yes, I guess that would work
[19:33:20] <fuzzie> the BG2EE screenshots don't look like they stipple the cover
[19:33:34] <wjp> do you have a link?
[19:33:46] <fuzzie> I am using google image search for 'baldur's gate 2' on 'large' mode
[19:33:57] <fuzzie> if I find a good example I will link it
[19:33:58] <wjp> (firefox on these pixies tend to die with pages with too many images...)
[19:34:03] <fuzzie> ah
[19:34:26] <fuzzie> http://neptunegames.it/wp-content/uploads/2013/11/Baldurs-Gate-II-Enhanced-Edition-for-iOS-7.jpg maybe?
[19:34:50] <fuzzie> annoying jpeg artifacts
[19:35:19] <wjp> that does look transparent
[19:35:44] <fuzzie> http://www.dappersharkstudios.com/wp-content/uploads/2013/09/Baldurs-2.jpg clearer
[19:35:47] <wjp> (the ones on the bottom left are fairly clear)
[19:36:09] <wjp> right, bottom left on that one too :-)
[19:36:12] <fuzzie> I wanted one with a VVC
[19:36:48] <wjp> in the first one the VVC looks entirely cut off
[19:37:01] <wjp> while the feet are transparent
[19:37:31] <wjp> (or, well, are those spell effects VVC?)
[19:37:59] <fuzzie> http://images.eurogamer.net/2013/articles//a/1/6/3/6/2/7/1/5.jpg.jpg is clearly VVC but I guess those are non-transparent polys.. :)
[19:38:28] <fuzzie> but yes I was just looking for something clearer
[19:38:42] <wjp> the first one is not too bad though
[19:39:56] <wjp> nobody around here with BG2EE with a properly advanced party and effects galore? :-)
[19:40:12] <wjp> or, say, Avenger? :-)
[19:40:29] <fuzzie> heh, *asking* how it works is cheating :-)
[19:40:56] <fuzzie> I do rather like how it looks though
[19:41:17] <fuzzie> will definitely buy the android releases
[19:42:25] <wjp> I have the Windows versions of both BGEE and BG2EE
[19:42:36] <fuzzie> do they run under wine?
[19:42:37] <Beholder> 50% alpha for masking area?
[19:42:42] <wjp> haven't had the chance to try yet
[19:42:56] <wjp> or, well, didn't try yet :-)
[19:43:15] <Beholder> i may add a simple hack to simulate it
[19:43:22] <Beholder> an one line of code
[19:43:54] <Beholder> but only for dots(
[19:44:01] <fuzzie> Beholder: in core/Video.cpp, check the 'if (dither) {' bit
[19:45:05] <fuzzie> if you replace all of that with a memset of 128, and then make the other (non-dither) memset output 255, then wjp's suggestion is that you can just use the output as alpha
[19:45:12] <fuzzie> would that help?
[19:45:19] <wjp> this cover code was my very first contribution to gemrb :-)
[19:45:26] <wjp> looong time ago now :-)
[19:45:37] --> brada has joined #gemrb
[19:51:25] <wjp> fuzzie: one complication is that it should only overwrite higher values
[19:51:38] <wjp> (but not too important for testing)
[19:52:08] <wjp> (for overlapping cover polygons with different flags)
[19:53:42] <fuzzie> you mean, lower, right?
[19:53:49] <wjp> maybe
[19:54:10] <wjp> but wouldn't fully covered be 0?
[19:55:32] <fuzzie> um
[19:55:46] <fuzzie> well, in the current code, 0 is transparent
[19:55:52] <Beholder> i added support for semi-transparent masks
[19:56:06] <fuzzie> as in, no cover
[19:56:20] <wjp> fully transparent = fully covered though?
[19:56:49] <wjp> or maybe I'm just missing context
[19:59:29] <-- brada has left IRC (Quit: brada)
[19:59:59] <Beholder> we need to call method with cover with 0-128-255 values, not with 0-1
[20:00:37] <Beholder> if we want a semi-transparency
[20:01:05] <fuzzie> yes
[20:01:10] <Beholder> now i convert 1 to 0 and 0 to 255
[20:01:13] <fuzzie> I just explained above how to modify this :-)
[20:01:16] <fuzzie> go do it!
[20:03:53] <Beholder> i need add color key support for 32 bit sprites
[20:06:45] <Beholder> SetColorKey(ieDword index)
[20:07:00] <Beholder> what mode of index?
[20:07:10] <Beholder> RGBA? ARGB? etc.
[20:07:40] <fuzzie> the same one as in the sprite
[20:08:40] <Beholder> ARGB
[20:09:04] <Beholder> but not important
[20:09:35] <Beholder> if color masks are valid i can convert all in rgba
[20:09:55] <fuzzie> color masks should be valid
[20:14:39] <fuzzie> I wonder if any of the BG:EE/BG2:EE builds have helpful debug symbols
[20:15:05] <fuzzie> I suspect unfortunately they're not quite that foolish, and ofc the Windows one definitely doesn't
[20:18:13] <Beholder> hm, never called for 32 bit
[20:18:42] <Beholder> i get an missing color key on loading screen
[20:18:50] <Beholder> where i can fix it?
[20:19:49] <fuzzie> hm
[20:20:38] <fuzzie> bg1, right?
[20:21:21] <fuzzie> it's the last parameter to CreateSprite
[20:21:33] <fuzzie> in SDLVideoDriver, that calls SetColorKey
[20:23:39] --> brada has joined #gemrb
[20:26:16] <wjp> do these mask texcoords behave properly wrt flipping?
[20:27:13] <wjp> the cover mask should not be flipped if the sprite is flipped
[20:27:14] <Beholder> hm
[20:27:18] <Beholder> no, i do not check it
[20:29:19] <fuzzie> just talking to someone else
[20:29:31] <fuzzie> and I looked at pst fixpack
[20:29:40] <fuzzie> and it has *81* exe patches (to fix 31 issues)
[20:29:54] <wjp> *shudder*
[20:34:33] <Beholder> chiv works with pst?
[20:35:20] <chiv> I'm still failing at building it :/
[20:35:53] <fuzzie> oh
[20:36:13] <fuzzie> let me take a look
[20:36:15] <chiv> only because I am dumb though
[20:36:16] <fuzzie> you're using cmake, chiv?
[20:36:33] <chiv> yeah but I know nothing about it really
[20:36:44] <fuzzie> no need to be so self-deprecating :)
[20:36:44] <Beholder> chiv: you are in windows?
[20:37:02] <chiv> no, linux
[20:37:55] <chiv> i cant seem to set the include dir for gles
[20:38:02] <fuzzie> don't use gles
[20:38:02] <Beholder> I am here alone with windows :)
[20:38:07] <fuzzie> yes :P
[20:38:17] <Beholder> add USE_GL macro
[20:40:39] <Beholder> i added support for 32bpp sprites color keying
[20:41:49] <lynxlynxlynx> pass -DUSE_GL=666
[20:42:00] --> edheldil_ has joined #gemrb
[20:42:22] <Beholder> i am interesting about performance on modern PC's
[20:42:26] <fuzzie> well beholder's branch needs SDL2
[20:42:47] <fuzzie> so you have to add FindSDL2.cmake and switch to SDL2 renderer
[20:42:48] <Beholder> of cource
[20:43:11] <fuzzie> well SDL2 doesn't do 16bpp so it's been bad for performance previously :)
[20:44:53] <fuzzie> once you do that, you'll see it failing to build
[20:45:31] <fuzzie> which is brada's fault
[20:46:24] <fuzzie> and then you'll see msvc-isms!
[20:46:43] <brada> what did/didn't I do?
[20:46:49] <fuzzie> Beholder: there's an extra 'Shader::' in Shader.h
[20:47:03] <fuzzie> brada: SDL2 backend is C++11 incompatible
[20:47:21] <brada> care to be more specific? :p
[20:47:40] <fuzzie> no
[20:47:41] <fuzzie> lazy
[20:50:31] <fuzzie> let me make this build first
[20:51:05] <brada> fuzzie: dont worry about it
[20:51:16] <brada> i can build with c++11 and see for myself
[20:51:37] <fuzzie> well I can apply a fix maybe
[20:52:05] <brada> if you arent too lazy :p
[20:52:05] <lynxlynxlynx> we shouldn't be requiring 11 anyway
[20:52:11] <brada> we dont
[20:52:13] <fuzzie> [SDL 2 GL Driver]: Creating display
[20:52:14] <fuzzie> Segmentation fault
[20:52:15] <fuzzie> :-)
[20:52:16] <brada> but its nice to be compatible
[20:52:22] <Beholder> Shader.h is a helper class
[20:52:31] <fuzzie> Beholder: yes, and it fails to build, because it's invalid syntax
[20:52:32] <Beholder> damn
[20:52:42] <Beholder> oho
[20:52:45] <fuzzie> but you have other bugs too
[20:52:45] <Beholder> where?
[20:52:50] <fuzzie> I'll provide a diff once it works
[20:52:51] <Beholder> may be
[20:54:08] <brada> fuzzie: is the fix for c++11 to use static_cast<int> for the struct members?
[20:54:20] <fuzzie> brada: if you mean '(int)', yes
[20:54:30] <brada> that works too
[20:54:34] <fuzzie> 'fix' :-)
[20:54:45] <fuzzie> damn, why does this crash
[20:55:22] <Beholder> but why SEGV?
[20:55:33] <fuzzie> I expect it's brada's fault
[20:55:59] <Beholder> i used my own constructor
[20:56:15] <Beholder> and my own createdisplay
[20:56:56] <fuzzie> it's just crashing in SDL_CreateWindow
[20:57:34] <Beholder> SDL20GLVideo.cpp line 167
[20:57:41] <Beholder> thats my code
[20:57:48] <fuzzie> yes but there's no reason for a crash
[20:58:00] <Beholder> unsupported flags?
[20:58:18] <fuzzie> so I don't think it's your fault
[20:58:44] <fuzzie> oh, um
[21:00:50] <fuzzie> o.O
[21:01:22] <fuzzie> so _this is NULL but that's impossible
[21:01:47] <fuzzie> SDL 1.2 is loaded into my address space, grr
[21:01:59] <brada> fuzzie: propably my fault
[21:02:01] <brada> :D
[21:02:29] <fuzzie> oh nice
[21:02:32] <fuzzie> so I excluded OpenALAudio
[21:02:35] <fuzzie> erm, SDLAudio
[21:02:39] <fuzzie> but OpenALAudio links against SDL 1.2 too
[21:02:52] <brada> um, how?
[21:02:56] <brada> i build on ios with it
[21:03:09] <wjp> oh, for threads?
[21:03:10] <brada> and there is no SDL 1.2...
[21:03:21] <brada> so you should be able to use SDL 2 for everything
[21:03:34] <fuzzie> brada: yes but cmake does it :)
[21:03:38] <Beholder> hm, i excluded all audio too
[21:03:40] <fuzzie> I am now working!
[21:03:41] <brada> oh right
[21:03:45] <fuzzie> I have a very cool screen!
[21:03:46] <Beholder> not needed for me
[21:03:47] <chiv> i cant even get sdl2...
[21:03:50] <fuzzie> I will make a screenshot!
[21:03:55] <wjp> yay!
[21:04:20] <fuzzie> http://fuzzie.org/bg2_gl_sadface.png
[21:04:30] <wjp> oh, not so yay
[21:04:34] <fuzzie> :-)
[21:04:36] <fuzzie> hey it's something
[21:04:37] <Beholder> :)
[21:04:46] <brada> its super retro!
[21:04:49] <chiv> bg2 start screen?
[21:04:55] <fuzzie> chiv: yes
[21:05:00] <fuzzie> bg1 is even more broken sadly
[21:05:03] <chiv> for atari 2600
[21:05:11] <fuzzie> ok can I get these patches applied maybe?
[21:05:57] <Beholder> can you debug?
[21:06:08] <fuzzie> http://pastebin.com/1p2KF7rY <- patches
[21:06:09] <Pepelka> diff --git a/CMakeLists.txt b/CMakeLists.txt index 07fbd60..912167d 100644 --- - Pastebin.com
[21:06:50] <fuzzie> for Beholder I guess the relevant ones are the GLTextureSprite2D.cpp ones, the Shader.cpp ones and the Shader.h ones
[21:07:14] <fuzzie> I guess the #pragmas should go inside a MSVC check
[21:11:05] <fuzzie> Beholder: hmm I thought it would be shader problems but they compile fine
[21:12:02] <fuzzie> ah, valgrind finds it
[21:12:51] <Beholder> no errors in message?
[21:13:04] <fuzzie> nvidia driver is reading off the end of the buffer
[21:13:07] <Beholder> program and shaders handles are valid?
[21:13:10] <fuzzie> yes
[21:13:23] <fuzzie> texture creation issue somehow..
[21:13:40] <-- Yoshimo has left IRC (Quit: Yoshimo)
[21:15:04] <Beholder> hm
[21:16:34] <fuzzie> I debug
[21:18:48] <fuzzie> I don't get it though
[21:19:10] <wjp> does GL_ALPHA exist for GL?
[21:19:16] <wjp> I can only find it in GLES ref docs
[21:19:34] <Beholder> works with GLEW
[21:19:37] <fuzzie> well, valgrind complains on the RGBA version
[21:20:29] <Beholder> you may change GL_ALPHA to any 8 bit one and add simply modification to a shaders
[21:20:45] <fuzzie> I guess that would be GL_RED?
[21:20:50] <fuzzie> I will try that
[21:20:53] <Beholder> right
[21:21:03] <Beholder> change .a to .r in shaders
[21:21:17] <Beholder> GL_RED not supported by GLES
[21:21:31] <fuzzie> for index?
[21:21:50] <Beholder> yes
[21:22:07] <fuzzie> nope
[21:22:13] <fuzzie> doesn't help
[21:22:14] <wjp> the diagonals are weird
[21:22:23] <fuzzie> they look like your typical pitch issue, right?
[21:22:27] <wjp> yes
[21:22:31] <wjp> except... where?
[21:22:46] <fuzzie> and the nvidia driver reading off the end of buffers, that would also match that
[21:22:52] <fuzzie> but I also don't see :)
[21:22:56] <wjp> (minor thing I just noticed: "delete buffer;" instead of "delete[] buffer;")
[21:23:15] <fuzzie> yes, valgrind whines about that too :) thanks
[21:23:26] <wjp> (same for mask later)
[21:23:40] <fuzzie> and some other ones too in the other file
[21:24:04] <fuzzie> I don't reach those though
[21:25:53] <fuzzie> the video works fine, mind
[21:26:05] <Beholder> video do not use gl
[21:26:09] <fuzzie> ah, drat
[21:26:25] <Beholder> is the SDL20Video part
[21:26:43] <fuzzie> probably this is a really obvious bug
[21:26:46] <fuzzie> but I don't see it right now
[21:27:17] <Beholder> i'll try to run in my work pc tomorrow
[21:27:28] <Beholder> windows 8 x64
[21:27:50] <brada> we should merge things anyway to make it easier for everybody to test
[21:27:52] <Beholder> core i5 with nvidia
[21:27:52] <fuzzie> I thought it would be a refcounting issue
[21:28:20] <fuzzie> brada: well right now it sabotages the SDL2 driver :)
[21:28:27] <brada> oh drat
[21:28:39] <Beholder> no
[21:28:40] <brada> i mean that should be the goal once everything works
[21:28:41] <fuzzie> otherwise I would
[21:28:50] <Beholder> it uses it for video and input)
[21:30:11] <Beholder> i plan to merge SDL20Video with my driver
[21:30:17] <Beholder> in future
[21:30:46] <fuzzie> it's really odd
[21:31:00] <fuzzie> it creates a (192,768) 32bpp texture and then it crashes reading off the end of it
[21:31:50] <wjp> it doesn't make any sense that the diagonal looks about the same everyhwere
[21:32:50] <wjp> could there be a few hidden glPixelStore params set?
[21:32:52] <fuzzie> ah maybe valgrind is lying to me, wait a moment
[21:33:24] <fuzzie> I get a 'conditional jump or move depends on uninitialised values' first
[21:33:44] <wjp> inside glTextImage2D?
[21:33:50] <wjp> s/xt/x/
[21:33:54] <fuzzie> invalid stack there
[21:34:01] <fuzzie> let me try track-orgins
[21:35:00] <fuzzie> ah heh
[21:36:52] <wjp> do tell :-)
[21:38:42] <fuzzie> hm
[21:39:22] <fuzzie> well, the copy constructor wasn't reiniting glPaletteTexture
[21:39:31] <fuzzie> obvious cause, right?
[21:40:11] <wjp> oops
[21:40:13] <fuzzie> except (a) it didn't help (b) there's no way that could've been it, GetPaletteTexture checks pal==currentPalette
[21:40:26] <fuzzie> oh I guess the first version doesn't
[21:40:29] <fuzzie> anyway it still didn't help :/
[21:40:41] <fuzzie> valgrind just says it's now caused by a malloc() deep inside the nvidia driver instead
[21:40:52] <brada> sounds yucky
[21:41:02] <wjp> which glTexImage2D call exactly?
[21:41:25] <fuzzie> well
[21:41:36] <fuzzie> as I said, invalid stack
[21:41:39] <fuzzie> so I can't see where the problem is
[21:41:43] <wjp> oh
[21:41:54] <wjp> hum
[21:42:01] <fuzzie> but the buffer that it reads off the end of, is the one in the RGBA case
[21:42:16] <fuzzie> and by using print statements, I see it happens rather suspiciously when the glTexImage2D happens
[21:42:35] <wjp> in createGlTexture or createGlTextureForPalette?
[21:42:46] <fuzzie> the first
[21:43:16] <fuzzie> GemRB::GLVideoDriver::blitSprite -> GemRB::GLTextureSprite2D::GetTexture -> GemRB::GLTextureSprite2D::createGlTexture -> operator new[]
[21:44:10] <wjp> but could there be persistent wrong glPixelStores?
[21:45:10] <fuzzie> yes but .. how?
[21:45:13] <wjp> SDL?
[21:45:29] <fuzzie> but we call glPixelStorei right there with the only param that matters I thought
[21:45:40] <fuzzie> oh drat, I should listen to you
[21:45:44] <fuzzie> this little tab here is the opengl es docs
[21:45:48] <fuzzie> no wonder I was confused
[21:46:21] <fuzzie> one grep later, sigh
[21:46:28] <fuzzie> src/render/opengl/SDL_render_gl.c: renderdata->glPixelStorei(GL_UNPACK_ROW_LENGTH, (pitch / 2));
[21:47:07] <wjp> and no convenient matching 0 I suppose
[21:47:18] <fuzzie> thankyou :p
[21:47:25] <fuzzie> doesn't help, mind
[21:47:31] <wjp> argh :-)
[21:47:38] <wjp> identical output?
[21:47:41] <fuzzie> but it was a lot better than my theory
[21:48:06] <wjp> next random idea: coordinate normalization problems?
[21:48:12] <fuzzie> the output is a bit random
[21:48:27] <fuzzie> but it looks a lot better now
[21:48:54] <fuzzie> and valgrind is happier
[21:49:37] <Beholder> are you using glew?
[21:49:42] <fuzzie> yes
[21:50:30] <edheldil_> lookslike github is DoSed by people downloading Beholder's plugin
[21:51:41] <Beholder> i don't know what may helps with textures(
[21:52:00] <fuzzie> oh I get a nice broken bg1 mouse cursor now
[21:52:09] <fuzzie> definitely pitch issue still
[21:52:14] <Beholder> it's a typical texture creating code
[21:52:39] <wjp> could you do an updated screenshot?
[21:53:25] <Beholder> i can try to run bg2, but i think it will work fine for me
[21:53:36] <fuzzie> http://fuzzie.org/bg1_gl_sadface.png
[21:53:51] <wjp> nice
[21:53:53] <lynxlynxlynx> :D
[21:54:07] <fuzzie> I think it's brilliant
[21:54:07] <lynxlynxlynx> looks like a mockup for a palette editor
[21:54:14] <fuzzie> this is great
[21:54:19] <wjp> I'm not convinced it's pitch
[21:54:30] <fuzzie> well the single-colour portraits etc are interesting
[21:54:34] <wjp> the colours being constant on the diagonals is too strange
[21:54:37] <fuzzie> yes
[21:54:53] <lynxlynxlynx> so the main menu stuff worked ok or did you just guess-enter the game?
[21:54:55] <fuzzie> it looks a bit different in other places but this really is suspicious
[21:54:58] <fuzzie> lynxlynxlynx: guess-enter
[21:54:58] <wjp> any texture coordinate normalization differences between GLES and GL?
[21:55:04] <fuzzie> yes
[21:55:55] <lynxlynxlynx> please don't overwrite that file, i'll post it to the forum
[21:56:00] <fuzzie> hehe
[21:56:41] <wjp> broken BUFFER_OFFSET or such could also have interesting effects
[21:56:45] <chiv> can we patch the sound driver to replace all the sounds with high pitched beeping?
[21:56:55] <fuzzie> wjp: right, I'm just looking at that right now
[21:57:33] <fuzzie> but it looks fine right?
[22:01:19] <lynxlynxlynx> done, it's there as an attachment, overwrite at will
[22:02:21] <fuzzie> oh
[22:02:26] <fuzzie> sigh
[22:03:21] <fuzzie> the texture coords are going out-of-scope :)
[22:03:34] <fuzzie> all magic now
[22:03:37] <wjp> argh
[22:03:41] <wjp> of course
[22:03:55] * wjp should've spotted that
[22:04:06] <fuzzie> colour-keying fails
[22:05:23] <Beholder> :(
[22:06:39] <fuzzie> precision issues..?
[22:07:24] <Beholder> precisions for gles only
[22:08:43] <Beholder> problem with paletted textures?
[22:08:54] <fuzzie> yes
[22:10:08] <Beholder> i use ieDword index, but for palette it must be Uint8, try a cast it to Uint8 when on createGlTextureForPalette
[22:11:19] <Beholder> and in loop set i to Uint8 too
[22:11:26] <fuzzie> that shouldn't be a problem
[22:12:13] <Beholder> I do not believe anything
[22:12:46] <Beholder> are you fix problem with textures?
[22:14:33] <fuzzie> the cover doesn't work either
[22:15:37] <Beholder> I'm not surprised
[22:15:58] <wjp> did you do the alpha/red thing for the mask?
[22:16:06] <fuzzie> yes (just now)
[22:16:16] <fuzzie> but mostly this is using plain fragmentPal
[22:16:28] <fuzzie> without mask
[22:16:55] <wjp> how does the colour keying look?
[22:17:01] <fuzzie> bright green
[22:17:17] <wjp> weird
[22:17:20] <Beholder> a right color
[22:17:30] <fuzzie> yes, just the alpha..
[22:17:39] <fuzzie> probably I do something really really stupid
[22:17:44] <Beholder> check for setting alpha in texture generation
[22:17:51] <fuzzie> good idea
[22:18:21] <fuzzie> it always sets alpha for every palette
[22:18:37] <Beholder> for palleted sprites is a first index typically
[22:19:03] <fuzzie> yes it's always zero
[22:19:07] <fuzzie> which is why I wondered about precision
[22:19:37] <fuzzie> ok, I set alpha to always zero and it doesn't help
[22:20:12] <Beholder> hm
[22:20:50] <Beholder> look at line 231 on SDL20GLVideo.cpp
[22:20:57] <Beholder> and 230 too
[22:20:59] <fuzzie> I think alpha is being ignored
[22:21:04] <fuzzie> I set it to zero in shader, no difference
[22:21:13] <fuzzie> right :)
[22:21:20] <fuzzie> oh heh
[22:21:21] <fuzzie> !
[22:21:21] <Beholder> this not works if blending disabled
[22:21:24] <fuzzie> you do it only at startup
[22:21:34] <Beholder> right
[22:21:49] <Beholder> not needed every frame
[22:22:07] <fuzzie> that was it
[22:22:20] <fuzzie> I put it in swapbuffers too
[22:22:30] <Beholder> damned nvidia
[22:22:38] <fuzzie> no, it is SDL2
[22:22:50] <fuzzie> it messes with this when it plays videos
[22:23:00] <brada> yes
[22:23:03] <Beholder> oh
[22:23:23] <fuzzie> I guess you probably disabled them? :)
[22:23:28] <Beholder> right)
[22:23:35] <fuzzie> it still doesn't quite work
[22:23:50] <Beholder> i'll fix this
[22:23:59] <fuzzie> well, I have all these fixes locally
[22:24:01] <Beholder> but what about textures
[22:24:04] <fuzzie> so try not to change too much :p
[22:24:37] <Beholder> what's wrong with them?
[22:24:44] <fuzzie> is it easiest for me just to push my changes to my repo?
[22:24:50] <fuzzie> or give me access to yours and I can push there?
[22:25:59] <Beholder> how to?
[22:26:25] <brada> cant you send him a pull request?
[22:26:32] <fuzzie> oh dammit it's all broken atm anyway
[22:26:36] <fuzzie> yes I'll send a pull request I guess
[22:26:40] <fuzzie> but github is being broken atm
[22:26:45] <fuzzie> I can do it tomorrow
[22:26:47] <fuzzie> morning
[22:27:01] <Beholder> ok, i go to sleep now
[22:27:14] <fuzzie> yes I have to go do stuff too
[22:27:18] <fuzzie> really nice work
[22:27:18] <Beholder> but ask me what are you do with textures
[22:27:42] <fuzzie> oh, the tex coordinates are bad in the C++ side
[22:27:50] <fuzzie> the pointer is invalid
[22:27:55] <Beholder> oO
[22:27:58] <Beholder> why?
[22:28:30] <fuzzie> because you do 'GLfloat *coords; if (something) { GLfloat mycoords = { <stuff> }; coords = mycoords; }'
[22:28:36] <Beholder> i create an array and set pointer to it
[22:28:40] <fuzzie> but mycoords is *invalid* as soon as you get to the '}'
[22:28:58] <fuzzie> it no longer points to your coords, it points to some random bit of the stack
[22:29:04] <brada> heh
[22:29:15] <Beholder> not true for msvc
[22:29:17] <fuzzie> so I just moved the four arrays outside the 'if' statements :)
[22:29:24] <fuzzie> are you in Release mode?
[22:29:29] <Beholder> no
[22:29:33] <fuzzie> it should happen in msvc in Release mode :)
[22:29:34] <Beholder> on debug
[22:29:41] <fuzzie> anyway I will give you patch tomorrow
[22:29:46] <fuzzie> or you can fix it yourself
[22:29:46] <Beholder> ok
[22:29:48] <fuzzie> simple enough :)
[22:29:53] <Beholder> what fps?
[22:30:03] <fuzzie> how do I tell?
[22:30:05] <Beholder> i'll fix it
[22:30:36] <Beholder> enable fps in config and comment out delay
[22:30:59] <fuzzie> the last problem is
[22:31:29] <fuzzie> that in fragmentPal, the 'texture2D(s_mask, v_texCoord).r' is always zero
[22:31:43] <fuzzie> so I just changed it to always use v_alphaModifier to make it work
[22:32:26] <Beholder> you use GL_RED?
[22:32:29] <fuzzie> yes
[22:33:03] <fuzzie> in first page of bg1 chargen, I get 153fps
[22:33:21] <fuzzie> in menu I get 1250fps
[22:33:30] <Beholder> i know thats deprecated in newer versions, try GL_LUMINANCE
[22:33:36] <Beholder> good
[22:34:00] <fuzzie> in complex bg2 area I get 90fps
[22:34:07] <fuzzie> needs optimisation :)
[22:34:13] <Beholder> right)
[22:34:16] <fuzzie> but I don't carry on now
[22:34:21] <fuzzie> I have to go do stuff before I can sleep
[22:34:24] <fuzzie> really good work
[22:34:31] <Beholder> thank you
[22:34:40] <fuzzie> will look for alternatives to GL_RED tomorrow
[22:35:04] <Beholder> i do not know why GL_ALPHA not work for you
[22:35:25] <Beholder> bye
[22:35:37] <-- Beholder has left IRC (Quit: Beholder)
[22:43:18] <-- fizzle has left #gemrb
[22:44:07] <brada> I havent looked at the code yet, but I doubt touch input will work immediately with this driver since it depends on SDL_RenderGetScale and SDL_RenderSetLogicalSize
[22:46:15] <chiv> i have altogether far too much time on my hands these days: http://forums.gibberlings3.net/index.php?showtopic=26017
[22:46:17] <Pepelka> Video overview of development in 2013 - The Gibberlings Three Forums
[22:46:18] <Pepelka> »GemRB development in 2013 as seen by gource in our main repository: This year was a bit slower than the previous one, but everyone got hyperactive...«
[22:47:08] <chiv> actually it was a better fit for this picture: http://fuzzie.org/bg2_gl_sadface.png
[23:04:23] --> i30817 has joined #gemrb
[23:05:02] <i30817> you know what time it is?
[23:05:04] <i30817> It's time for my annual request for gemrb to incorporate a ToEE engine.
[23:07:23] <i30817> Become a scummvm of isometric rpg you might say.
[23:07:40] <Lightkey> depends on whether you have Portuguese versions of LucasArts adventure games, maybe then ;p
[23:18:18] <-- brada has left IRC (Quit: brada)
[23:36:10] <edheldil_> how should I compile Beholder's work? cmake -DUSE_GL=666 .. ; make ?
[23:50:15] <-- edheldil_ has left IRC (Ping timeout: 257 seconds)
[23:56:03] --> edheldil_ has joined #gemrb