#gemrb@irc.freenode.net logs for 9 Aug 2010 (GMT)

Archive Today Yesterday Tomorrow
GemRB homepage


[01:30:26] <-- nickdaly has left IRC (Quit: ERC Version 5.2 (IRC client for Emacs))
[01:30:27] <-- edheldil_ has left IRC (Ping timeout: 246 seconds)
[06:31:50] <pupnik> hey fuzzie - got me before schleep
[06:32:21] <pupnik> damn devkit VM is too big for my system here
[06:43:25] * fuzzie yawns.
[06:49:15] <-- |Cable| has left IRC (Remote host closed the connection)
[06:50:43] <pupnik> mornin fuzzie
[06:50:48] <pupnik> "This process will take a long time to collect very much at all, and will be of quite low strength, but is probably better than nothing. Easily adapted for use in your cell." http://homedistiller.org/notstill.htm#bucket
[06:58:02] <fuzzie> pupnik: you seen SiENcE's dingoo port?
[06:58:37] <fuzzie> http://www.youtube.com/watch?v=nFlQCSmX67s
[07:01:13] <pupnik> well that's running nice and fast, isn't it
[07:02:16] <fuzzie> it is impressive given it is downscaling :p
[07:36:33] <pupnik> i broke nokia's n900 so i'm kind of down on the whole ports thing
[07:37:16] <fuzzie> yes, access to hardware is annoying :(
[07:37:25] <fuzzie> i'd love to see an iPad port, but it's way too expensive to think about
[07:38:27] <pupnik> yep
[07:41:26] <fuzzie> maybe i can get the uni's multimedia people to let me play with one
[07:42:23] <fuzzie> they did an iPhone programming workshop, so they've got to have the equipment
[07:45:18] <fuzzie> ok, exam time
[07:48:28] <pupnik> good luck, sir!
[07:58:20] <-- raevol has left IRC (Quit: Leaving.)
[08:02:58] <Edheldil> pupnik: please, would you care to write about GemRB ports you work on on our wiki? Looks like there's quite a few of them now and it would be nice to have pointers to them
[08:06:31] <pupnik> i can summarize status for the wiki yes
[08:06:38] <pupnik> but is not my port
[08:09:01] <Edheldil> it would be great, thank you ;-)
[08:09:23] <Edheldil> do you have write rights to the wiki?
[08:10:43] <pupnik> hey i see fuzzie is on talk.maemo.org
[08:10:59] <pupnik> i do not
[08:11:33] <Edheldil> ask wjp, then :)
[08:11:39] <pupnik> would be a good way for me to put my fanboyism to use
[08:12:04] <Edheldil> yep! :)
[08:13:24] <pupnik> with SiENcE's dingoo downscaling, someone might be able to do a symbian (s60) port
[08:13:50] <pupnik> but that's just too extreme imo... better to work on android, iphone, meego, maemo
[08:16:42] <Edheldil> or winM ;-)
[08:18:11] <pupnik> Edheldil: it seems there ought to be a 'ports' page linked from the main wiki page http://gemrb.sourceforge.net/wiki/doku.php?id=start#gemrb
[08:18:44] <pupnik> ahh http://gemrb.sourceforge.net/wiki/doku.php?id=gallery#exotic_platforms
[08:18:53] <pupnik> yeah i see what you mean
[08:20:35] <pupnik> http://www.youtube.com/watch?v=EnQe5KlqGPg Pandora - GemRB - Baldurs Gate 1 - German Preview.avi
[08:22:36] <pupnik> one thing that is occasionally harsh is load times
[08:25:52] <pupnik> hoping to get a pandora someday
[08:31:35] <Edheldil> my idea is something that could be used by those exotic platforms users to find port for their platform. Soooo if you write something, it will be great :)
[08:32:18] <pupnik> well i have borrowed another n900 for a couple of weeks so i'll test gemrb now
[08:47:34] <Edheldil> :)
[11:12:31] * fuzzie flails
[11:12:35] <pupnik> at what
[11:12:59] <fuzzie> well, you'll do :)
[11:13:19] <fuzzie> i just did 3 hours of exam and everyone irl is on vacation, so consider yourself flailed at
[11:22:56] --> SiENcE has joined #GemRb
[11:26:43] <pupnik> what subject fuzzie ?
[11:26:48] <pupnik> hey nice youtube SiENcE - congrats
[11:28:30] <SiENcE> thanks :)
[11:28:38] <Edheldil> SiENcE: gz ;-)
[11:28:53] <SiENcE> i'm currently improving the vncserver
[11:29:16] <SiENcE> like having an http vnc stream people can watch pver the internet
[11:32:53] <pupnik> i think video options are better - things that generate mpeg4 streams
[11:33:37] <fuzzie> pupnik: Data Structures
[11:34:13] <pupnik> i loved my data structures and algorithms class :)
[11:34:20] <pupnik> but it was hard!
[11:34:26] <fuzzie> well, i didn't turn up to my data structures or algorithms classes
[11:34:39] <SiENcE> fuzzie, when do you release the new gemrb?
[11:35:01] <fuzzie> but hopefully the exam went ok anyway
[11:35:08] <fuzzie> SiENcE: lynx is in charge, but i thought it was the 21st
[11:36:09] <SiENcE> ok
[11:36:20] <SiENcE> i wait for an dingux release
[11:36:43] <SiENcE> later when Caanoo comes out, i will also do an Caanoo port
[11:37:16] <fuzzie> huh, we lost igi? :(
[11:37:33] <fuzzie> meh
[11:38:06] <fuzzie> by the time gemrb works, no-one will be left
[11:38:45] <Edheldil> igi left g3?
[11:39:08] <fuzzie> he hasn't logged in since the start of March
[11:39:56] <Edheldil> ah
[11:41:10] <Edheldil> SiENcE: I am already pestering pupnik - would you care to create a page(s) on our wiki dedicated to ports you are responsible for (or knowledgeable of)?
[11:41:31] <SiENcE> shure but later
[11:41:51] <SiENcE> first i want to build the current git
[11:41:59] <SiENcE> currently i'm working with 0.61
[11:42:07] <Edheldil> ok, thank you
[11:43:15] --> lynxlynxlynx has joined #GemRb
[11:43:16] --- ChanServ gives channel operator status to lynxlynxlynx
[11:43:52] <SiENcE> i will also write down what i need for bette porting ;)
[11:51:19] <Edheldil> was it a promise or a threat? ;-)
[12:05:30] <lynxlynxlynx> http://www.youtube.com/watch?v=XHosLhPEN3k
[12:42:14] <Edheldil> hehe
[13:15:25] <fuzzie> this instants stuff sure is hard to decode
[13:15:41] <fuzzie> the trouble with trying to work it out by experimentation is that everything is delayed
[13:16:07] <fuzzie> if you, say, display a string, give gold, display a string and then do something else, even linearly, it can be pretty random in which order those appear in the original message window
[13:17:19] <pupnik> hahah :)
[13:18:48] <lynxlynxlynx> you just need to retry it enough times to be able to do a statistical analysis on the results heh
[13:18:57] <fuzzie> well, it's also deterministic :(
[13:19:00] <fuzzie> just seems random
[13:20:40] <fuzzie> i think instants must be added to the queue and then executed as encountered, but gosh is that difficult to debug
[13:23:55] <fuzzie> and i guess i should implement it the other way, for now
[13:39:27] <fuzzie> oh hmm, maybe ActionOverride is a message
[13:39:33] <fuzzie> that would be ideal
[13:51:49] <fuzzie> Avenger's latest notes seem to have a lot missing
[13:57:15] <fuzzie> but i think that ActionOverride is a message and if so, this is not so hard to implement
[13:58:55] <fuzzie> and would fix most of our remaining scripting problems i hope
[14:19:36] <fuzzie> lynxlynxlynx: you think DisplayMessage would be a reasonable place to put the message buffer?
[14:21:38] <lynxlynxlynx> message as in what-i'm-writing-now sense? sure
[14:25:16] <fuzzie> sorry, yes, the "message window" one
[14:25:32] <fuzzie> i'm not sure we have any situations where we need to worry otherwise
[14:27:42] <fuzzie> right now we do DisplayConstantString for some other stuff (e.g. WhyCantEquip in DropDraggedItem), but i think those are all bugs
[15:00:14] * fuzzie peers
[15:00:57] <pupnik> I see there is a libSDL for symbian s60 :)
[15:01:11] <pupnik> GF doesn't seem to want to work on it though
[15:01:27] <fuzzie> somehow, don't think SDL is difficult bit :)
[15:02:30] <pupnik> yeah :(
[15:03:38] <fuzzie> anyone know anything about the search maps here?
[15:03:43] <fuzzie> i don't understand them, after looking
[15:05:21] <wjp> which aspect?
[15:05:30] <fuzzie> well, this 'side wall' stuff
[15:06:41] <fuzzie> gemrb maps 0 and 8 to it, and 8 (iesdp says 'impassable, non light blocking') does seem to appear in the middle of 0 ('impassable, light blocking')
[15:07:46] <fuzzie> really odd
[15:08:07] <wjp> would that be to avoid aliasing issus?
[15:08:10] <wjp> issues
[15:08:18] <wjp> or discretization, I should call it
[15:08:27] <fuzzie> i have no idea what that is :)
[15:08:53] <wjp> if you discretize a diagonal (on-screen) wall you get jagged rectangular edges
[15:09:13] <wjp> and you may not want some of those bits to block light
[15:10:26] <fuzzie> ah. not sure. it's really odd. maybe artifacts from their editing tools, honestly.
[15:12:06] <fuzzie> would help if dltcep displayed them right :)
[15:15:48] <fuzzie> (underlying problem is: i want to be able to do "can we see point X from point Y?" for visibility purposes)
[15:16:29] <fuzzie> (we do some weird sidewall thing right now where we allow the player to see through only some limited amount of sidewall, and i'm trying to work out whether i have to care about that at all, or if i can just block immediately)
[15:16:46] <wjp> can't one of our debugging hotkeys display this?
[15:16:59] <wjp> I'd check, but I don't have any IE stuff around here
[15:17:28] <fuzzie> no
[15:17:31] <fuzzie> but that is an excellent point
[15:17:35] <fuzzie> let me hack it to do so :)
[15:17:38] <wjp> :-)
[15:17:50] <lynxlynxlynx> we do have one to display an overlay, but i'm not sure of what
[15:18:04] <lynxlynxlynx> maybe the search map?
[15:18:07] <fuzzie> it just checks PATH_MAP_PASSABLE atm
[15:18:16] <wjp> ah, right, it's binary
[15:18:27] <fuzzie> but i'll add some other colours
[15:18:40] * wjp puts in a feature request for a legend :-)
[15:23:49] <fuzzie> that works rather delightfully well
[15:31:01] --> barra_library has joined #GemRb
[15:33:09] <fuzzie> gemrb is doing an admirable job of mangling this stuff at load time, though
[15:33:59] <fuzzie> and *nothing* in the start area has 'no see' set
[15:35:59] <fuzzie> in fact, gemrb's mangling doesn't ever use that value
[15:37:27] <fuzzie> unsurprisingly enough, Avenger disabled this in a commit labelled "updated pathfind.2da" :)
[15:38:26] <fuzzie> but, yes, ok, somehow i am not surprised that scripts can see through walls if pathfind.2da mangles everything to be visible
[15:39:21] <pupnik> wow
[15:45:04] <fuzzie> "meh", i would say
[15:46:57] <fuzzie> but, well, Map::ExploreMapChunk being such a pain certainly makes a lot more sense now
[15:48:59] <CIA-26> GemRB: 03fuzzie * r5c9f693c877b 10gemrb/gemrb/core/Map.cpp: add red (impassible) and blue (sidewall) tints to search map overlay
[15:49:00] <fuzzie> roll up, roll up, try your hand at spotting the nonexistant vision-blocking search tiles :)
[15:59:28] <-- SiENcE has left IRC (Ping timeout: 252 seconds)
[16:00:43] <wjp> :-)
[16:01:00] <fuzzie> http://fuzzie.org/nfs/gemrb/fix_isvisible.txt
[16:01:04] <fuzzie> ^- fixes eveything for me
[16:01:14] <fuzzie> but i'm tired and cranky and don't want to think about it any more
[16:03:56] <wjp> cool
[16:04:55] <fuzzie> *terrible* :P
[16:07:08] <fuzzie> should probably be a big warning 'this is broken and never works' in ExploreMapChunk too
[16:07:14] <fuzzie> i will put things into lynx's hands and go make dinner
[16:10:26] <lynxlynxlynx> why me? oO
[16:15:58] <-- barra_library has left IRC (Ping timeout: 276 seconds)
[16:20:33] --> SiENcE has joined #GemRb
[16:20:36] <fuzzie> well, you're more familiar with how it's all meant to work than i am
[16:20:50] <fuzzie> and i expect wjp will be IE-less for weeks, so i can't try and put another hat on his head
[16:23:33] <lynxlynxlynx> do you have an example of something that is broken now?
[16:23:51] <fuzzie> i don't think this breaks anything
[16:24:38] <fuzzie> the ExploreMapChunk warning comment is because the fog-of-war code checks NO_SEE and that seems to never ever be set, has been that way for 5 years
[16:25:04] <fuzzie> but, well, just that fact makes everything a bit of a mess :(
[16:25:28] <fuzzie> so it needs someone else to think about it (see how pathfind.2da never uses 4, which is NO_SEE)
[16:26:06] <-- SiENcE has left IRC (Quit: @all: cya)
[16:29:59] <fuzzie> and i guess you can verify the patch i posted by looking at some example areas with ctrl-7 and ctrl-8 and making sure that anything coloured blue shouldn't be visible for scripts/etc
[16:30:10] <fuzzie> but i'm not really quite sure where to look
[16:34:45] * wjp only has a tourist-hat on this week :-)
[16:37:42] <fuzzie> so go be a tourist!
[16:42:12] <lynxlynxlynx> looks ok
[16:42:49] <lynxlynxlynx> brynlaw is a nice example, since it has multistorey buildings
[16:43:12] <lynxlynxlynx> pretty clever way of making it work
[16:49:15] <fuzzie> ok.
[16:50:47] <CIA-26> GemRB: 03fuzzie * rd83932584e18 10gemrb/gemrb/core/Map.cpp: use PATH_MAP_SIDEWALL in Map::IsVisible, since PATH_MAP_NO_SEE is never present
[16:50:52] <fuzzie> that seems much better than the previous situation; would appreciate some testing
[16:51:29] <lynxlynxlynx> the four room monster prisoners in spellhold didn't see me, so that's new
[16:51:45] <fuzzie> :)
[16:51:49] <lynxlynxlynx> =i didn't get attacked
[16:52:29] <fuzzie> well, that's good, right?
[16:52:38] <fuzzie> assuming you don't mean that you opened the doors and it still didn't work
[16:52:48] <lynxlynxlynx> that's better than good
[16:53:05] <lynxlynxlynx> the last time i was there, one of them was out of the room too heh
[16:56:09] <fuzzie> but i think the coloured search map view is really useful too
[16:56:36] <fuzzie> will have to experiment with it, some other time
[17:02:41] <fuzzie> no longer sure of the hold thing, since i noticed actsleep.ids exists
[17:02:49] <fuzzie> so back to drawing board there
[17:09:28] --> ubik has joined #GemRb
[17:09:55] --- ubik is now known as Guest67985
[17:10:13] --- Guest67985 is now known as micru
[17:10:39] --> Maighstir_laptop has joined #GemRb
[17:25:19] <micru> hi there, i just watched a german review of gemrb over pandora but i'm not sure if what they say happens only on the console
[17:26:04] <lynxlynxlynx> what do they say?
[17:26:09] <micru> the test was with gemrb 060 and bg1
[17:26:15] <micru> one moment
[17:26:24] <micru> i'm writing it on the text pad
[17:26:53] <lynxlynxlynx> i'm sure some of the complaints are not relevant anymore
[17:27:48] <micru> 1.- white text over white background (bio and some other menus)
[17:27:49] <micru> 2.- text a bit displaced in the gui
[17:27:51] <micru> 3.- it is not possible to adjust the sond
[17:27:53] <micru> 4.- font size is too small to read on a small screen. They suggest that there should be an option that, when activated, would make the text bigger when talking to someone
[17:28:25] <micru> http://www.youtube.com/watch?v=EnQe5KlqGPg
[17:29:46] <micru> sorry, the test version is 0.61
[17:30:18] <lynxlynxlynx> there's a font mod for 4
[17:30:35] <micru> yep, i know, the bigger fonts mod
[17:30:36] <lynxlynxlynx> 3 - it is, but we probably don't save the ingame values
[17:30:54] <lynxlynxlynx> 2 - too vague and doesn't sound like a problem
[17:30:58] <lynxlynxlynx> 1 - huh
[17:31:21] <micru> that´s strange, i think it has more to do with the port
[17:31:38] <lynxlynxlynx> yeah, me too
[17:32:07] <micru> but other than that (and the complain about the sluggish sound in some sequences) they give a positive review :)
[17:32:26] <fuzzie> it is difficult to tell what's a problem with their port
[17:32:57] <fuzzie> i can never find out if anyone puts any thought into these openal ports for pandora/n900, for example
[17:33:05] <pupnik> ~/dev/gemrb$ git pull // fatal: Not a git repository (or any of the parent directories): .git
[17:33:23] <pupnik> maybe i should just clone
[17:33:29] <fuzzie> probably
[17:33:33] <fuzzie> that isn't a server error
[17:33:40] <fuzzie> it just means you don't have a git repository
[17:37:55] <lynxlynxlynx> in the current dir
[17:52:53] <fuzzie> the colour thing is, hm
[17:56:24] <fuzzie> i thought it might be an alignment issue, but the code looks ok at a glance
[17:59:29] <fuzzie> SDLVideoDriver::DrawRectSprite seems unlikely to do what the comment says
[18:14:02] <fuzzie> [CharAnimations]: Invalid or nonexistent avatar entry:0000
[18:14:02] <fuzzie> [Actor]: Missing animation for Altar
[18:14:05] <fuzzie> ^- this is kind of silly
[18:22:55] <lynxlynxlynx> is that the shadow altar?
[18:24:36] <fuzzie> it's in the area south of where Bodhi drops you under Spellhold
[18:24:52] <fuzzie> i rather assume that an anim id of '0' just means 'don't draw', though
[18:30:24] <lynxlynxlynx> oh, the goblin thingie
[18:31:21] <lynxlynxlynx> or the one on the way to the big mouth door?
[18:31:41] <fuzzie> i don't know, my binary has too many hcks for me to look at anything except output :|
[18:34:16] <lynxlynxlynx> anyway, your assumption sounds sane to me
[18:34:22] <fuzzie> trying to work out if i can rewrite this text area stuff to be less messy
[18:35:01] <fuzzie> at the moment the text logic, the dialog code and the actual control/drawing stuff is all here in one big mess
[18:35:30] <fuzzie> making it pretty inconvenient to do what Edheldil suggested and just make it point to an external buffer for the messages
[18:37:04] <fuzzie> and there are various other bugs here (text wrapping, drop capitals, etc)
[18:39:09] <fuzzie> it looks quite possible to fix it all, but it would be much easier split up a bit..
[18:41:27] <micru> fuzzie, your to-do list is growing huge :) i wish i could help more
[18:45:28] <fuzzie> it's nice to be fixing things again
[18:50:36] <fuzzie> magic missiles move *way* too fast
[18:51:09] <fuzzie> also Tiax just killed Jon on his own
[18:51:28] <fuzzie> some script issue, i guess
[18:52:21] <fuzzie> i guess he was probably waiting on the clone spawning
[18:58:59] <fuzzie> casting WIZARD_POWER_WORD_SILENCE (spwi612) fails, too: "[Spell]: Cannot retrieve spell header!!! required header: 3, maximum: 1"
[19:00:13] <fuzzie> oh, never mind, the combat code is at fault, as usual
[19:01:34] <lynxlynxlynx> for this?
[19:01:37] <fuzzie> yes
[19:04:07] <fuzzie> the action code is assuming that the combat code doesn't mess with actor vars when it's not in combat
[19:05:46] --> |Cable| has joined #GemRb
[19:13:32] --> SiENcE has joined #GemRb
[19:16:08] <fuzzie> ok, so who broke clicking on piles when you have a party?
[19:17:51] <fuzzie> oh, right
[19:17:56] <fuzzie> this stupid close container hack
[19:18:10] <fuzzie> i *wondered* what was crashing there
[19:24:31] <fuzzie> and, right, it doesn't give feedback on 0 damage
[19:25:59] <fuzzie> that is odd actually
[19:26:00] <fuzzie> meh
[19:26:45] * pupnik just stares :/
[19:27:28] <lynxlynxlynx> i found my bug :)
[19:27:35] <lynxlynxlynx> the original one
[19:27:51] <fuzzie> spell Source?
[19:29:30] <lynxlynxlynx> i don't know from where yet, but we give the same clab entries multiple times (could be the move of the LSR_KNOWN check) and RemoveSpell clears all memorisations too, so unmemorize+remove can only work once; hasspell will return true, since it is checking known spells
[19:29:45] <fuzzie> ah
[19:29:54] <fuzzie> all the kit/innate/etc stuff is just magic to me :)
[19:29:57] <lynxlynxlynx> i found the source one when debugging this
[19:30:28] <lynxlynxlynx> you master a lot of other areas though :)
[19:30:45] <fuzzie> well, yes, i'm happy we can split things like this :)
[19:32:56] <fuzzie> i'm just getting very confused here
[19:34:34] <lynxlynxlynx> what was with the damage thing?
[19:34:42] <fuzzie> i have no idea
[19:34:55] <fuzzie> no effects being applied to the party
[19:35:06] <fuzzie> i'll look at it with printfs in a bit
[19:39:56] <CIA-26> GemRB: 03lynxlupodian * rfa0bff8e1405 10gemrb/gemrb/GUIScripts/GUICommon.py: GUICommon: no need to manually unmemorize spells in RemoveClassAbilities
[19:41:13] <fuzzie> hm, Jon is working well this time around
[19:42:03] <fuzzie> but this container thing is terrible
[19:43:52] <lynxlynxlynx> remove it then
[19:44:03] <fuzzie> sorry
[19:44:08] <lynxlynxlynx> i'm serious
[19:44:21] <fuzzie> i mean, i want to know why this happens :)
[19:44:27] <fuzzie> since i guess it is some combat bug
[19:45:11] <fuzzie> something just killed my fighter in about 1 second, my guess is that something else is resisting that
[19:45:45] <lynxlynxlynx> any fireshields in play?
[19:45:48] <fuzzie> yes
[19:45:53] <lynxlynxlynx> oh yes
[19:46:08] <lynxlynxlynx> like i noted on the badspells page, the effect is horrible
[19:46:26] <fuzzie> ah
[19:46:46] <lynxlynxlynx> combined with our occasional let's-make-owner-the-target it is hard to play with keldorn and his original gear
[19:48:27] --> edheldil_ has joined #GemRb
[19:50:28] <fuzzie> // 0x10f AvatarRemoval
[19:50:31] <fuzzie> //bg2 calls this SummonDisable
[19:50:44] <fuzzie> ^- that looks fairly suspicious :P
[19:53:48] <fuzzie> wow, that fireshield is amazing :)
[19:55:17] <micru> which one? red or blue?
[19:56:11] <fuzzie> so, it has a fx_cast_spell_on_condition which checks LastDamage
[19:56:21] <fuzzie> and which doesn't self-destruct on death
[19:56:40] <lynxlynxlynx> that's only one problem
[19:56:55] <fuzzie> and LastDamage is never unset, so that'll last forever
[19:57:07] <lynxlynxlynx> it's so fast i doubt this can be accounted just to buggy timekeeping
[19:57:16] <fuzzie> (until the target dies, which in this case is my main char..)
[19:57:22] <lynxlynxlynx> oh, so every fx application
[19:57:32] <lynxlynxlynx> doh :)
[19:58:22] <fuzzie> i note also that fx_apply_effect_repeat says 'seconds' but actually does things every fx application
[19:59:30] <fuzzie> but my check in there hasn't fired in these two big spell battles i tried, so it's not the cause of any bug here
[20:01:54] <lynxlynxlynx> do you know off-the-top-of-your-head any spell that uses it?
[20:02:11] <fuzzie> i have no idea if it's used at all
[20:02:16] <fuzzie> let me try dltcep
[20:03:15] <lynxlynxlynx> oh
[20:03:21] <lynxlynxlynx> warding whip maybe?
[20:04:07] <fuzzie> summon insects, insect plague, creeping doom, earthquake?
[20:04:50] <fuzzie> looks like it
[20:06:14] <pupnik> i loved creeping doom as a kid
[20:07:01] <lynxlynxlynx> ouch yes, i think they do kill you pretty fast
[20:07:29] <lynxlynxlynx> earthquake was not so bad the last time i tried it though, the memory is from the bugs
[20:10:16] <fuzzie> for earthquake, it has resource 'shake1'
[20:12:02] <fuzzie> hmm, the original engine's effect queue really is buggy :)
[20:27:51] <pupnik> :)
[20:31:19] <pupnik> getting heavy sound stuttering in bg2 at 600mhz
[20:32:00] * pupnik curses pulseaudio again
[20:34:17] <fuzzie> yes, 'pulseaudio' and '600mhz' do not go together :p
[20:47:40] <pupnik> yep, 34% of cpu running gemrb
[20:47:46] <pupnik> 54% is gemrb
[20:48:32] <pupnik> the rest is X and other stuff
[20:49:13] <fuzzie> in normal use, gemrb spends most of its time in SDL, drawing
[20:49:33] <fuzzie> if you have up-to-date pulse and an up-to-date openal configured to use it, pulseaudio is not so bad on top of that
[20:50:45] <fuzzie> well, on x86
[20:54:44] <pupnik> yeah i'll configure this device to go bypass pulse and do some tests
[20:54:53] <pupnik> but that endangers/breaks speakers on n900s
[20:54:59] <pupnik> so headphones-only for now
[20:55:07] <pupnik> other people are overclocking from 600 to 800-900mhz
[20:55:28] <pupnik> damn nokia
[20:55:55] <fuzzie> well, openal on maemo is just silly
[20:55:59] <fuzzie> it's not officially supported, right?
[20:56:14] <fuzzie> it's just people building the heavy-fpu-usage for-desktop-use openal soft?
[20:56:47] <fuzzie> i think if all the porters are just going to do that, some lower-quality sdl audio plugin is a much better idea
[20:57:16] <pupnik> yes
[20:57:24] <pupnik> ok thanks for that tip
[20:58:00] <lynxlynxlynx> porters unite and write us that plugin :)
[20:58:12] <pupnik> this isn't doing much good excludefx = eaxreverb,reverb
[20:58:17] <-- micru has left IRC (Remote host closed the connection)
[20:58:21] <pupnik> ok
[20:58:28] <fuzzie> let me write one
[20:59:08] <lynxlynxlynx> why inhibit the unlikely contribution?
[20:59:22] <fuzzie> they can make it not suck :)
[20:59:43] <lynxlynxlynx> fair enough
[21:00:53] <pupnik> do you know how i can pass the playback device (not "default") to gemrb?
[21:01:04] <fuzzie> i don't think you can
[21:01:18] <pupnik> ok i need to force export AUDIODEV='hw:0'
[21:04:56] <pupnik> AL lib: alsa.c:486: Could not open playback device 'default': Connection refused
[21:08:40] <lynxlynxlynx> oh, what a waste of time, silly me!
[21:09:32] <pupnik> sorry man
[21:10:22] <lynxlynxlynx> maybe i was wrong, let's see :)
[21:19:11] <CIA-26> GemRB: 03lynxlupodian * ra5591b440449 10gemrb/gemrb/plugins/GUIScript/GUIScript.cpp: removed nonexistent bonus parameter from the doc of GemRB_SetMemorizableSpellsCount
[21:19:13] <CIA-26> GemRB: 03lynxlupodian * r505ccb5185ea 10gemrb/gemrb/GUIScripts/GUICommon.py:
[21:19:13] <CIA-26> GemRB: fxied AddClassAbilities to properly memorize already known spells
[21:19:13] <CIA-26> GemRB: fixes a bug and a regression from 232be72deb
[21:27:57] <lynxlynxlynx> ouch
[21:28:13] <lynxlynxlynx> another SaveGameIterator crash
[21:28:38] <lynxlynxlynx> something like load + delete game x2 + load
[21:29:15] <lynxlynxlynx> the last gemrb symbol is for RescanSaveGames
[21:34:47] <fuzzie> what after that?
[21:38:03] <lynxlynxlynx> libgemrb_core.so(+0x1c0299)[0xb8064299]
[21:38:10] <lynxlynxlynx> libgemrb_core.so(+0x1c03ec)[0xb80643ec]
[21:38:14] <lynxlynxlynx> and then glibc
[21:38:17] <fuzzie> you have line numbers?
[21:38:28] <lynxlynxlynx> no
[21:38:59] <lynxlynxlynx> the first glibc function was opendir though
[21:39:31] <fuzzie> you probably want to fix your CMAKE_BUILD_TYPE
[21:39:48] <lynxlynxlynx> Debug needs another define?
[21:39:53] <fuzzie> no
[21:40:12] <fuzzie> just needs to be RelWithDebInfo
[21:40:16] <lynxlynxlynx> opendir is only used from DirectoryIterator::Rewind()
[21:40:48] <lynxlynxlynx> how would that help?
[21:41:01] <fuzzie> 'Deb' = Debug
[21:41:40] <lynxlynxlynx> RelWithDebInfo uses -O2 by default, so it would screw gdb
[21:41:58] <lynxlynxlynx> Debug is just -g
[21:42:01] <fuzzie> no moreso than Release, though
[21:42:08] <fuzzie> which is presumably what you're using
[21:42:16] <lynxlynxlynx> release uses -O3 >>
[21:42:19] <lynxlynxlynx> i'm on debug
[21:42:34] <fuzzie> huh
[21:42:36] <fuzzie> weird, then
[21:43:29] <-- pupnik has left IRC (Quit: leaving)
[21:47:52] <fuzzie> which caller?
[21:47:55] <fuzzie> oh
[21:47:56] <fuzzie> nm
[21:48:06] <fuzzie> this is likely caused by Debug mode actually :P
[21:51:57] <fuzzie> the thing is: using DirectoryIterator here doesn't really work
[21:52:14] <fuzzie> because you can't have more than one at once, and this code calls PathJoin internally, which promptly makes another one
[21:53:15] <fuzzie> oh, hm, no
[21:53:21] <fuzzie> this should work, i guess
[21:53:40] <fuzzie> too tired for this, maybe :)
[21:55:45] <lynxlynxlynx> well, if it is a problem only in debug mode, it isn't a big deal
[21:55:54] <fuzzie> no, i was thinking along entirely the wrong lines
[21:56:09] <fuzzie> i don't think i can look at it without a traceback though, so i'll look at it another time
[21:56:33] <lynxlynxlynx> good night
[21:56:37] <fuzzie> night
[22:02:39] <-- lynxlynxlynx has left IRC (Remote host closed the connection)
[22:20:49] <-- SiENcE has left IRC (Quit: cya @all)
[22:28:19] <CIA-26> GemRB: 03fuzzie * rc5c583040282 10gemrb/ (5 files in 3 dirs): add the bare skeleton of a SDLAudio plugin
[22:28:32] <fuzzie> that took way too long
[22:28:57] <fuzzie> stupid SDL_mixer will only let you feed it data in the device format, so had to work out how SDL's audio conversion stuff worked
[22:39:22] <-- Maighstir_laptop has left #GemRb
[22:39:22] <fuzzie> wow, it works
[22:41:50] <CIA-26> GemRB: 03fuzzie * r33af1f2129f8 10gemrb/gemrb/plugins/SDLAudio/ (SDLAudio.cpp SDLAudio.h): add hacky music support to SDLAudio too
[22:41:56] <fuzzie> so, there you go, it plays music and UI sounds and doesn't crash in the first 10 seconds i tested it for :)
[22:42:04] <fuzzie> an ideal replacement for openal
[23:54:06] <-- edheldil_ has left IRC (Ping timeout: 246 seconds)