#gemrb@irc.freenode.net logs for 30 Dec 2013 (GMT)

Archive Today Yesterday Tomorrow
GemRB homepage


[00:58:31] <-- murray has left IRC (Quit: Leaving)
[04:17:51] <-- brada has left IRC (Quit: brada)
[04:40:34] --> brada has joined #gemrb
[05:50:24] --> Eli2 has joined #gemrb
[06:19:27] <-- brada has left IRC (Quit: brada)
[07:13:21] --> brada has joined #gemrb
[07:30:11] <-- brada has left IRC (Quit: brada)
[09:21:07] --> WingedHussar has joined #gemrb
[09:38:27] --> lynxlynxlynx has joined #gemrb
[09:38:27] <-- lynxlynxlynx has left IRC (Changing host)
[09:38:27] --> lynxlynxlynx has joined #gemrb
[09:38:27] --- ChanServ gives channel operator status to lynxlynxlynx
[09:45:46] --> mrugiero has joined #gemrb
[09:49:40] <-- mrugiero has left IRC (Client Quit)
[09:50:00] --> mrugiero has joined #gemrb
[09:50:18] <-- mrugiero has left IRC (Client Quit)
[10:02:55] <-- WingedHussar has left IRC (Quit: WingedHussar)
[10:09:48] --> Eli2_ has joined #gemrb
[10:11:35] <-- Eli2 has left IRC (Ping timeout: 245 seconds)
[10:16:40] <-- |Cable| has left IRC (Ping timeout: 246 seconds)
[10:25:24] <fuzzie> so do we concentrate on pst now?
[10:28:26] <lynxlynxlynx> partially, through chiv's push
[10:29:51] --> |Cable| has joined #gemrb
[10:30:05] <fuzzie> hm
[10:30:32] <fuzzie> I really know very little about the game
[10:30:49] <fuzzie> oh yes, I remember what my problem is, I don't have the data
[10:31:24] <fuzzie> I have three copies of the 4 CD version here, and no optical drive.
[10:32:50] <fuzzie> but trying to add proper support for the pst animations might be a fun thing for January
[10:32:59] <fuzzie> maybe I should ask chiv about priorities though
[10:34:27] <lynxlynxlynx> that's one of the related wierd things he noticed
[10:34:40] <lynxlynxlynx> some actors have wrong palettes, but no pattern yet
[10:34:54] <lynxlynxlynx> could just be wrong types, but not enough data yet
[10:35:04] <fuzzie> pst actors are 99% one single type
[10:35:30] <fuzzie> so I have "burn the pst avatars.2da" on my todo
[10:38:41] <lynxlynxlynx> it was not clown colors, so something is likely wrong with our resdata stuff
[10:53:33] <wjp> weren't we also wrongly assuming all cycles in an actor's BAM have the same palette or something like that?
[10:57:50] <wjp> scanning through the logs, we briefly talked about that possibility on 26 Feb 2012
[10:58:04] <fuzzie> thx
[10:59:35] <fuzzie> the code in question is rather hideously complicated
[10:59:42] <wjp> yes
[11:02:34] <fuzzie> if we're going to do this whole opengl thing maybe we should think about notifying the backend about palettes too anyway
[11:03:40] <wjp> if this different-palette-per-cycle thing only shows up on sprites without clown colours it might be relatively doable
[11:04:07] <fuzzie> trying to boot my netbook which should have pst on it
[11:04:19] <fuzzie> every time I do, it unhibernates, then gives me a warning that it's about to hibernate, then hibernates
[11:04:20] <wjp> would just have to bypass the whole CharAnimation palette mess for those sprites then
[11:04:26] <fuzzie> this is inconvenient
[11:04:38] <wjp> "It's still winter, don't wake me up yet!" ?
[11:05:09] <fuzzie> now it's telling me that it's on the LIACS network. sigh.
[11:09:03] <fuzzie> searching for 'pst' finds me three gemrb working copies and zero copies of the data. drat.
[11:10:32] <fuzzie> ah, it's in backups/pst on the windows drive.
[11:10:39] <fuzzie> victory.
[11:10:56] --> chiv has joined #gemrb
[11:11:47] <chiv> hi all
[11:12:01] <fuzzie> hm I don't have weidu either
[11:14:19] <fuzzie> Fall-From-Grace is certainly pretty normal in the resdata
[11:15:22] <chiv> if we're talking about the wrong actor palette thing, oddly it only ever seems to be the idle sequence that is messed up
[11:15:31] <fuzzie> probably that's coincidence
[11:15:41] <fuzzie> but I'd have to check the BAMs
[11:16:38] <chiv> oh i got lots of the missing animations into the game btw
[11:17:09] <fuzzie> every time you use avatars.2da for pst, a kitten cries
[11:17:21] <fuzzie> did you hack up the stance thing?
[11:17:43] <chiv> i just followed the existing method in charanimations.cpp
[11:18:08] <fuzzie> I mean, MAX_ANIMS
[11:18:11] <fuzzie> and looks like you did
[11:18:14] <chiv> had to bumb max_anims upto 35
[11:18:24] <fuzzie> imo it's totally the wrong thing to do
[11:18:41] <fuzzie> but as you say in a comment, 'meh' :)
[11:19:04] <chiv> if you can suggest a right way, im open to that
[11:19:39] <fuzzie> as I said the other day, use resdata.ini instead
[11:19:56] <fuzzie> it's not beautiful either though
[11:20:18] <fuzzie> but for e.g. the spellcasting stances this is all so very complicated anyway
[11:21:02] <fuzzie> what's the deal with the death check? did someone break it again?
[11:22:09] <chiv> I haven't fully looked at it but it snatches most deaders away before they even hit the ground
[11:23:00] <chiv> so its probably a problem elsewhere, but that is as far as i traced it to yesterday
[11:23:04] <fuzzie> obviously skipping the function breaks everything though
[11:23:41] <fuzzie> is there ever a legitimate reason for removign an actor?
[11:23:41] <chiv> if it seems strange to commit that - the commit is much more reliable than my memory
[11:23:53] <fuzzie> I think there isn't
[11:24:23] <chiv> i think they are supposed to disappear after a day of being a corpse
[11:24:31] <fuzzie> yes, but we handle that on map change
[11:24:46] <fuzzie> I don't think they disappear in front of you
[11:24:55] <fuzzie> I guess that's easy to check in original if you can just sleep a few times
[11:27:58] <fuzzie> probably this was originally intended to handle chunking but that's actually handled by switching animation
[11:29:18] <chiv> i can't find this resdata.ini
[11:29:55] <lynxlynxlynx> they're individual files afaik
[11:30:28] <lynxlynxlynx> what i'd like to know is what happens in the original if you visit the top/destination portal in that ragpicker's save
[11:30:53] <chiv> the dialog opens straight in front of you
[11:31:40] <lynxlynxlynx> that's odd, the other infopoint has no scripted dialog
[11:31:43] <fuzzie> chiv: can't find it..?
[11:31:52] <fuzzie> I mean, it's in inifiles.bif by default, you have to extract it with something
[11:32:36] <chiv> oh yeah, that's right... I thought this was a gemrb thing for a second
[11:34:45] <chiv> if you are talking about the smaller one that is completely inside the large one, that is the actual portal, which torment.exe draws sparkles and such around the edges
[11:35:06] <lynxlynxlynx> no, where it throws you out
[11:35:37] <chiv> hmm.. it has a script though right?
[11:35:53] <lynxlynxlynx> yes, but only for the portaling
[11:36:31] <fuzzie> what's the script names?
[11:36:48] <lynxlynxlynx> 0101prt1 and 2
[11:36:49] <chiv> are we looking at the same thing? the actual doorway has nothing afaik, but the floor trap in front of it handles the travel
[11:37:20] <lynxlynxlynx> yeah
[11:37:35] <lynxlynxlynx> what i'm trying to say is, portals are two-way
[11:37:54] <lynxlynxlynx> what happens if you return in the original the opposite way?
[11:37:56] <fuzzie> only the standard to0100 has a special-case there
[11:38:38] <lynxlynxlynx> from the dialog itself, it seems more appropriate for the dialog to trigger on the top one, even though it's assigned to the bottom one
[11:38:53] <chiv> the traps activate the portal regions which are then just standard travel regions
[11:39:02] <chiv> with exit points etc
[11:39:14] <chiv> until they time out again
[11:40:35] <chiv> from_portal leads to fr0101b and to portal leads to fr0101a
[11:44:41] <lynxlynxlynx> so nothing special
[11:44:47] <chiv> hmm.. resdata.ini looks insane. I think I must have repressed the memory of reading it...
[11:45:05] <chiv> yeah the portals themselves are fine, just the activation dialog isn't working
[11:46:27] <fuzzie> resdata.ini is basically trivial
[11:46:39] <fuzzie> you have one section for every animation id
[11:47:19] <fuzzie> and it contains the speed (walkscale/runscale), the sounds (plus which frame they fire on), some misc info, and the filename for every supported stance
[11:47:22] --> WingedHussar has joined #gemrb
[11:47:52] <chiv> i understand it, I just don't get why they made it when the avatars all have a uniform naming scheme...
[11:48:04] <fuzzie> well, why do we have more than one pst anim type?
[11:48:27] <fuzzie> (ignoring GHOST which is special-cased)
[11:49:44] <fuzzie> that's what I'd like to be rid of
[11:50:00] <chiv> does this get used at all yet, for the hit sounds etc?
[11:50:07] <fuzzie> no
[11:50:16] <fuzzie> we read some data from it
[11:50:36] <fuzzie> the walkscale, the runscale and the bestiary
[11:53:13] <fuzzie> and you'd expect the sounds to be dealt with like the iwd ones
[11:54:15] <fuzzie> right, in Actor::GetSoundFromINI
[11:54:18] <chiv> I'll have to just keep it in mind, I've only just started looking at these parts and I can't think of a good way to use that yet
[11:54:26] <fuzzie> which handles at1sound/hitsound/dfbsound
[11:54:31] <fuzzie> so ok I guess 'yes'
[11:54:35] <fuzzie> it's been a long time :)
[11:54:38] <chiv> oh well thats a good start
[11:54:56] <fuzzie> but CharAnimations is a huge mess
[11:55:09] <fuzzie> I'd much rather dehardcode the pst version to use resdata if this is ok
[11:56:31] <chiv> but I don't know a good entry point off hand - the function I hacked is called a jillion times a second so that would be no good
[11:56:39] <fuzzie> but it's not vital, it's only modded to add one single anim id and we can add that ourselves
[11:58:02] <chiv> maybe saving the sequence names in the actor?
[11:59:46] <fuzzie> I didn't look at what you did
[12:09:46] <-- exultbot has left IRC (signing off...)
[12:11:24] --> exultbot has joined #gemrb
[12:11:24] --- Topic for #gemrb is: GemRB 0.8.0 | http://gemrb.org | Something wrong? State your exact version and CHECK THE GEMRB LOG | Be wary of your thoughts for there are Illithid present: http://log.usecode.org/gemrblog.php
[12:11:24] --- Topic for #gemrb set by lynxlynxlynx!~quassel@sourcemage/warlock/lynxlynxlynx at Sat Apr 27 10:20:08 2013
[12:11:33] <fuzzie> once upon a time I carefully decoded the whole standard hierarchy in the bg2 engine
[12:11:57] <fuzzie> only to discover that I'd picked the wrong branch at the top and I was actually looking at a broken version of the bg1 anim system. alas.
[12:12:16] <fuzzie> (which is still used in bg2 for some small subset I think..)
[12:12:16] <lynxlynx> heh
[12:12:34] <lynxlynx> yeah, plus some iwd-style too iirc
[12:12:36] <fuzzie> but I have to do that for the right branch sometime, because it has all the data for blood colour etc
[12:13:04] <fuzzie> unless that's unhardcoded in bgee? I didn't look at it at all yet
[12:14:28] <lynxlynx> we have blood color data already, at least for some - bloodclr.2da
[12:14:34] <-- WingedHussar has left IRC (Ping timeout: 252 seconds)
[12:14:39] <fuzzie> is that all of them?
[12:14:39] <fuzzie> I gave avenger the vtable so maybe he went through and did it
[12:14:39] <lynxlynx> we're lacking shadows, speed, color for chunking (if it is different from blood?) and perhaps more
[12:14:50] <fuzzie> I wonder if he can give me his last idb before being eaten by the beamdog hive
[12:14:52] --> WingedHussar has joined #gemrb
[12:16:57] <lynxlynx> most of the data is from 2010 and we have the whole range covered, though some of it by the default
[12:17:07] <lynxlynx> so it's hard to say whether it is complete or not
[12:17:45] <lynxlynx> no data for pst, if it cared at all
[12:18:08] <-- edheldil has left IRC (Ping timeout: 246 seconds)
[12:18:36] <fuzzie> avenger says that fuzzie says there's chunk color too
[12:18:42] <fuzzie> so I guess there's chunk color too
[12:19:13] <fuzzie> and yes I guess avenger dug out the blood color values from my work
[12:19:36] <fuzzie> so that should be complete
[12:21:14] <lynxlynx> oh, is something still wrong with CombatCounter? I saw we have a bug reminder from you on the sf tracker
[12:21:38] <lynxlynx> 1. (DONE?) fix combat round timings <-- and this in the TODO
[12:21:58] <fuzzie> um
[12:22:03] <fuzzie> yes
[12:22:22] <fuzzie> we don't actually handle the round timings properly right now at all I think?
[12:22:26] <fuzzie> unless someone added it
[12:22:33] <fuzzie> e.g. the times for weapons to launch and for animations to trigger and so forth
[12:24:51] <fuzzie> if not, then: what's what I would mean by that
[12:25:01] <lynxlynx> never thought about it
[12:25:10] <fuzzie> it's a completely insane system :)
[12:25:15] <fuzzie> but I know how it works!
[12:25:18] <lynxlynx> i'll fix the last TODO item and finally move it online
[12:25:34] <chiv> jumping on the bandwagon, some spells in pst take ages to cast
[12:26:00] <fuzzie> I have a 100-page paper pad here full of notes
[12:26:09] <fuzzie> pst casting speed is probably a whole different story though
[12:26:41] <chiv> I'll find out if nobody else does first anyway
[12:26:47] <fuzzie> pst spells are terrifying :)
[12:27:00] <chiv> hah, cos of the visual fx?
[12:27:08] <fuzzie> because they're all v.complicated hardcoded things
[12:27:43] <lynxlynx> chiv: we don't care about the casting animation or robe of vecna and similar would have no effect
[12:27:57] <lynxlynx> fuzzie: in dutch, huh?
[12:28:23] <fuzzie> no, so I could scan them, but they're pretty incoherent
[12:29:15] <lynxlynx> i'll let you be the judge of their usefulness
[12:31:35] <lynxlynx> oh, now i remmber another thing from animations: footstep sounds may be there too
[12:31:36] <fuzzie> but they're useful for me :)
[12:31:37] <fuzzie> ah yes
[12:33:39] <fuzzie> http://fuzzie.org/cgameanim_vtable.txt is the list :)
[12:33:41] --> vampi-the-frog has joined #gemrb
[12:38:53] <fuzzie> I don't know how much of it is relevant but I think annoyingly all of those are overridden
[12:39:35] <lynxlynx> GetCastHeight <-- another thing we only estimate now
[12:39:43] <fuzzie> if any of that info is immediately useful then I can pull it out in january. but it's somewhat time-consuming.
[12:40:17] <fuzzie> Avenger was always better at it than I was.
[12:41:41] <lynxlynx> the walking speeds are the holy grail there
[12:42:08] <lynxlynx> we still have 9 or frame[0] everywhere
[12:42:52] <fuzzie> I am going to be in big trouble if I don't pass my jan 6th exams so I am not looking until then ;p
[12:43:04] <fuzzie> it's the kind of thing where you lose days digging all the data out
[12:43:55] <chiv> I hadn't actually realised you were guessing so many things
[12:44:30] <fuzzie> well, 'guessing
[12:44:30] <fuzzie> '
[12:44:43] <fuzzie> if you ask then we can tell you what's based on reality and what isn't
[12:45:38] <fuzzie> but yes, that's why I say that you should just change stuff if it's obviously different in the original by observation, we're not magic :)
[12:45:40] <chiv> just pointing it out, haven't played bg2 for a while but apparently I was satisfied with the 'best guess' version
[12:46:02] <fuzzie> well, often we use the one which is right for most anims
[12:46:11] <fuzzie> but there's a lot of special corner-cases for individual animations in non-pst
[12:47:30] <fuzzie> and in some places they're very obvious :/
[12:51:54] <chiv> I guess I just didn't notice or possibly wilfully ignoring them
[12:53:36] <chiv> the only thing that felt strange to me as a player was the actors rigidly sticking to certain walk vectors instead of having 360 degree freedom
[13:00:11] <chiv> in fact the only reason I started looking at the code in the first place is that is felt a bit _too_ faithful, heh
[13:00:58] <chiv> anyway, off to work, laterz
[13:05:35] <-- chiv has left IRC (Ping timeout: 272 seconds)
[13:09:08] <lynxlynx> that's probably "just" an animation issue
[13:09:20] <lynxlynx> i noticed the same recently for the eyestalks in tob
[13:11:09] --> edheldil has joined #gemrb
[13:11:14] --- ChanServ gives channel operator status to edheldil
[13:29:44] <lynxlynx> +a bit more structure
[13:29:53] <Pepelka> [wiki] todo - added stuff from TODO http://www.gemrb.org/wiki/doku.php?id=todo&rev=1388409777&do=diff
[13:30:04] <-- edheldil has left IRC (Ping timeout: 272 seconds)
[13:49:01] <Pepelka> [wiki] todo - added styled recheck marks http://www.gemrb.org/wiki/doku.php?id=todo&rev=1388411219&do=diff
[14:07:10] <Pepelka> [wiki] todo - added research tag for when we're lacking RE data http://www.gemrb.org/wiki/doku.php?id=todo&rev=1388412182&do=diff
[14:36:17] <lynxlynx> huh, suddenly more stuff to prod brad about
[15:08:20] --> brada has joined #gemrb
[15:18:21] <-- exultbot has left IRC (signing off...)
[15:19:57] --> exultbot has joined #gemrb
[15:19:57] --- Topic for #gemrb is: GemRB 0.8.0 | http://gemrb.org | Something wrong? State your exact version and CHECK THE GEMRB LOG | Be wary of your thoughts for there are Illithid present: http://log.usecode.org/gemrblog.php
[15:19:57] --- Topic for #gemrb set by lynxlynxlynx!~quassel@sourcemage/warlock/lynxlynxlynx at Sat Apr 27 10:20:08 2013
[15:35:16] <lynxlynx> console text is not visible
[16:22:52] <brada> strange, i can see it broke in 16c81f7, but the code looks right to me
[16:27:23] <-- vampi-the-frog has left IRC (Quit: Leaving)
[16:32:10] <brada> its not as if much changed in console there
[16:32:19] <brada> how can i not see the problem?
[16:35:20] <fuzzie> um
[16:38:10] <fuzzie> where do you set Changed on the console?
[16:38:10] <brada> its not that
[16:38:10] <fuzzie> since you changed the Draw code so that it no longer draws in the non-Changed case
[16:38:10] <brada> i can see it entering drawinternal
[16:38:10] <fuzzie> oh, nutty
[16:38:10] <brada> and its drawing the background
[16:38:10] <brada> just not the text
[16:38:10] <brada> plus console always returns true for needsredraw
[16:38:10] <brada> because it doesnt have a window :p
[16:38:10] <fuzzie> isn't that irrelevant?
[16:38:10] <brada> sort of
[16:38:10] <brada> i check the window in the base case
[16:38:10] <brada> i just moved the existing check that was in the other controls, i think it has something to do with the PST float window
[16:38:10] <fuzzie> ye
[16:38:10] <brada> but certainly irrelevant to the console problem
[16:38:25] <fuzzie> well if it hits DrawInternal, yes
[16:38:43] <brada> right
[16:38:51] <fuzzie> although you changed the behaviour there, right?
[16:39:12] <fuzzie> it was checking an Owner flag and now it checks for owner Validity too
[16:39:36] <brada> ?
[16:39:48] <fuzzie> you check '!Changed && Owner && <stuff>'
[16:39:48] <brada> oh that isnt needed
[16:39:52] <brada> i should remove it
[16:39:55] <fuzzie> rather than '!Changed && (!Owner || <stuff>)'
[16:42:14] <fuzzie> or whatever is right
[16:42:14] <fuzzie> my suggestion is dumb toio
[16:42:14] <fuzzie> anyway irrelevant also if the console has no Owner
[16:42:14] <fuzzie> as you said
[16:42:14] <brada> well its irrelevant but not dumb
[16:42:14] <brada> also that line doesnt exist in HEAD
[16:42:14] <fuzzie> oh well
[16:42:14] <brada> but surely i am blind today becasue there isnt a lot of code to look over :/
[16:42:14] <-- exultbot has left IRC (Ping timeout: 240 seconds)
[16:42:14] <-- exultbot has left IRC (signing off...)
[16:43:28] --> exultbot has joined #gemrb
[16:43:28] --- Topic for #gemrb is: GemRB 0.8.0 | http://gemrb.org | Something wrong? State your exact version and CHECK THE GEMRB LOG | Be wary of your thoughts for there are Illithid present: http://log.usecode.org/gemrblog.php
[16:43:28] --- Topic for #gemrb set by lynxlynxlynx!~quassel@sourcemage/warlock/lynxlynxlynx at Sat Apr 27 10:20:08 2013
[16:43:33] --> raevol has joined #gemrb
[16:43:52] <fuzzie> are you sure your commit removing all the Invalidate calls is good btw?
[16:44:43] <brada> reasonably
[16:45:09] <brada> that happened after console broke if thats what you are suggesting
[16:45:12] <fuzzie> no
[16:45:18] <fuzzie> it's just a bit of a weird commit
[16:46:05] <fuzzie> I really have to trust you on the commit which adds MarkDirty() too, since it changes the logic
[16:48:47] <brada> fuzzie: it could be that the invalidate should stay and the markdirty call should be removed
[16:49:26] <brada> i cant think of a situation where we need to redraw the entire window for a single control changing
[16:49:37] <brada> except for auto scrolling text
[16:49:39] <fuzzie> when the control overlaps
[16:49:48] <brada> where does that happen?
[16:49:50] <fuzzie> esp. when the control is partially transparent
[16:50:08] <fuzzie> the usual test is the ticking clock in one of the games
[16:50:50] <brada> i know the levelup icon still "works"
[16:50:53] <fuzzie> with the UI hidden and/or visible
[16:51:00] <fuzzie> I forget which game
[16:51:10] <brada> probably one i dont have
[16:51:15] <brada> or at least not installed
[16:51:30] <fuzzie> which do you have?
[16:51:38] <brada> BG2 and IWD2
[16:51:43] <fuzzie> pst is the real problem but it was buggy anyway
[16:51:46] <fuzzie> but bg2 shuld show it
[16:52:10] <brada> well if its pst chiv will bug me soon enough
[16:52:23] <brada> i dont see any problems with BG2...
[16:52:32] <brada> but i have established i am blind :(
[16:53:14] <fuzzie> it should be pretty obvious if you hide/show the gui on that sid
[16:53:19] <fuzzie> if it works you didn't break it though ;p
[16:54:16] <brada> I did try to get you to look at that patch btw
[16:54:59] <fuzzie> which one?
[16:55:06] <fuzzie> I mean the mark dirty stuff is just impossible to review as committed
[16:55:09] <fuzzie> because you do two things at once
[16:55:36] <fuzzie> the invalidation one I'm just trusting you on, you're more of a dev than I am :)
[16:56:20] <fuzzie> the console is being created at the right position
[16:56:41] <brada> yeah, the background draws as expected
[16:56:45] <fuzzie> but you're drawing it in the wrong place now
[17:00:34] <fuzzie> kind of hilarious this ever worked
[17:01:10] <fuzzie> you can see that it used to draw the background at y but the font at y+YPos
[17:02:00] <brada> i already saw that but its not relevant (in bg2)
[17:02:01] <fuzzie> but your new Control::Draw is using a region 'y + XPos' where y is an unsigned short -> promotion to int
[17:02:17] <fuzzie> but y is negative!
[17:02:21] <brada> the background im talking about is just the black fill
[17:02:32] <brada> oh
[17:02:34] <fuzzie> and your refactoring has removed the old hack in the draw code which cast y back to a signed shrt
[17:02:38] <fuzzie> -> broken
[17:02:44] <fuzzie> (I did not test this theory)
[17:02:47] <brada> so it worked before because of the cast
[17:02:54] <brada> ?
[17:02:58] <fuzzie> that's my theory
[17:03:08] <fuzzie> can't test it but it looks right?
[17:03:26] <fuzzie> my real question is .. why on earth do we pass a y location for the draw at all?
[17:03:40] <fuzzie> why not just put it in the right place and then draw at (0,0)?
[17:03:53] <brada> is 0,0 the top?
[17:03:55] <brada> i forget
[17:03:58] <fuzzie> yes
[17:04:07] <brada> well so we need 0, height of console
[17:04:20] <fuzzie> do we?
[17:04:29] <brada> if you dont it will draw offscreen
[17:04:38] <brada> wont it
[17:04:39] <brada> ?
[17:04:40] <fuzzie> can't you just set YPos to the right pos?
[17:04:46] <brada> oh
[17:04:48] <brada> probably
[17:04:54] <fuzzie> it just seems overcomplex as it is
[17:04:56] <brada> i forget about that
[17:05:01] <fuzzie> nice refactor though
[17:05:44] <brada> thanks
[17:05:48] <brada> i hope you mean it
[17:05:53] <brada> cuz sometimes i dont know :p
[17:06:43] <fuzzie> yes
[17:06:58] <fuzzie> and yes, #gemrb always gets grumpy`tired`fuzzie, it's a bit sad :/
[17:08:18] <brada> thank you thogh for figuring that out for me
[17:08:34] <brada> i was too lazy to get into the debugger
[17:08:41] <brada> havent had coffee yet
[17:09:24] <brada> i dont know why it didnt occur to me as i was refactoring to give the console a good frame that makes the code much better
[17:09:48] <lynxlynx> if you will be changing the position, maybe put the whole thing a line lower, since the fps counter overlaps
[17:10:05] <lynxlynx> or just make it draw later or move the other one or something else
[17:12:23] <brada> i think i have an idea for that
[17:12:38] <brada> or maybe just move the FPS counter to the bottom
[17:12:54] <brada> the console moved to the top so we can use it on ios/android
[17:16:59] <lynxlynx> oh right, it was below before
[17:17:41] <lynxlynx> while there, i think refocusing isn't working (anymore? I think fizzle fixed it at one point)
[17:18:04] <lynxlynx> iow when closing it, gamecontrol is not refocused, so keyboard shortcuts don't work
[17:18:19] <brada> hmm
[17:18:35] <brada> strange
[17:18:41] <brada> because the console never gets focus
[17:18:48] <brada> we hack input to it :(
[17:19:00] <brada> we should give it its own window or something
[17:19:08] <brada> but thats beside the point
[17:19:36] <brada> also it works for me o_O
[17:19:41] <brada> or i dont understand the problem
[17:28:55] <lynxlynx> it doesn't work always, noticably if you teleport to a different area
[17:29:03] <lynxlynx> mta("ar1000") for example
[17:29:07] --> edheldil has joined #gemrb
[17:29:07] --- ChanServ gives channel operator status to edheldil
[17:29:26] <brada> ah i see
[17:31:35] <brada> still works here....
[17:35:19] <lynxlynx> i'm testing with space for pause
[17:36:56] <lynxlynx> the key is recognised, just not handled by the right control
[17:58:01] <brada> ok i can reproduce that
[18:14:06] <fuzzie> brada: it's always nice to try diagnosing things by inspection, it at least means I usually understand the code better afterwards
[18:14:18] <fuzzie> and I can do it from anywhere which also helps..
[18:15:18] --> pugachevat has joined #gemrb
[18:19:56] <-- pugachevat has left IRC (Remote host closed the connection)
[18:25:35] <-- edheldil has left IRC (Ping timeout: 272 seconds)
[18:29:08] --> edheldil has joined #gemrb
[18:29:08] --- ChanServ gives channel operator status to edheldil
[19:16:36] --> Yoshimo has joined #gemrb
[19:41:51] <-- Yoshimo has left IRC (Ping timeout: 240 seconds)
[20:13:56] <lynxlynx> splendid new year everybody, i'll be away until it happens
[20:15:34] <fuzzie> enjoy :)
[20:25:32] <wjp> have fun
[20:25:54] <wjp> FFG's palette issues are gone if I disable CharAnimation palettes entirely
[20:26:12] <wjp> so that strongly suggests it's indeed caused by different cycles having different palettes
[20:57:49] <edheldil> lynxlynx: where did the TODO go?
[21:01:31] <lynxlynx> wiki
[21:01:42] <lynxlynx> half of the stuff was useless anyway
[21:01:58] <lynxlynx> http://www.gemrb.org/wiki/doku.php?id=todo
[21:02:00] <Pepelka> todo [GemRB wiki]
[21:04:37] <edheldil> eh, I believed the commit message, silly me :)
[21:05:02] <fuzzie> please don't move anything to sf that doesn't need to be there :-p
[21:05:26] <lynxlynx> you know my stance on this :)
[21:07:22] <edheldil> veeery funny
[21:11:01] <-- Eli2_ has left IRC (Quit: Ex-Chat)
[21:12:09] <fuzzie> the world map seems to be missing exits?
[21:15:58] <fuzzie> I mean, local map
[21:19:15] <fuzzie> brada: you kind of broke the map drawing :P
[21:19:42] <brada> ok
[21:19:51] <fuzzie> it needs to always redraw
[21:20:03] <brada> world map?
[21:20:06] <fuzzie> local map
[21:20:17] <fuzzie> I can just override NeedsDraw to always return true?
[21:20:48] <brada> that wont fix it if its the problem lynx and i were discussing
[21:20:53] <fuzzie> oh
[21:20:58] <fuzzie> what is the problem you and lynx were discussing?
[21:21:09] <brada> the points of interest are missing
[21:21:15] <fuzzie> yes, well, that is annoying
[21:21:24] <fuzzie> but I'd just like it to redraw at all
[21:22:07] <brada> i can appreciate that :)
[21:22:15] <brada> im not seeing a problem, however
[21:22:18] <fuzzie> well
[21:22:22] <fuzzie> set some actors moving across the map
[21:22:23] <fuzzie> then open the map
[21:22:27] <brada> ah
[21:22:28] <fuzzie> then watch their positions not update :)
[21:22:29] <brada> yup
[21:22:39] <brada> didnt think to check that honestly
[21:22:46] <fuzzie> it works fine if you do the NeedsDraw thing
[21:22:46] <brada> i didnt even know it did that!
[21:23:13] <brada> neat feature that i never knew about
[21:23:16] <lynxlynx> one of the nice novelties in bg2 - game wasn't paused during mapviewing
[21:23:17] <fuzzie> hehe
[21:23:21] <fuzzie> yes :)
[21:23:32] <fuzzie> anyway is this a good fix brad
[21:23:44] <fuzzie> or should I be trying to invalidate it when actors move or something?
[21:23:53] <brada> im thinking
[21:24:04] <lynxlynx> %15 at least
[21:24:07] <brada> i mean its not a big deal to just redraw all the time in that view
[21:24:31] <fuzzie> and what is the problem with the points of interest?
[21:24:38] <brada> they dont draw at all
[21:24:41] <fuzzie> I don't really mind that they don't appear
[21:24:49] <fuzzie> but it's way annoying that it's more broken than that
[21:24:57] <fuzzie> and I don't understand why that is
[21:26:36] <brada> i didnt look into it yet
[21:26:48] <lynxlynx> the label for their names doesn't update properly either
[21:29:48] <brada> i would say its best to just do NeedsDraw returns true for it then
[21:30:27] <brada> we could try to work out when the game is paused, but meh
[21:30:39] <brada> same goes for gamecontrol in that regard
[21:31:44] --> Beholder has joined #gemrb
[21:31:52] <Beholder> hi
[21:32:46] <brada> hey
[21:32:59] <brada> pull out all your hair yet?
[21:35:26] <Beholder> no)
[21:35:34] <brada> good :)
[21:36:53] <Beholder> there many works that i need to do before new year begins
[21:38:18] <Beholder> about gemrb, ugly fonts is a palette issue
[21:38:36] <wjp> hi
[21:38:51] <brada> yeah
[21:38:55] <brada> you mean the green?
[21:39:02] <brada> the colorkey right?
[21:39:07] <Beholder> right
[21:39:19] <Beholder> i'll pos some screenshots now
[21:39:21] <Beholder> post
[21:40:33] <brada> cool
[21:43:16] <Beholder> http://i-fotki.info/16/6407196f5b298919b6b9934663832da0bc8616171099842.png.html
[21:43:17] <Pepelka> Ifotki.info - бесплатный хостинг фотографий для форумов, дневников, гостевых книг
[21:43:18] <Pepelka> »Ресурс для публикации фотографий на форумах, дневника, личных страницах«
[21:43:22] <Beholder> look at this
[21:44:01] <Beholder> the first - palette applied (no transparency), the second without palette
[21:44:56] <Beholder> what are you think?
[21:45:28] <Beholder> color key does not broke image
[21:46:16] <Beholder> i see a bad palette
[21:46:28] <wjp> it's a bit hard to tell with the jpeg compression, but it looks like the pixels with red channel 0xFF disappear?
[21:46:40] <brada> depends what you mean by bad palette
[21:47:03] <wjp> could there be an off-by-one where you only set 255 instead of 256 palette colours?
[21:47:12] <brada> wjp the red is due to the shader
[21:47:13] <wjp> or some other special handling of palette index 255?
[21:47:18] <wjp> brada: I know
[21:47:20] <brada> that is the image with no palette
[21:47:22] <brada> oh
[21:47:22] <Beholder> wjp, image in png, scaled with nearest mode
[21:47:23] <brada> sorry
[21:47:34] <wjp> Beholder: the image host turns it into jpeg probably
[21:47:43] <Beholder> damn
[21:47:57] <wjp> but it does look like the missing pixels have index 255
[21:48:03] <Beholder> red image is an indexes saved in red channel
[21:49:34] <Beholder> i think so, 255 index is missing
[21:49:39] <brada> its really hard to tell in that image but its something to look at
[21:49:49] <Beholder> min
[21:49:53] <Beholder> i repost
[21:51:08] <Beholder> https://dl.dropboxusercontent.com/u/80162785/Untitled-2.png
[21:51:25] <Beholder> in dropbox it looks properly
[21:52:02] <-- Beholder has left #gemrb
[21:52:03] <wjp> hm, could it be wrapping in the palette texture access?
[21:52:06] <wjp> oh
[21:52:24] --> Beholder has joined #gemrb
[21:52:41] <wjp> could it be wrapping in the palette texture access?
[21:53:04] <wjp> (since it has GL_TEXTURE_WRAP set)
[21:53:44] <Beholder> think do not
[21:54:04] <wjp> but why would it become green otherwise?
[21:54:09] <Beholder> i'll examine the palette source now
[21:54:20] <wjp> (green is palette index 0 right?)
[21:54:22] --> Yoshimo has joined #gemrb
[21:54:25] <Beholder> right
[21:54:40] <wjp> maybe quickly try with clamp instead of wrap to see if that affects it?
[21:55:42] <Beholder> no effect, i tried
[21:56:08] <wjp> weird; I really don't understand where the green comes from then
[21:58:16] <Beholder> odd, palette has an index 255 its value a 255 in all channels (fps text)
[21:58:36] <wjp> that doesn't sound very green
[21:58:57] <wjp> could you maybe put the indexed shader you're using online?
[22:00:18] <Beholder> const char* fragmentPal =
[22:00:18] <Beholder> "uniform sampler2D s_texture; // own texture \n"
[22:00:18] <Beholder> "uniform sampler2D s_palette; // palette 256 x 1 pixels \n"
[22:00:18] <Beholder> "varying vec2 v_texCoord;\n"
[22:00:18] <Beholder> "void main()\n"
[22:00:18] <Beholder> "{\n"
[22:00:18] <Beholder> " vec4 index = texture2D(s_texture, v_texCoord);\n"
[22:00:19] <Beholder> " gl_FragColor = texture2D(s_palette, vec2(index.r, 1.0));\n"
[22:00:19] <Beholder> "}\n";
[22:01:40] <wjp> hm, isn't that off by one half?
[22:02:02] <wjp> (i.e., shouldn't it be vec2(index.r + 0.5, 0.5) ? )
[22:02:49] <wjp> oh, are they normalized, by the way?
[22:03:05] <wjp> my glsl knowledge is a bit poor
[22:04:33] <Beholder> it normalized, right i may change 1.0 to 0.0 or -1.0 its the same
[22:04:46] <Beholder> texture has 1 pixel height
[22:05:00] <Beholder> gles does not support 1D textures
[22:06:03] <wjp> but the x coordinate runs from 0.0 to 256.0 right?
[22:06:30] <Beholder> 0 to 255
[22:06:36] <Beholder> 255 max
[22:20:34] <fuzzie> my first thought would also be to make sure you clamp on the S coord
[22:21:36] <fuzzie> for the palette texture, obv.
[22:26:03] <Beholder> damn damn damn
[22:26:11] <Beholder> i do it!
[22:27:06] <Beholder> in gles glsl uses precisions
[22:27:25] <Beholder> it default is lowp for textures
[22:29:10] <Beholder> and if in color conversion i get value closer to 0 it converts to 0
[22:30:17] <fuzzie> is that a problem..?
[22:31:13] <Beholder> i set a highp for float manually
[22:31:47] <Beholder> http://i-fotki.info/16/e12195d4b4b191708e83400b30b195e1bc8616171102749.jpg.html
[22:31:47] <Pepelka> Ifotki.info - бесплатный хостинг фотографий для форумов, дневников, гостевых книг
[22:31:48] <Pepelka> »Ресурс для публикации фотографий на форумах, дневника, личных страницах«
[22:32:06] <fuzzie> ok well I'm confused, but, good? :)
[22:32:15] <Beholder> yeah)
[22:32:31] <Beholder> but fps.....
[22:32:32] <fuzzie> you might want to look into the 1bpp texture thing
[22:32:39] <wjp> hurray :-)
[22:32:41] <Beholder> i do
[22:32:47] <fuzzie> the fps will have to wait until we rewrite things, I think
[22:33:14] <wjp> is it better in screens with not much text?
[22:35:17] <fuzzie> although really it shouldn't be *that* slow, should it?
[22:35:53] <brada> isnt that is some simulator?
[22:37:06] <Beholder> in menu with a small text i got >150fps
[22:37:55] <Beholder> i set up a 8 bit texture (GL_ALPHA) but it cause a visual glitch
[22:38:30] <fuzzie> hm, texture uploads are that expensive?
[22:38:47] <fuzzie> don't you have to use the non-alpha one?
[22:39:17] <Beholder> http://i-fotki.info/16/17c3242066ff8d4849a00d2e751d5c91bc8616171103193.jpg.html
[22:39:17] <Pepelka> Ifotki.info - бесплатный хостинг фотографий для форумов, дневников, гостевых книг
[22:39:18] <Pepelka> »Ресурс для публикации фотографий на форумах, дневника, личных страницах«
[22:39:30] <Beholder> GL_LUMINANCE looks same
[22:39:42] <Beholder> i cant use GL_R
[22:39:43] <fuzzie> are you padding them to 2 byte alignment?
[22:40:17] <Beholder> GL_alpha is a 8 bit per pixel
[22:40:21] <fuzzie> or whatever the rule is
[22:40:23] <Beholder> one byte
[22:40:40] <fuzzie> no, I mean per line
[22:40:49] <Beholder> hm
[22:41:10] <fuzzie> I don't know what the alignment is, the opengles spec says UNPACK_ALIGNMENT
[22:41:37] <fuzzie> I guess it's 4 by default?
[22:41:53] <Beholder> i think so
[22:41:57] <fuzzie> but you can set it to 1
[22:43:56] <-- lynxlynx has left IRC (Remote host closed the connection)
[22:46:58] <Beholder> right, i fixed it by glPixelStorei
[22:47:52] --> nutron has joined #gemrb
[22:48:10] <fuzzie> so does that help?
[22:49:14] <Beholder> yes, work fine now
[22:49:34] <fuzzie> I mean, performance-wise
[22:49:35] <fuzzie> I assume not
[22:49:42] <fuzzie> but if it's upload bandwidth limited then maybe?
[22:50:42] <wjp> but if it's strongly dependent on amount of text, we can assume it'll get magically fixed when we rewrite fonts?
[22:50:56] <Beholder> got an extra 3 fps:)
[22:51:11] <brada> not magiacally, but we will have an easier time deciding which direction to go
[22:51:13] <Beholder> right
[22:51:17] <fuzzie> wjp: well but in-game we'll be doing just as much work..
[22:51:24] <brada> in the meantime beholder still has to write shaders for game spritees
[22:51:31] <brada> so that should keep him busy till then :)
[22:52:28] <Beholder> stop
[22:52:43] <Beholder> i can do some optimisation
[22:53:14] <Beholder> it helps with performance i think
[22:54:01] <fuzzie> hm
[22:54:14] <fuzzie> you only upload the texture for each glyph once, right?
[22:54:51] <fuzzie> so.. it's really not that many draws
[22:55:40] <Beholder> the problem that palette is changes very often during rendering
[22:55:43] <Beholder> why?
[22:55:55] <brada> you mean for text?
[22:56:00] <Beholder> i need to yes
[22:56:02] <brada> to change the font color
[22:56:15] <fuzzie> the palette, or the Palette object?
[22:56:18] <Beholder> hm
[22:57:14] <fuzzie> we .. really don't change the palette much at all
[22:57:41] <fuzzie> although if you store the palette with the Sprite2D, that will probably go badly
[22:58:06] <Beholder> i disabled generate new palettes during rendering, but fps so small to
[22:58:19] <Beholder> 22 instead of 15
[22:58:51] <Beholder> this is not a bottleneck
[22:59:10] <fuzzie> what is the bottleneck, just the rendering?
[23:00:05] <Beholder> i'll think about it tomorrow
[23:00:09] <fuzzie> ok
[23:00:13] <fuzzie> looks great though
[23:00:13] <Beholder> but now i go to sleep
[23:00:18] <fuzzie> good work!
[23:00:27] <Beholder> 3.00 AM in Russia
[23:00:39] <brada> goodnight
[23:00:50] <wjp> good night
[23:01:06] <Beholder> my eyes closed by itself
[23:01:12] <Beholder> bye
[23:01:21] <-- Beholder has left #gemrb
[23:01:24] <brada> do you think it would much to render to a texture as a sort of backbuffer so you dont need to draw all controls every frame?
[23:13:28] <-- Yoshimo has left IRC (Quit: Yoshimo)
[23:29:14] <brada> is anybody still awake?
[23:29:31] <wjp> yes
[23:30:04] <wjp> fuzzie: I just tried this palette shader here in a small test program, and I get the same behaviour with GL_REPEAT, but not with GL_CLAMP
[23:30:26] <wjp> multiplying the red channel with 255.0/256.0 seems to improve things
[23:36:51] <wjp> no, that scale factor is wrong
[23:37:06] <wjp> (get the wrong palette entries that way)
[23:37:35] <brada> wjp: do you have a min to look at this patch?
[23:39:56] <brada> https://github.com/bradallred/gemrb/compare/gemrb:master...master
[23:39:58] <Pepelka> Comparing gemrb:master...bradallred:master · bradallred/gemrb · GitHub
[23:40:00] <Pepelka> »gemrb - Game Engine Made with preRendered Background«
[23:46:24] <edheldil> hmmm, polish Planescape is dubbed to Polish :o)
[23:47:04] <wjp> right, so the texture coordinates of that 256x1 texture are indeed spread over [0.0, 256.0)
[23:47:45] <wjp> so to get a [0.0,1.0] color value into there it's a (0.5+r*255.0)/256.0
[23:48:15] <wjp> um, when I said [0.0, 256.0) I meant 256 equally wide blocks in [0.0, 1.0)
[23:48:19] <wjp> (normalized)
[23:50:41] <wjp> brada: focusControl is a bit of a misleading name
[23:50:50] <brada> i thought thats what you suggested
[23:51:05] <brada> anyway just tell me what to call it :)
[23:51:16] <wjp> I think I said something like focusLock or such
[23:51:49] <wjp> but it needs a comment in the header anyway so we don't use it for other things
[23:51:55] <brada> ok
[23:53:39] <wjp> I think I'll take your word for the coordinate arithmetic
[23:53:58] <wjp> too late to go through that in detail :-)
[23:55:14] <wjp> hrm, I wonder why http://www.opengl.org/wiki/Common_Mistakes#Paletted_textures doesn't mention anything about that texel coordinate issue
[23:55:16] <Pepelka> Common Mistakes - OpenGL.org
[23:56:31] <wjp> but since I'm just talking to myself now I think it's probably time for bed :-)
[23:56:37] <wjp> good night
[23:56:59] <brada> night