#gemrb@irc.freenode.net logs for 17 Aug 2010 (GMT)

Archive Today Yesterday Tomorrow
GemRB homepage


[04:12:40] --> raevol has joined #GemRb
[04:23:37] --> tomprince_loki has joined #GemRb
[04:34:02] <tomprince_loki> Hello again.
[06:02:05] <-- raevol has left IRC (Quit: Leaving.)
[07:11:17] <-- |Cable| has left IRC (Remote host closed the connection)
[07:38:56] --> Avenger has joined #GemRb
[07:39:12] --- ChanServ gives channel operator status to Avenger
[07:39:24] <Avenger> meh g3 is gone
[07:39:39] <Edheldil> again?
[07:39:48] <Edheldil> Hi, Avenger
[07:39:52] <Avenger> hi
[07:40:44] --> SiENcE has joined #GemRb
[07:41:27] <Edheldil> during RE, have you stumbled over structures that look like this? struc { int interLockedCounter; int ; int size; void* something ; LPCRITICALSECTION critSection; } ?
[07:42:00] <Edheldil> or st. like that
[07:44:44] <Avenger> yes, that's locking
[07:44:49] <fuzzie> morning tomprince, Avenger, Edheldil, SiENcE :)
[07:44:51] <Avenger> synchonizing
[07:45:09] <Avenger> hello fuzzie, now any iesdp source?
[07:45:21] <Avenger> i'm already addicted :P
[07:46:05] <Edheldil> well, it of course is, but do you have any idea what is it? It looks like some kind of smart pointer, either from C++ or from Lua
[07:46:05] <fuzzie> g3 will be back soon, i hear
[07:46:22] <Avenger> heh, you got good sources
[07:46:49] <Avenger> i just need to know what they know about spell flags
[07:47:05] <Edheldil> I have copy of iesdp from last year. Unfortunately it's still the current version, grrr
[07:47:21] <Edheldil> or almost current
[07:47:30] <SiENcE> hey morning
[07:47:36] <fuzzie> but "site:iesdp.gibberlings3.net" in google will give you the cached pages, you know
[07:47:36] <Edheldil> morning all
[07:48:43] <Avenger> haha, didn't know
[07:49:11] <Avenger> i mean, i always used it for search, but never thought about looking for the all
[07:49:49] <Edheldil> I use archive on some sites that require registration to read them :)
[07:50:10] <Edheldil> many times they are open to google bot to index them
[07:51:16] <fuzzie> but it's the forum which is important, so meh
[07:51:23] <fuzzie> i should quietly automate a bot, maybe
[07:51:51] <Edheldil> and do a local mirror of the forums?
[07:52:30] <fuzzie> mhm
[07:53:44] <Edheldil> before they disappear like DLTC
[07:54:26] <fuzzie> just wget would go in circles and get the host to take them down again for overloading the machine, though
[07:55:00] <Avenger> dltc (dragonlancetc.com) will reappear. it is already in friendly hands
[07:56:19] <SiENcE> fuzzie, how do i clone your working git to github?
[07:56:30] <fuzzie> huh, so it is
[07:56:54] <fuzzie> SiENcE: you have a github account setup?
[07:57:19] <SiENcE> yes
[08:00:02] <fuzzie> hm, i have no github public key on this machine
[08:00:44] <SiENcE> i found
[08:00:45] <SiENcE> git clone git@github.com:username/reponame.git
[08:02:46] <fuzzie> what do you want to do, exactly?
[08:03:49] <SiENcE> clone your gemrb repro from sf to my github
[08:04:28] <SiENcE> because every day i have to switch on 3 locations
[08:04:36] --> lynxlynxlynx has joined #GemRb
[08:04:36] --- ChanServ gives channel operator status to lynxlynxlynx
[08:07:22] <fuzzie> i think you have to clone to your machine first, and then push to github
[08:08:14] <fuzzie> or you could just use the github gui and fork my github gemrb repos, but maybe 'clone' is not what you mean
[08:09:08] <SiENcE> yes you are right fork i what i mean
[08:09:21] <SiENcE> i'm new to git...i only use svn
[08:09:33] <SiENcE> where is your gemrb github?
[08:09:55] <SiENcE> ah found
[08:10:28] <SiENcE> but i wanna use your official release repo.
[08:10:37] <fuzzie> then you have to do it manually
[08:11:05] <fuzzie> take a copy of the git repository, and 'git push git@github.com:username/reponame.git'
[08:11:50] <SiENcE> so yours is not in sync with the release one?
[08:12:18] <fuzzie> right now, it is identical to the release one, but it isn't always
[08:12:25] <Edheldil> you can track more than one repository, btw
[08:12:27] <fuzzie> not automated
[08:12:34] <SiENcE> mh
[08:12:56] <SiENcE> whats the best? i dont want a local copy and than copy to github
[08:12:58] <fuzzie> it's only up-to-date because there's a release coming up and so i'm probably going to do my work on github now
[08:13:09] <fuzzie> well, don't use github :)
[08:13:13] <SiENcE> arg
[08:13:20] <Edheldil> use our official SF repos
[08:13:48] <fuzzie> but i think in any case, if you want to also keep your local changes, there is nothing automated
[08:15:15] <Edheldil> nah... git pull sf ; git push --all github or st. like that
[08:15:32] <fuzzie> but that isn't automated :) you have to have your own machine do it
[08:16:17] <Edheldil> well, you have to have your own machine anyway, to do any work ;-). But I get your point
[08:16:27] <SiENcE> ok :)
[08:16:44] <fuzzie> well, i got used to pulling repositories everywhere because i use different machines all the time
[08:17:37] <Edheldil> true, that's why I installed my private CVS pserver many years ago :)
[08:17:58] <SiENcE> i have a private svn server
[08:18:31] <SiENcE> i have to read how i can manage this easily
[08:18:39] <fuzzie> yes, i store everything in a private git repos :)
[08:18:45] <fuzzie> branches, branches everywhere
[08:19:18] <fuzzie> since it's private i can be very untidy, just as long as i can keep track of the refs i am happy
[08:20:48] <SiENcE> first i fork yours fuzzie, and you can take a look what i added and maybe merge it to official
[08:21:10] <SiENcE> i used defines
[08:25:13] <Edheldil> fuzzie: I have migrated to git as well, and it was a relief
[08:25:39] <fuzzie> if only the scummvm people would get on with their migration :)
[08:31:45] <SiENcE> why scummvm?
[08:32:38] <fuzzie> because their svn repository has more than 52.000 commits and even that doesn't have all the history
[08:33:08] <fuzzie> and it gets enough updates that even just updating git-svn gets slow if you're not careful
[08:34:00] <SiENcE> true
[08:34:23] <fuzzie> all the other cvs/svn repositories i have to use are pretty small :)
[08:34:32] <fuzzie> although git-cvs is one-way, it is terrible :(
[08:36:45] <fuzzie> anyway, i guess everything left on my todo list is too invasive for the release
[08:49:50] <SiENcE> could you try to display the text on a conversation from start and not from the bottom?
[08:50:50] <SiENcE> if the messagewindow is to small to display the whole text, it scrolls down to the last sentences...you have to scroll up every time to read the start of the conversation
[08:51:10] <fuzzie> i want to rewrite that whole thing
[08:51:20] <SiENcE> uhm
[08:52:41] <Edheldil> is there still that annoying glitch that reply which just scrolled under your mouse cursor is not highlighted and active until you move your mouse a bit?
[08:53:19] <fuzzie> that too
[08:53:49] <fuzzie> it is a more general problem: nothing is updated until you move your mouse cursor, this also applies to in-game stuff
[08:53:49] <SiENcE> and scrolling up/down via cursor keys :)
[08:54:05] <fuzzie> i do not consider it annoying enough to care about
[08:54:23] <fuzzie> but this is why i ask for priorities :)
[08:55:05] <SiENcE> if i want todo a release for dingoo i need this small enhancements, but i can hack it into my release anyway :)
[08:55:49] <lynxlynxlynx> hacks will be needed for the time being
[08:56:21] <Edheldil> if good enough, somebody will gladly apply them back ;-)
[08:58:32] <Edheldil> fuzzie, how did you envision the plugin lists? Was the reason for them that a plugin would fail not during load, but in Init? Why to add another namespace (sdl vs. libSDLVideo)
[09:05:54] <fuzzie> they're not mine, remember :)
[09:06:06] <fuzzie> but tomprince supports static builds now (everything in one single binary)
[09:07:01] <fuzzie> and even with the plugin system, how do you say "try openal first, then sdlaudio, then nullsound"?
[09:07:50] <fuzzie> so it makes sense, it's just that they need to be *lists* and the defaults need to not be hard-coded into the binary (pick the first driver which isn't 'none', maybe?), plus the case-sensitivity thing and debug output and etc..
[09:09:34] <fuzzie> (debug output because it needs to not fail silently! no driver in the list works --> error out, i think sounds best - people can have 'none' as last choice if they want it)
[09:11:03] <fuzzie> but as i said, the whole thing annoys me because it's broken right now, so..
[09:11:11] <fuzzie> i'll fix it if no-one else gets around to it before release
[09:11:42] <Edheldil> ah, static builds ....
[09:13:12] <Edheldil> with plugins, you could achieve that with DelayPlugin or "PreferPlugin", but I will keep static build in mind
[09:14:06] <fuzzie> DelayPlugin only allowed one delay
[09:14:21] <fuzzie> i guess you could add delay values etc, but then yes, no static builds
[09:14:39] <fuzzie> i think he just hacked the whole thing in for static builds, it isn't written too well :)
[09:17:03] <fuzzie> so if you want to rewrite it and make it a new system, i think that's no problem
[09:17:33] <fuzzie> just would be nice if it would work ok with an empty config, so avoiding 'none' plugins would be good
[09:17:49] <fuzzie> and now i am going off to study push-down automata..
[09:18:00] <Edheldil> I will have to get my head around the Holder<> thingy. I never used templates above stl::Vecor<int>
[09:18:22] <fuzzie> it is just a smart reference-counted pointer
[09:18:25] <Edheldil> automata are fun :)
[09:18:52] <Edheldil> st. I think is that struct mentioned above :)
[13:08:33] <-- tomprince_loki has left IRC (Read error: Connection reset by peer)
[13:46:06] <fuzzie> hi
[13:56:24] <lynxlynxlynx> i wonder if sarevok's cheese strike is hardcoded
[13:57:19] <fuzzie> in bg1?
[13:57:29] <lynxlynxlynx> bg2
[13:58:51] <fuzzie> ah. i don't know what is, then, so not too helpful :)
[13:59:13] <lynxlynxlynx> something like 5% chance of 300 damage
[13:59:19] <lynxlynxlynx> "deathbringer assault"
[13:59:40] <fuzzie> oh, tha
[13:59:40] <fuzzie> t
[14:00:02] <lynxlynxlynx> enemies can still attack me when i go invisible
[14:00:03] <fuzzie> i wondered what that was :)
[14:00:06] <lynxlynxlynx> we don't clear the targets?
[14:00:22] <fuzzie> um, it's complicated
[14:00:37] <lynxlynxlynx> they're drow and don't have any seeinvisible effect
[14:01:04] <fuzzie> what's their script?
[14:01:21] <lynxlynxlynx> drofod01
[14:01:26] <fuzzie> but yes, we don't clear any target stuff
[14:01:37] <lynxlynxlynx> err, no that's the dialog
[14:01:46] <lynxlynxlynx> they only have shout and wtasight
[14:02:31] <lynxlynxlynx> we maybe don't need to clear, but it should definitly abort on the next attack
[14:02:42] <fuzzie> yeah, but our triggers are broken
[14:03:22] <fuzzie> i'm not sure if we should definitely abort :) but easy to check
[14:03:44] <lynxlynxlynx> we currently check only the death state
[14:04:36] <fuzzie> where are you looking?
[14:05:04] <lynxlynxlynx> PerformAttack
[14:05:20] <fuzzie> that should never get called with dead/invisible/etc actors, in theory
[14:08:03] <fuzzie> but, right, there's no invisible check
[14:08:17] <lynxlynxlynx> it was the first place i checked
[14:09:08] <fuzzie> i'm just not sure where exactly it should go
[14:10:16] <fuzzie> all the printMessage stuff in PerformAttack should be followed by an abort(), really :|
[14:15:06] <lynxlynxlynx> we can change that after the release
[14:15:20] <lynxlynxlynx> should give more incentive to fix it then
[14:15:44] <fuzzie> but i guess put the invisibile check wherever you want
[14:16:16] <fuzzie> it's a bit complicated to do correctly, so another post-release thing to do it right
[14:21:32] <lynxlynxlynx> yeah
[14:22:22] <fuzzie> > Phaere saved against effect: Damage
[14:22:29] <fuzzie> ^- another game-breaking resist..
[14:25:10] <fuzzie> oh, maybe not
[14:27:33] <fuzzie> she's scripted to die. oh dear.
[14:29:05] <fuzzie> and wow, Adalon is completely misrendered
[14:30:16] <fuzzie> also, broken
[14:32:15] <fuzzie> TextScreen is broken at 800x600, too
[14:32:30] <fuzzie> (rendered off-center)
[14:34:36] <lynxlynxlynx> all the dragons have some animation problems
[14:34:55] <fuzzie> and the broken thief skill stuff is known, right?
[14:35:07] <fuzzie> we don't apply it dynamically, or something
[14:37:21] <fuzzie> and i mean, her script is broken, but maybe my fault
[14:42:37] <lynxlynxlynx> what thief skill stuff?
[14:42:59] <lynxlynxlynx> on the last replay, i don't remember problems with adalon
[14:43:14] <fuzzie> the thief skills are wrong
[14:43:18] <fuzzie> for everyone in my party
[14:43:38] <fuzzie> off by up to 25 or so, it seems
[14:44:04] <lynxlynxlynx> smaller than expected?
[14:44:10] <fuzzie> it looks like it
[14:44:21] <fuzzie> so this is the known issue?
[14:44:25] <lynxlynxlynx> yeah
[14:44:32] <fuzzie> ok
[14:50:51] <fuzzie> i think there's something i
[14:50:56] <fuzzie> 'm missing in the dialog thing
[14:55:40] <SiENcE> fuzzie, http://github.com/SiENcE/gemrb/commit/1220d8f89ab4f93bfd40d9fd50f3fc0adb9b7477
[14:55:54] <SiENcE> i think you can safely add this commit
[14:56:06] <SiENcE> if you want
[14:56:46] <SiENcE> to get an fully autogen without manual modifications it need more work
[14:57:02] <fuzzie> ah, i think the Adalon bug is that our ReallyForceSpell is AF_BLOCKING
[14:58:24] <fuzzie> will bbiab
[14:58:54] --> tomprince_loki has joined #GemRb
[14:59:45] <lynxlynxlynx> SiENcE: it's practically identical to the current one
[15:00:08] <lynxlynxlynx> you could just make that one take another paramter for the extra flags
[15:01:30] <lynxlynxlynx> why don't we have SDL_HWSURFACE anywhere? :|
[15:01:34] <SiENcE> so no extra script?
[15:02:09] <SiENcE> hwsurfaces have some problems depending on the graphic driver
[15:02:28] <SiENcE> maybe thats why you choose the safer method
[15:02:39] <fuzzie> well, it ends up generally being slower, i think
[15:02:45] <Edheldil> config option it should be
[15:03:16] <SiENcE> fuzzie, ? why it should be slower? it should be faster in hardware
[15:03:18] <fuzzie> someone has to review the locks and surface code in SDLVideo if they want to turn it on
[15:03:35] <fuzzie> SiENcE: sure, but in reality everyone uses X11 or windib or quartz
[15:03:43] <fuzzie> so it ends up being not hardware at all
[15:03:48] <SiENcE> ah ok
[15:03:58] <fuzzie> just some different software surface
[15:05:04] <SiENcE> so..i modify my fork for my needs to build gemrb for dingux and you choose what to add to official, right?
[15:05:25] <CIA-26> GemRB: 03lynxlupodian * r53531fe40334 10gemrb/gemrb/core/Scriptable/Actor.cpp: added invisibility check hack to PerformAttack
[15:08:34] <fuzzie> i have no problem applying the #ifdef stuff, but would prefer you got an ok from lynx or someone too, especially when ignoring things like mouse warping
[15:08:40] <fuzzie> but cooking dinner right now
[15:09:28] --> Maighstir has joined #GemRb
[15:11:46] <fuzzie> and a much *better* solution would just be to provide a modded gui like http://www.de-doc.com/Projects/BGEdit/BGPS1_19.jpg ;p
[15:14:24] <SiENcE> fuzzie, you are right but therefore i need to get into and it needs a lot more changes
[15:14:44] <fuzzie> yes, i am ok with being pragmatic
[15:15:06] <fuzzie> (which is to say: take whatever works right now)
[15:15:21] <SiENcE> my second commit is not so nice
[15:15:29] <SiENcE> just a hack for me
[15:15:30] <SiENcE> http://github.com/SiENcE/gemrb/commit/13857881cd689a569ee8f024b30bbc900ee79990
[15:15:42] <SiENcE> so dont merge this
[15:16:06] <fuzzie> why not just add --disable-werror to your autogen?
[15:16:31] <SiENcE> i'm doing ...this is old stuff
[15:16:32] <fuzzie> or just get rid of the stupid -Wcast-align
[15:16:35] <Edheldil> fuzzie: what engine did it use?
[15:16:53] <fuzzie> Edheldil: the bg port?
[15:17:17] <SiENcE> i think -Wcast-align has nothing todo with my errors
[15:17:22] <fuzzie> i'm pretty sure the -Wcast-align shouldn't be there, it shows up bugs but it will never work on MIPS
[15:17:52] <fuzzie> because there's absolutely no way to convince gcc that casts are correct alignment-wise
[15:18:03] <fuzzie> every new release they add more and more strict checks
[15:18:08] <SiENcE> i think my mips crosstoolchain ignores this
[15:18:20] <fuzzie> so -Wcast-align is basically just a "produce thousands of stupid warnings" option
[15:18:34] <SiENcE> how do i disable this?
[15:18:52] <fuzzie> but any *other* warnings, i would like to hear about
[15:19:40] <SiENcE> k
[15:19:45] <fuzzie> you can just remove the -Wcast-align from the configure.in flags
[15:19:55] <SiENcE> ok
[15:20:47] <SiENcE> i also ignore the python check because in my cross toolchain it don't find my python
[15:21:13] <SiENcE> thats really bad...i have to recompile python and find the correct way to install
[15:21:49] <fuzzie> our python check is just broken, i think
[15:21:59] <fuzzie> it tries *running* the target python or something
[15:22:05] <fuzzie> this is why we don't support autotools any more :P
[15:22:35] <SiENcE> for this release i need it
[15:22:42] <SiENcE> later i try to get into cmake
[15:22:48] <SiENcE> i promise :)
[15:23:03] <fuzzie> well, this is why we don't remove it :) but it will just be more and more broken, i guess
[15:24:15] <SiENcE> fuzzie i aggree with you to keep buildsystem clean
[15:24:29] <SiENcE> i also maintain our big Iris2 project
[15:24:39] <SiENcE> and we have the same problems
[15:27:16] <SiENcE> currently we use premake
[15:27:26] <SiENcE> and on windows visual studio express
[15:27:57] <tomprince_loki> I noticed when poking around at the include.py code, that we import __main__ into every main GUIScript, which has GUIDefines, GUIClasses, and GemRB imported into it (the PyDict_Merge line ~1920 in GUIScript.cpp)
[15:35:59] <Edheldil> Hi, tom
[15:38:16] <tomprince_loki> Hello.
[15:38:18] <fuzzie> tomprince_loki: 1920?
[15:38:30] <fuzzie> i see one on 9710
[15:39:06] <tomprince_loki> ..... yes. Apprently I can't read or something.
[15:39:07] <fuzzie> but i thought that merged everything into pMainDic rather than vice-versa
[15:39:33] <fuzzie> so, eek
[15:40:48] <fuzzie> october 2003, heh :)
[15:40:55] <tomprince_loki> It is vice-versea, so I would like to remove that eventually, but there is code that breaks without it.
[15:40:56] <Avenger> hmm fuzzie, just some info: the IE code also has these 'HandleContainer', 'HandleDoor' etc functions where it constructs an action. I just found the one handling containers
[15:41:14] <fuzzie> Avenger: it just constructs UseContainer with the global id in int0?
[15:41:27] <fuzzie> tomprince_loki: it isn't simple to fix the code?
[15:41:57] <fuzzie> i have not gotten around to refactoring any of the GUIScript stuff - obviously - i have a rather long list of other important refactoring first
[15:41:59] <Avenger> it constructs action types, it simply copies the whole object of the container into the action's object
[15:42:18] <fuzzie> Avenger: ok. no need to copy this.
[15:42:58] <Avenger> well, the whole object transfer is overkill, since containers don't have those targeting id fields anyway :)
[15:43:50] <fuzzie> well, they just copy whole objects around, which makes sense
[15:43:58] <Avenger> in memory, each action has 3 complete objects in it, each of them with the size of 0x1c
[15:44:01] <fuzzie> but we can be less lazy
[15:44:33] <tomprince_loki> Yes. But it has the potential to break code, and I don't have any time to test right now, and won't really for another month or so.
[15:45:08] <fuzzie> tomprince_loki: ok. if you put it in a branch, i'd be happy to test and merge after the next release (Friday).
[15:45:25] <Avenger> they also have lots of action constructors followed by an action destructor, i have no idea what are those. So many unused local vars?
[15:46:19] <Avenger> they keep lots of them on the stack too :)
[15:46:34] <fuzzie> tomprince_loki: your 'invert' branch is somewhat out of date (i refactored the code completely) and i merged 'banter', modified a bit
[15:48:23] <Edheldil> Avenger: perhaps it's a leftover from compiled out code
[15:50:33] <Avenger> i like the structure definitions, now that i defined the action struct, the stack variable counts shrink significantly.
[15:56:01] <fuzzie> hm, still no G3
[15:57:38] <Avenger> is this a new isp, or an old isp hung it up?
[15:57:48] <fuzzie> old isp i think
[16:01:34] <fuzzie> huh, i hadn't seen http://www.rockpapershotgun.com/2010/08/16/not-forgotten-bioware-on-baldurs-gate/
[16:10:18] <Avenger> nice they still remember their best game :P
[16:11:44] <fuzzie> but yet another confirmation on the "the lawyers will doom any re-release" front :(
[16:13:58] <Avenger> hmm, that is hasbro lawyers, i guess?
[16:14:50] <fuzzie> and whoever owns Interplay now, i expect
[16:16:45] <Avenger> well, to interplay i would offer them to release the games on linux and whatever they wish :)
[16:17:21] <fuzzie> and their lawyers would say that they're not quite sure who owns the rights, but give them a heap of money and they'll look into it
[16:17:45] <fuzzie> it is annoying
[16:17:47] <Avenger> if they don't know who own the rights, then meh :P
[16:18:05] <Avenger> i would say, come back when you know
[16:18:53] <fuzzie> GameTap used to have PS:T/BG/etc
[16:19:11] <fuzzie> but then some company got bought by another company and the new company lawyers realised that no-one has any clue
[16:19:16] <Avenger> well gametap can sell the game on linux from now :)
[16:19:28] <fuzzie> of course the copies on gametap weren't very useful anyway, windows-only encrypted-image stuff
[16:20:25] <fuzzie> but there was a PS:T DVD re-release in 2009, so clearly *someone* knows the situation and isn't saying
[16:20:49] <Avenger> since we don't supply any original content, code or art, it is basically the d&d license anyway. Which we can avoid somewhat by making the engine generic enough
[16:20:51] <fuzzie> and that makes me suspicious it is Hasbro at fault
[16:21:11] <fuzzie> well, the d&d licensing is all about the game data anyway, really
[16:21:34] <fuzzie> they don't want you having Beholders in your game, or copying all their tables
[16:22:04] <lynxlynxlynx> luckily the originals don't match to good at all
[16:22:10] <fuzzie> and, well, i hope we don't ship the tables with gemrb anyway
[16:22:34] <Avenger> well, the 'beholder' text is in the original tlk table which we don't supply, and the beholder bitmaps are also in the original avatar graphics, which we don't supply
[16:24:03] <Avenger> there may be some case where the original engine contained a d&d table hardcoded
[16:24:13] <Avenger> that would be the problematic part :)
[16:24:41] <fuzzie> as lynx says, i think the hardcoded tables aren't really d&d anyway
[16:26:56] <SiENcE> anyone knows how to get the cross_compiling=yes in configure=
[16:27:01] <SiENcE> ?
[16:27:12] <SiENcE> all my exports dont work
[16:29:13] <fuzzie> it doesn't work if you pass the correct --host/--target/--build?
[16:30:18] <Avenger> whoa, at one point they stuff a complete action on the stack
[16:30:31] <Avenger> i mean, passing action by value
[16:30:49] <SiENcE> btw. http://www.golem.de/1008/77243.html
[16:31:07] <SiENcE> fuzzie, doesn't work :(
[16:31:24] <fuzzie> SiENcE: all the cross-compiling code comes from your local install of autotools, so we can't do anything about it
[16:31:39] <fuzzie> my local install of autotools will set cross_compiling as long as you set all of those options, but maybe yours is different
[16:33:08] <SiENcE> how can i override my local autotools environments?
[16:33:23] <fuzzie> by modifying the configure script
[16:33:51] <fuzzie> it is an internal variable, so they don't want you changing it
[16:34:12] <SiENcE> so i cant override it? ok
[16:34:17] <fuzzie> at least, it is here :)
[16:34:39] <fuzzie> i was in Rotterdam today, they had a stack of books on autotools
[16:34:46] <fuzzie> at some bookstore in the town centre
[16:35:05] <fuzzie> it was very strange, i wonder if they have a lot of demand for it :)
[16:35:52] <SiENcE> :)
[16:38:18] <fuzzie> but i have no spare 50eur for one, sorry :P
[16:38:29] <SiENcE> ;)
[16:38:34] <SiENcE> ok cya later
[16:38:57] <-- SiENcE has left IRC (Quit: @all: cya)
[17:02:59] <fuzzie> i was wondering why i hadn't noticed white circles when talking to people before
[17:03:18] <fuzzie> but i guess they only show for dialog when a feet circle is rendered at all
[17:21:31] <Avenger> wildmages get some spells automatically?
[17:26:51] <fuzzie> don't ask me :)
[17:27:09] <Avenger> well i hoped lynx knew it
[17:33:41] <lynxlynxlynx> you mean in the spellbook?
[17:34:18] <lynxlynxlynx> probably yes, i don't remember any scrolls with their few specialised spells
[17:35:23] <lynxlynxlynx> my guess is they did it this way so other spellcasters couldn't learn the spells, besides the description of the kit makes it sound more like wild magic is an innate, so the limit sounds fine from there too
[17:43:34] <lynxlynxlynx> Avenger: do you know what fx_play_visual_effect should do if the second param is 2?
[17:43:46] <lynxlynxlynx> dltcep and we only knows about 0 and 1
[17:43:59] <lynxlynxlynx> ugh
[17:44:10] <Avenger> i thought it is implemented
[17:44:37] <lynxlynxlynx> we don't differentiate between 1 and 2
[17:44:46] <fuzzie> should w?
[17:45:17] <lynxlynxlynx> i don't know, that's why i ask
[17:45:34] <Avenger> there is difference yes, and we didn't implement it
[17:45:57] <lynxlynxlynx> it's definitely a candidate for another PlayOnce though, the dragon's breath spell's animation is sticky
[17:46:32] <Avenger> param 0: creates vvc in area, at coordinates
[17:46:49] <fuzzie> surely sticky animation effects shouldn't be PlayOnce?
[17:47:09] <lynxlynxlynx> it has this param as 2, so maybe it shouldn't be sticky
[17:47:17] <fuzzie> ah
[17:47:23] <lynxlynxlynx> it definitely should play only once
[17:47:24] <fuzzie> i guess Avenger will explain
[17:47:38] <fuzzie> the effect is permanent?
[17:47:45] <Avenger> param 0: creates vvc in area, at the target actor's coordinates, param 1: sticky, param 2: creates vvc in effect's target coordinates
[17:47:54] <lynxlynxlynx> now i'm stuck with this appearing dragon head in my save :s
[17:48:08] <fuzzie> oh, well, that is easy :)
[17:48:37] <fuzzie> Avenger: and 2 is not sticky, i assum?
[17:49:04] <Avenger> yes, only 1 is tied to actor, 0,2 are area based
[17:51:27] <lynxlynxlynx> is one of you two fixing it or should i?
[17:51:58] <fuzzie> you seem best placed to fix it :)
[17:54:16] <fuzzie> i am playing Watcher's Keep in the original
[17:55:49] <lynxlynxlynx> that one's fun
[17:56:16] <lynxlynxlynx> if you do it right at the start of tob, you come out a true godling
[17:58:33] <fuzzie> also useful, lots of varied things to experiment with
[17:58:49] <CIA-26> GemRB: 03lynxlupodian * rdd4f4a1a96c7 10gemrb/gemrb/plugins/FXOpcodes/FXOpcodes.cpp:
[17:58:49] <CIA-26> GemRB: fx_play_visual_effect: fixed the handling of 0/2 Parameter2
[17:58:49] <CIA-26> GemRB: fixes eternal Dragon's breath spell animation
[18:05:25] <Avenger> hmm fuzzie, if you ever use ida on the IE, first define all the actions/objects/triggers structures, then define them over stack variables, the whole stuff is so much more readable.
[18:06:08] <fuzzie> or i can just leave the work to people like you :)
[18:06:25] <fuzzie> i don't think i've ever seen code with so many stack structures as you talk about
[18:06:49] <Avenger> the whole complexity in the code is coming from constructing messages, actions and triggers. Otherwise it is simple
[18:07:07] <Avenger> all the huge chunks are those
[18:11:18] <fuzzie> what are you looking at?
[18:11:29] <fuzzie> i mean, in particular
[18:13:58] <lynxlynxlynx> bah, my quicksave got corrupted
[18:14:10] <lynxlynxlynx> now crashes on load
[18:14:33] <lynxlynxlynx> invalid seeks aplenty
[18:16:31] <fuzzie> :(
[18:17:01] <fuzzie> and you have asserts enabled?
[18:21:51] <fuzzie> hmm
[18:22:45] <fuzzie> got saved locations, maybe?
[18:23:27] <fuzzie> looks like they'd corrupt a game
[18:23:33] <fuzzie> (should be easy to check with ielister)
[18:26:51] <lynxlynxlynx> maybe i do
[18:27:17] <lynxlynxlynx> it was before i was supposed get them afresh
[18:27:42] <fuzzie> the '* 20' for the saved location count and the pocket plane count is bad
[18:27:53] <lynxlynxlynx> if it makes any difference, the last succesful load was of a bcs script, so it looks like it died on reading a creature
[18:27:56] <fuzzie> they're only 12 bytes each, so you go off the end of the file
[18:28:17] <lynxlynxlynx> i remember that too
[18:29:40] <fuzzie> if that's the bug, you could just hexedit the save, if it's important
[18:30:00] <lynxlynxlynx> nah, i'll replay, this lich is kinda fun
[18:31:35] <lynxlynxlynx> is it fine if i change the 20 to sizeof(GAMLocationEntry) instead of hardcoding 12?
[18:31:51] <fuzzie> um
[18:32:03] <fuzzie> no
[18:32:06] <fuzzie> i think
[18:32:17] <lynxlynxlynx> something in that manner?
[18:33:03] <fuzzie> i think hardcoding 12 is the easiest way
[18:33:38] <lynxlynxlynx> it is
[18:33:48] <fuzzie> there are two issues: structs are not necessarily the same size on each platform, and GAMLocationEntry is bigger than 12 anyway
[18:34:59] <fuzzie> (ieResRef is 8 + 1 for null-terminator, plus padding for alignment)
[18:36:07] <fuzzie> you could do '* (8 + 2*sizeof(ieWord))' or something
[18:36:17] <fuzzie> but i think that's as nice as it gets
[18:36:45] <lynxlynxlynx> nah, i'll leave it as 12, not much worse
[18:40:12] <lynxlynxlynx> oh and i found out i wasn't wrong about the white circles
[18:40:22] <lynxlynxlynx> it's the centering that isn't always following
[18:40:39] <fuzzie> oh?
[18:40:52] <fuzzie> i mean, i don't understand :)
[18:41:14] <lynxlynxlynx> i thought the circle updates could lag, but it's the centering that doesn't always keep track of the talker
[18:41:16] <fuzzie> oh, you mean, in gemrb, the centering doesn't always happen?
[18:41:19] <fuzzie> right
[18:41:32] <fuzzie> if i didn't say, all that code needs rewriting, so i wouldn't worry too much
[18:42:31] <lynxlynxlynx> i don't
[18:42:43] <lynxlynxlynx> i was just happy it wasn't my fault :)
[18:43:52] <fuzzie> Avenger: if you want yet another thing to look at, the magic missiles stuff would be nice, but difficult
[18:44:09] <Avenger> why? our magic missiles are not good?
[18:44:22] <fuzzie> nope
[18:44:54] <fuzzie> i can fix it by just hacking the projectile code a bit, probably
[18:45:08] <fuzzie> but they're oddly complicated
[18:45:34] <fuzzie> and not as irrelevant as i thought
[18:45:59] <fuzzie> (they're all too fast, and they're not meant to hit at the same time, and they actually slow down quite a bit in the middle depending on how many there are, so that the damage is spread)
[18:47:04] <fuzzie> and, well, i am lazy, and i guess we'll never reproduce this properly by guesswork
[18:51:39] <CIA-26> GemRB: 03lynxlupodian * r85e23693632e 10gemrb/gemrb/plugins/GAMImporter/GAMImporter.cpp: GAMImporter::GetStoredFileSize: fixed bad header size, thanks fuzzie
[18:54:17] <-- Maighstir has left IRC (Quit: Maighstir)
[18:55:37] --> Micru has joined #GemRb
[18:56:08] <Micru> hi
[18:56:10] <fuzzie> hi
[18:56:36] <Micru> i'm still working on it, it is *almost* done :)
[18:57:10] <Micru> how is the release preparation going?
[18:57:14] <Micru> todo list empty?
[19:01:45] <fuzzie> i think there is very little on there
[19:02:49] <lynxlynxlynx> mine is empty
[19:03:26] <Micru> btw, what happens with g3?
[19:03:53] <Micru> financial crisis?
[19:05:03] <lynxlynxlynx> periodic crisis
[19:06:25] <fuzzie> the perils of cheap hosting providers
[19:07:30] <Micru> ah, if that is the only thing happening, then i think they will solve it soon... hopefully....
[19:07:52] <fuzzie> it sounds like the cheap host got fed up with the database load
[19:18:28] <-- tomprince_loki has left IRC (Read error: Connection reset by peer)
[19:36:04] <lynxlynxlynx> bah
[19:36:12] <lynxlynxlynx> now another quicksave just segfaults
[19:36:48] <lynxlynxlynx> nicely though
[19:37:40] <fuzzie> did you test the locations stuff?
[19:37:50] <lynxlynxlynx> fx->CasterID is 0 so GetCaster returned null
[19:40:39] <fuzzie> this is just a bug where nothing checks for that?
[19:41:03] <lynxlynxlynx> yeah
[19:42:13] <fuzzie> hm, our fx_set_charmed_state checks caster?
[19:42:49] <lynxlynxlynx> nope
[19:43:03] <lynxlynxlynx> that's where it crashes
[19:43:36] <fuzzie> i mean, should it?
[19:44:17] <lynxlynxlynx> this is caused by a corpse - more stuff to remove on death, i guess
[19:44:24] --> Maighstir has joined #GemRb
[19:44:39] <fuzzie> i'm just wondering how this is meant to work
[19:44:52] <fuzzie> references to actors can't persist after a save/load
[19:45:04] <fuzzie> the original engine just doesn't let you save when something is charmed?
[19:45:10] <lynxlynxlynx> yep
[19:45:41] <fuzzie> doesn't look like it gets removed on death
[19:46:37] <lynxlynxlynx> i'll fallback on Owner in the effect
[19:46:53] <fuzzie> Avenger: you have this reversed?
[19:47:10] <Avenger> what?
[19:47:26] <Avenger> disabled saving when something is charmed?
[19:48:00] <fuzzie> the charm opcode
[19:48:06] <fuzzie> (5)
[19:50:54] <Avenger> i saw it already, but it is complicated
[19:51:35] <Avenger> there are copies of an object's 'identity' (idscopy1/2)
[19:51:50] <Avenger> those are also involved
[19:52:06] <Avenger> i still didn't figure out how
[19:53:45] <Avenger> ahh, one thing: the calculated EA is stored in the effect, so theoretically the charm effect survives a save/load
[19:54:38] <Avenger> i think you can even force a save
[19:54:38] <fuzzie> stored where? :)
[19:54:52] <Avenger> try this:
[19:55:00] <fuzzie> i mean, it seems silly for lynx to add some hack if you know the correct way
[19:55:00] <Avenger> create a simple dummy cre
[19:55:06] <Avenger> which has no scripts set
[19:55:20] <fuzzie> i think you can *always* force a save
[19:55:27] <Avenger> then go to some area where you can save by leaving the area
[19:55:33] <fuzzie> engine sometimes crashes when you come back though
[19:55:37] <Avenger> then charm the dummy, and leave
[19:55:44] <Avenger> which engine?
[19:55:51] <fuzzie> bg2
[19:55:56] <Avenger> really?
[19:56:01] <Avenger> wow
[19:56:03] <Avenger> :)
[19:56:14] <fuzzie> well, it doesn't expect you to hack saves by script :P
[19:56:48] <fuzzie> and it can disable use of party-required transitions to avoid the autosave
[19:57:11] <fuzzie> but i forget how to make it crash, i'm sure charm doesn't do it
[20:01:57] <Avenger> btw, if i recall correctly, there is hostile and friendly charm
[20:02:22] <Avenger> if you use param2=0, the target becomes neutral after charm is gone
[20:03:01] <lynxlynxlynx> i got as far as sendai's inner sanctum
[20:03:23] <lynxlynxlynx> something breaks there
[20:03:54] <fuzzie> in ToB?
[20:04:11] <fuzzie> i wonder if we should branch for the release now
[20:05:10] <lynxlynxlynx> http://pastebin.com/FPACnC4J <-- this block doesn't do everything
[20:05:11] <fuzzie> mh, i guess it doesn't matter
[20:05:20] <lynxlynxlynx> maybe another instant foo
[20:05:43] <fuzzie> where does it fail?
[20:06:21] <lynxlynxlynx> not sure yet
[20:06:29] <lynxlynxlynx> the actionoverride works
[20:06:47] <lynxlynxlynx> TakeItemReplace is fishy
[20:06:55] <lynxlynxlynx> the statue now has all four rings
[20:07:27] <fuzzie> SetGlobal and CreateCreature are instants
[20:07:27] <lynxlynxlynx> others have just the two hold/stone ones
[20:07:38] <fuzzie> TakeItemReplace will get executed on the queue, and then the override happens
[20:08:15] <lynxlynxlynx> but it looks like the action is buggy - the items were not replaced, but added
[20:08:25] <fuzzie> or i broke it :)
[20:09:20] <lynxlynxlynx> mnja
[20:09:39] <fuzzie> the Sender->Pos looks bad, no?
[20:09:43] <fuzzie> but not relevant for your case
[20:10:20] <lynxlynxlynx> that's only on failure to do it directly
[20:10:39] <lynxlynxlynx> i'll be in there soon
[20:10:41] <fuzzie> but i assume this is a problem where the ring is equipped
[20:11:14] <fuzzie> i guess we should methodically check every action and get rid of this IE_INV_ITEM_UNDROPPABLE|IE_INV_ITEM_EQUIPPED stuff
[20:11:41] <fuzzie> maybe just rethink the whole design
[20:12:44] <lynxlynxlynx> this could be the same bug as with the portal key
[20:12:52] <fuzzie> yes
[20:12:57] <lynxlynxlynx> the rings probably are undroppble, so you couldn't cheat by pickpocketing
[20:13:00] --> |Cable| has joined #GemRb
[20:13:06] <fuzzie> i mean, we don't need this fancy flags stuff
[20:13:42] <fuzzie> just have INV_REMOVE_CHECK_UNDROPPABLE and INV_REMOVE_CHECK_EQUIPPED and INV_REMOVE_CHECK_PICKPOCKETABLE or something
[20:14:07] <-- |Cable| has left IRC (Client Quit)
[20:14:18] --> |Cable| has joined #GemRb
[20:14:30] <lynxlynxlynx> yeah, it has undroppable and whatever 64 is set
[20:15:22] <lynxlynxlynx> err no, that's the mask, only undroppable
[20:15:41] <Avenger> btw, some opcodes got some 'hacking' when they get reloaded
[20:15:53] <Avenger> for example, disease and poison get their duration recalculated
[20:15:57] <lynxlynxlynx> so why not just add IE_INV_ITEM_UNDROPPABLE as the flags to be called with?
[20:16:12] <Avenger> charm also has some hacking
[20:16:49] <lynxlynxlynx> well, it would help here, but i know it wasn't so easy the last time
[20:17:13] <fuzzie> well, i hacked it since last time
[20:19:11] <lynxlynxlynx> so do you have any objection to adding IE_INV_ITEM_UNDROPPABLE to the call?
[20:19:14] <fuzzie> no
[20:19:59] <fuzzie> just worried it'll break again
[20:21:50] <fuzzie> would appreciate it if you mentioned the issue in the commit msg though (that it wasn't removing undroppable items)
[20:22:57] <lynxlynxlynx> it matches the comment in the header of RemoveItem
[20:23:31] <lynxlynxlynx> it worked (eek)
[20:24:37] <fuzzie> since then, i think Avenger changed it twice, then i changed it, then someone else changed it, then i changed it again :-)
[20:24:46] <fuzzie> i am hoping the last time i got it righ
[20:24:47] <fuzzie> t
[20:25:40] <Avenger> i changed it at least twice, i think it was more than once, it is one of the worst flip flops :)
[20:27:39] <CIA-26> GemRB: 03lynxlupodian * r69e327c07310 10gemrb/gemrb/core/GameScript/Actions.cpp:
[20:27:39] <CIA-26> GemRB: TakeItemReplace: also remove undroppable items
[20:27:39] <CIA-26> GemRB: fixes the final Sendai battle
[20:27:39] <CIA-26> GemRB: 03lynxlupodian * rcdbda88c4630 10gemrb/gemrb/plugins/FXOpcodes/FXOpcodes.cpp: fx_set_charmed_state: added fallback for when the caster is not known
[20:37:48] <Avenger> takeitemreplace removes undroppables? nice
[20:37:57] <Avenger> i didn't see that :)
[20:38:39] <fuzzie> i think most of the item actions do
[20:47:08] <fuzzie> heh
[20:47:23] <fuzzie> i just got poison damage when paused, in original engine
[20:53:59] <Avenger> haha there are at least 5 copies of derivedstats in the perform action function (each derivedstats object maintains some 14 lists)
[20:54:26] <Avenger> ok, they are not constructed, just keeping space
[21:02:56] <fuzzie> space is free :p
[21:09:54] <Avenger> ok, i'm done for today
[21:09:55] <Lightkey> I wish :-/
[21:09:55] <-- Avenger has left IRC (Quit: ChatZilla 0.9.86 [Firefox 3.6.8/20100722155716])
[21:11:48] <fuzzie> ok. stack space is free. until you run out.
[21:21:38] <Lightkey> even if I stack 'em, the room seems to be getting smaller and smaller
[21:43:34] --> Avenger has joined #GemRb
[21:43:34] --- ChanServ gives channel operator status to Avenger
[21:44:21] <Avenger> please remind me, what to do if a file was changed while i made a local change on it and i cannot download it
[21:45:03] <fuzzie> 'git stash' will stash local changes, then 'git stash apply' will try to re-apply them later
[21:45:24] <Avenger> i'm not sure it is enough now :(
[21:45:28] <fuzzie> or just 'git pull' should try and merge the changes, but then you maybe get complications and an ugly merge commit
[21:45:33] <fuzzie> so it is a bad idea usually
[21:45:57] <Avenger> git pull didn't work
[21:46:34] <fuzzie> so, what happened
[21:46:35] <lynxlynxlynx> conflicts?
[21:46:41] <fuzzie> you did 'git pull' and it left you in the middle of a failed merge?
[21:46:49] <Avenger> actually git stash seems to work
[21:46:53] <fuzzie> oh, great :)
[21:52:01] <-- Avenger has left IRC (Quit: bye!)
[21:56:06] <CIA-26> GemRB: 03avenger_teambg * r28220982f5c2 10gemrb/gemrb/ (3 files in 2 dirs): RemoveLevelEffects:return the source of the removed spell for further processing
[21:59:55] --> SiENcE has joined #GemRb
[22:00:43] <lynxlynxlynx> another typical commit
[22:04:16] <-- lynxlynxlynx has left IRC (Remote host closed the connection)
[22:08:59] <fuzzie> ye, that is not a very nice commit
[22:09:57] <fuzzie> oh, with '-w' it seems fine
[22:20:52] <-- SiENcE has left IRC (Quit: cya @all)
[22:36:55] <-- Micru has left IRC (Remote host closed the connection)
[22:43:46] --> SiENcE has joined #GemRb
[22:49:32] <SiENcE> for you folks
[22:49:34] <SiENcE> http://sience.schattenkind.net/dingoo/iwd_vnc_test.html
[22:50:02] <SiENcE> vnc session recorded directly from dingoo handheld with gemrb running
[22:50:23] <SiENcE> thats why fps is a little slow
[22:50:27] <SiENcE> ok ..good night
[22:50:39] <-- SiENcE has left IRC (Quit: cya @all)
[23:24:59] <-- Maighstir has left IRC (Quit: Maighstir)