#gemrb@irc.freenode.net logs for 13 Aug 2010 (GMT)

Archive Today Yesterday Tomorrow
GemRB homepage


[00:22:18] --- Gekz_ is now known as Gekz
[00:55:22] --> raevol has joined #GemRb
[00:58:33] <-- Maighstir has left #GemRb
[01:13:16] <-- edheldil_ has left IRC (Ping timeout: 246 seconds)
[01:18:54] --> edheldil_ has joined #GemRb
[01:30:32] <-- edheldil_ has left IRC (Ping timeout: 255 seconds)
[06:55:36] <-- raevol has left IRC (Quit: Leaving.)
[07:24:04] --> SiENcE has joined #GemRb
[07:38:18] <-- |Cable| has left IRC (Remote host closed the connection)
[09:04:18] --> lynxlynxlynx has joined #GemRb
[09:04:18] --- ChanServ gives channel operator status to lynxlynxlynx
[09:04:32] <fuzzie> morning
[09:05:09] <CIA-26> GemRB: 03fuzzie * r95d21abf17df 10gemrb/gemrb/plugins/DLGImporter/DLGImporter.cpp: add a comment to DLGImporter's GetStrings with notes about test cases
[09:07:27] <fuzzie> i think the actual fix there is probably to simply treat Test("hi, 1) as Test("hi", 1), but i don't have the original engine here to try it
[09:09:21] <CIA-26> GemRB: 03fuzzie * r2d9a9ae3c885 10gemrb/gemrb/core/Interface.cpp: don't print game features at startup
[09:12:05] <lynxlynxlynx> oj
[09:18:20] <fuzzie> so, if i make a kensai, i see only the kit effects
[09:18:29] <fuzzie> this is the problem?
[09:22:17] <fuzzie> i guess so
[09:24:04] <fuzzie> well, that isn't a very difficult problem to debug
[09:25:09] <fuzzie> the chargen is creating an actor which doesn't support effect sources
[09:26:43] <fuzzie> so no wonder i couldn't debug it :)
[09:32:47] <lynxlynxlynx> how can an actor not support it?
[09:33:08] <fuzzie> bg1 actors have bg1 effects, which don't have a source
[09:33:30] <fuzzie> i added an "override version" to force an upgrade, testing it now
[09:34:20] <fuzzie> (maybe we should simply force versions globally, but we already do it this way for games, can always be changed)
[09:43:51] <CIA-26> GemRB: 03fuzzie * r7e447d937585 10gemrb/gemrb/ (6 files in 4 dirs):
[09:43:51] <CIA-26> GemRB: allow VersionOverride of new actors made by CreatePlayer, use in bg2 chargen
[09:43:51] <CIA-26> GemRB: (maybe we should replace this and the one in LoadGame with a config option?)
[09:44:19] <fuzzie> lynxlynxlynx: seems to work for me now
[09:47:08] <lynxlynxlynx> cool
[09:50:19] <fuzzie> but new ToB games are still broken in a lot of other ways, of course..
[09:53:17] <fuzzie> i don't know if you care about fixing that before the release
[09:55:20] <fuzzie> re healing in temples: you can end up with everything being disabled, so i'm pretty sure that means you can't buy healing when uninjured
[09:56:23] <lynxlynxlynx> is this info from bg2? I saw one bg1 screenshot like that, but i don't remember being limited in bg2
[09:56:44] <fuzzie> yes, bg2
[09:56:59] <lynxlynxlynx> for tob i'll maybe only look into that pp exit issue
[09:57:26] <fuzzie> i did an original-engine fixpacked SoA complete-ish run (as many subquests as I could find) in July
[09:58:29] <fuzzie> the idea being that i can then come and fix SoA in gemrb with it all relatively fresh in my mind, i like to think it's going well :)
[10:08:33] <fuzzie> hmm
[10:09:37] <fuzzie> IESDP is really unhelpful sometimes
[10:10:03] <lynxlynxlynx> is nalia's keep and imnesvale fixed? i remember you mentioned problems there
[10:10:20] <fuzzie> i haven't checked either
[10:10:48] <fuzzie> i think nalia's keep probably still has that heightmap bug
[10:11:03] <fuzzie> Taimon said the original engine doesn't do anything complicated there, so i don't understand it
[10:13:56] <fuzzie> i guess the temple stuff is just hardcoded, meh
[10:14:34] <fuzzie> sppr303, 307, 417, 713, 212, 404, 504 and 712..
[10:14:49] <fuzzie> right in the temple code of the original. bah! :)
[10:18:31] <lynxlynxlynx> bleh
[10:22:42] <fuzzie> if it's in the spells then i can't see any relevant flag differences, anyway
[11:09:26] <fuzzie> lynxlynxlynx: do you have time to confirm my search map bug?
[11:10:23] <lynxlynxlynx> the jar one?
[11:10:31] <fuzzie> no
[11:10:59] <fuzzie> my theory is that fog-of-war/etc checks are done starting at a point too far to the right (maybe also down)
[11:11:27] <fuzzie> so if you stand next to a wall, you can no longer see further than a couple of squares, because it's starting inside a wall
[11:11:59] <fuzzie> i checked this using the Spellhold entrance in a few places (if you turn on search map view, you can see it very clearly - try for example the right wall just before you come to the first prison cell)
[11:12:10] <fuzzie> but would like to make sure it's not just my compiler being broken or something
[11:13:00] <lynxlynxlynx> will check, but later
[11:13:12] <fuzzie> no rush
[11:14:37] <fuzzie> i am just going down my todo list
[11:14:48] <fuzzie> just posted a request on the IESDP forum for more information about triggers
[11:22:30] <fuzzie> but i don't want to ask them about anything which is pretty easily researched by me
[13:03:10] <lynxlynxlynx> i do get a bit of shade from the wall
[13:05:29] <lynxlynxlynx> http://lynxlynx.info/bugs/spellhold-fog.jpg
[13:05:43] <lynxlynxlynx> it looks more like it is due to the wall uncovering we do though
[13:15:29] <fuzzie> heh
[13:15:40] <fuzzie> you have to pick a wall which is to your right, i think
[13:16:42] <fuzzie> the one on the right side of the doorway there (so, sloping upward and to the right) was i think what i was looking at
[13:19:31] <lynxlynxlynx> yeah, there it is definitely wrong
[13:20:52] <lynxlynxlynx> eyesight should be computed from the center of the actor anyway; i can't remember any mosters with off-center eyes
[13:23:03] <fuzzie> it tries, but then this weird code picks the wrong squares
[14:15:50] <Edheldil> I wonder what is all the Windows GDI/GUI stuff doing in PS:T. Is it just linked in, or was it used for some ingame editor?
[14:17:27] <lynxlynxlynx> assert dialog and also the exit one if they have one
[14:17:55] <lynxlynxlynx> (if you close by x in windowed mode)
[14:21:24] --> Maighstir has joined #GemRb
[14:25:37] <fuzzie> Edheldil: depends which stuff you're talking about :)
[15:03:38] <SiENcE> hey, fuzzie ... when do you plan to move the keybinding to py?
[15:03:56] <SiENcE> should i wait for the first release port?
[15:04:55] <fuzzie> i want to fix scripting first
[15:05:06] <fuzzie> so not entirely sure
[15:05:16] <SiENcE> ~1week 1month?
[15:05:41] <fuzzie> and maybe fixing SDLAudio is important, if it is better
[15:06:16] <fuzzie> but maybe openal is also fine, now i fixed the ACM, maybe you want to check
[15:06:21] <fuzzie> it all depends on what is most important :)
[15:06:37] <fuzzie> lynxlynxlynx: is the Source thing fixed, btw?
[15:06:49] <lynxlynxlynx> yes
[15:07:05] <SiENcE> fuzzie, ah of couse i can test openal again
[15:07:10] <SiENcE> good idea
[15:07:33] <SiENcE> yep sdlaudio would be good. movie sound is not working and soundfx have a big lag
[15:07:39] <SiENcE> but you know
[15:08:09] <fuzzie> the soundfx leak memory, you don't want to ship it like that :)
[15:08:25] <SiENcE> uhm really?
[15:08:27] <SiENcE> :/
[15:13:42] <fuzzie> it only had an hour of work so far
[15:13:57] <fuzzie> but i don't want to spend time fixing it, unless it is really better than openal
[15:19:00] <SiENcE> i will test it
[15:55:28] <-- SiENcE has left IRC (Quit: @all: cya)
[17:10:33] --> SiENcE has joined #GemRb
[17:11:10] <SiENcE> fuzzie, openal sounds bad. lags and really bad fps
[17:11:38] <SiENcE> sdlaudio music is fine and only 1fps less than noaudio plugin
[17:11:57] <SiENcE> i also have two small suggestions for the messagebox
[17:13:21] <SiENcE> first, when message box is opend during a conversation, please center the last textlines on the buttom....not at the top of the messagebox
[17:14:00] <SiENcE> an second, please make the messagebox scrolling using the cursor Up, Down keys
[17:18:30] <lynxlynxlynx> what do you mean with the first?
[17:22:09] <lynxlynxlynx> the second relies on keybindings, but even more improved than in the original - we'd need them context aware too or you'd lose the ability to pan the view with them
[17:27:57] <fuzzie> for a dialog, it would be ok, i think
[17:28:06] <fuzzie> we have to be context aware anyway
[17:29:00] <fuzzie> but should probably fix sdlaudio etc first
[17:29:15] <fuzzie> and #1 fix is cutscenes which i still didn't do because today i turn out to be quite busy so far :(
[17:31:36] <fuzzie> but i guess no-one except me is too worried :)
[17:32:27] <lynxlynxlynx> you're the only one that knows what is wrong
[17:38:05] <SiENcE> lynxlynxlynx, i mean my two suggestions
[17:38:16] <SiENcE> fuzzie, you understand my points?
[17:38:47] <fuzzie> not the first: i think you have to explain what goes wrong for you at the moment
[17:39:07] <fuzzie> it shouldn't matter whether it's at the bottom or the top, i would think
[17:39:09] <SiENcE> ok. i make two screenshots to explain
[17:39:47] <SiENcE> ah no...there is too much space left where is nothing...and i dont see the first choice
[17:39:58] <SiENcE> so i everytime i have to scroll up
[17:40:12] <fuzzie> ok, so the bug is the space :)
[17:40:30] <SiENcE> i do two screens to show you
[17:40:34] <SiENcE> :)
[17:40:38] <fuzzie> that would be good
[17:41:13] <SiENcE> openal is always bad
[17:42:28] <fuzzie> ok
[17:42:44] <fuzzie> will fix sdlaudio, and see if the other porters find it better too
[17:43:20] <fuzzie> did you make it work properly with autotools or did you just hack it?
[17:47:42] <SiENcE> it works prop. with autotools
[17:49:50] <fuzzie> you have a patch? :)
[17:51:55] <SiENcE> not now
[17:52:00] <SiENcE> later
[17:52:26] <SiENcE> here are two screenshots
[17:52:29] <SiENcE> http://img840.imageshack.us/g/screenshot2g.png/
[17:52:45] <SiENcE> second is when conversation starts
[17:52:53] <SiENcE> only choices 2,3,4
[17:53:02] <fuzzie> ah, i see
[17:53:08] <SiENcE> i have to scroll up and down to see 1,2,3,4
[17:53:26] <SiENcE> the text should be on the bottom
[17:53:28] <SiENcE> :)
[17:53:41] <SiENcE> and please scroll up/down using cursor up/down
[17:53:47] <SiENcE> this makes it easier
[17:57:16] --> edheldil_ has joined #GemRb
[18:28:19] <Lightkey> SiENcE: so unusual to see you all talkie-talkie compared to elsewhere :p
[18:35:16] <-- edheldil_ has left IRC (Ping timeout: 276 seconds)
[18:51:36] <SiENcE> Lightkey :-)
[18:52:21] <SiENcE> time is short...to short
[18:53:00] <lynxlynxlynx> to short to talk to Lightkey? ;)
[18:54:37] <Lightkey> zing..
[18:54:54] <SiENcE> ah no :)
[18:55:08] <SiENcE> to short todo all things i want
[18:55:24] <Lightkey> that's a given
[19:17:05] --> edheldil_ has joined #GemRb
[19:21:41] <lynxlynxlynx> well, that was easy
[19:21:57] <lynxlynxlynx> the first print i did after the breakpoint was hit showed the problem
[19:22:42] <lynxlynxlynx> when my pc talks to cespenar, edwin is the sender
[19:23:50] <lynxlynxlynx> showed the problem with my memory :)
[19:23:59] <fuzzie> do you have the whole party selected?
[19:25:45] <fuzzie> if so, that is Avenger's fault :P
[19:26:15] <fuzzie> i interrupted him for something else while he was coding it, i think
[19:26:29] <lynxlynxlynx> nah
[19:29:47] <CIA-26> GemRB: 03fuzzie * r999c41280741 10gemrb/gemrb/core/GUI/GameControl.cpp: use the first PC when more than one is selected
[19:34:32] <fuzzie> maybe more sensible to keep the selected list ordered, then we can simplify quite a bit of the code there
[19:48:09] --> Avenger has joined #GemRb
[19:48:10] --- ChanServ gives channel operator status to Avenger
[19:48:59] <lynxlynxlynx> oj
[19:49:30] <lynxlynxlynx> fuzzie: dlg/bg2/botsmith.d:1300
[19:49:32] <Avenger> hello
[19:49:54] <Avenger> lynx: do you give the -2 saving throw bonus for paladin?
[19:50:12] <lynxlynxlynx> i'm debugging why cespenar doesn't take the ingredients and while he does take the hammer when run in gdb (hmpf), the second TakePartyItemNum never gets run
[19:50:20] <lynxlynxlynx> Avenger: no, that's why it is on the todo
[19:51:07] <lynxlynxlynx> it's easily hardcoded into chargen, but that wouldn't cover original saves if they did it dynamically (which i don't know yet)
[19:51:21] <lynxlynxlynx> the tip was from the forum
[19:51:49] <lynxlynxlynx> fuzzie: oh, the DestroyGold does get run
[19:52:47] <Avenger> i looked into the engine, it works like this: get the savewar value, and subtract 2
[19:52:49] <lynxlynxlynx> err, didn't check that one, but TakePartyGold does work
[19:53:07] <Avenger> the best would be to add a new column to clskills: savebonus
[19:53:36] <lynxlynxlynx> yeah
[19:53:39] <Avenger> i could add it if you are busy
[19:53:58] <lynxlynxlynx> sure, i'm busy ;P
[19:59:04] <Avenger> btw, i think the original engine allows you to have clabma01 (maybe even school specific clabs), they just didn't supply any
[20:00:13] <lynxlynxlynx> with action debugging enabled, the next thing it does after the first TakePartyItemNum is the cutscene start in the next dialog state
[20:00:16] <Avenger> eep, iwd's clskills is not compatible with the rest
[20:00:26] <lynxlynxlynx> yep
[20:00:43] <Avenger> oops actually pst is wrong too
[20:01:06] <lynxlynxlynx> if you're going to move this around, you'll need to check all the users
[20:01:07] <Avenger> wouldn't this cause problem with the common code?
[20:01:39] <Avenger> i will modify only bg1 and bg2 then
[20:01:46] <lynxlynxlynx> there's no common users that i can think of
[20:02:22] <lynxlynxlynx> another big todo for the guiscripts: use col/row names for lookup as much as possible
[20:04:53] <fuzzie> hmm
[20:05:05] <fuzzie> that can get a bit slow if you're doing it a lot
[20:05:12] <fuzzie> maybe not relevant, don't know
[20:06:14] <Avenger> lynx: LuCommon:SetupSavingThrows uses clskills with numeric columns :)
[20:07:27] <lynxlynxlynx> yes, but until the used columns match, there's no problem
[20:08:30] <lynxlynxlynx> fuzzie: i think we have both values in few places already, so for some there would be no extra lookup needed
[20:09:50] <fuzzie> ok
[20:10:30] <fuzzie> re botsmith.d: the issue is that DestroyItem is never run?
[20:13:05] <fuzzie> hmph
[20:13:43] <lynxlynxlynx> i don't know about that and it shouldn't really matter, since it is not about the party
[20:13:56] <lynxlynxlynx> the second TakePartyItemNum doesn't get run
[20:14:08] <fuzzie> well, actions are executed in order
[20:14:25] <fuzzie> and DestroyItem isn't an instant --> is never run
[20:14:25] <lynxlynxlynx> http://pastebin.com/P1q0P2pc
[20:14:30] <fuzzie> i'll have to do more research
[20:14:39] <fuzzie> this that is 'intended behaviour' in gemrb right now
[20:15:26] <lynxlynxlynx> it basicly cleared the queue
[20:15:35] <fuzzie> dialogs clear the queue when you go to a new dialog
[20:15:55] <fuzzie> and so if the game is paused, your actions don't get run
[20:16:20] <lynxlynxlynx> but the first two are instants and ignore this limit?
[20:16:35] <fuzzie> yes, they're executed immediately, even if paused or held or whatever
[20:17:27] <fuzzie> three possibilities: dialog pause weirdness (since this isn't on a party member), the queue clearing being more complex than i thought, or instants in dialogs always being executed (but that last one is unlikely)
[20:18:16] <fuzzie> oho
[20:18:24] <fuzzie> it is probably #2
[20:18:48] <lynxlynxlynx> i was just about to say "too bad, i kinda hoped you'd immediately point out the problem" :)
[20:19:37] <fuzzie> the next actions executed are instants, and someone claims on the forum that the existing queue isn't wiped if only instants are queued, although i really have to verify that first
[20:19:44] <fuzzie> just as well, i need to verify some other dialog stuff
[20:20:12] <lynxlynxlynx> ok
[20:20:20] <lynxlynxlynx> sounds plausible too
[20:21:50] <fuzzie> > I've since cleaned things up so that non-instant actions in dialogues I modify only ever occur in EXIT states (with the exception of a couple things that would be too heinous to modify). BioWare seems to have been pretty careful about this
[20:21:54] <fuzzie> ^- so much for that
[20:27:56] <fuzzie> it is entirely possible that this is using messages, actually
[20:27:56] <fuzzie> sigh
[20:28:16] <fuzzie> but i'll see if we can do something for now, until Avenger has the whole binary mapped out :p
[20:29:23] <lynxlynxlynx> there's no rush
[20:29:38] <lynxlynxlynx> on the other hand we don't know how many times such situations occur
[20:32:03] <lynxlynxlynx> now i've found part of the pocket plane exit problem
[20:32:26] <lynxlynxlynx> i teleported there and checked that the locations are properly stored and they were
[20:32:36] <lynxlynxlynx> save and reload and they're gone
[20:33:17] <fuzzie> :(
[20:34:18] <fuzzie> so GetPlaneLocationCount() is fine before saving?
[20:35:27] <lynxlynxlynx> i didn't check, but i'll retry - the entries were fine
[20:35:42] <lynxlynxlynx> there's a note in the loader that suggests the offsets may be wrong
[20:36:06] <fuzzie> stream->WriteDword( &SavedLocOffset);
[20:36:19] <fuzzie> ^- it doesn't write the count after this
[20:36:41] <fuzzie> also using Signature for padding is *really* silly
[20:36:43] <lynxlynxlynx> GetPlaneLocationCount() is actually at 2 now
[20:36:48] <fuzzie> it doesn't cost us more to make a padding variable
[20:38:10] <fuzzie> oh and of course it should be PPLocOffset anyway!
[20:38:49] <fuzzie> well, i can fix that for you, if you don't do it
[20:39:11] <lynxlynxlynx> PutPlaneLocations is never called
[20:40:13] <fuzzie> well, that too
[20:40:20] <fuzzie> but it would be a bit worthless without writing an offset or a count :)
[20:42:05] <lynxlynxlynx> ok, i'll try to fix this
[20:44:04] <fuzzie> obviously you need to pad with 4 less bytes, after writing the PPLocCount
[20:44:18] <fuzzie> otherwise it looks doable
[20:44:57] <Avenger> ok, bg1 and bg2 paladins got their bonus
[20:45:10] <CIA-26> GemRB: 03avenger_teambg * r431028556ca1 10gemrb/gemrb/ (3 files in 3 dirs): implemented SAVEBONUS (in bg1/bg2 use it to give the paladin's saving throw bonus)
[20:46:05] <fuzzie> oh dear, you're on Windows again?
[20:46:49] <fuzzie> i should work out how to make git ignore line-ending changes
[20:48:23] <fuzzie> or i guess just leave the whitespace detection on my default
[20:52:27] <fuzzie> Avenger: ok, how do i edit dialogs in DLTCEP?
[20:52:35] <fuzzie> it won't let me edit half the fields, they're greyed out
[20:53:06] <Avenger> if you run it under wine, forget it ;(
[20:53:18] <fuzzie> no, native Windows XP
[20:53:21] <Avenger> oh good
[20:53:22] <fuzzie> i have to use weidu to edit dialogs?
[20:53:29] <Avenger> weidu???
[20:53:31] <Avenger> no ;)
[20:53:38] <fuzzie> i want to change the 'linking dialog' field
[20:54:15] <fuzzie> ah!
[20:54:19] <fuzzie> there is a 'modify link' option
[20:54:23] <fuzzie> that is confusing :P
[20:54:44] <Avenger> ahh, just wanted to tell
[20:54:53] <Avenger> i haven't used this for a long time
[20:55:12] <fuzzie> maybe weidu is easier though
[20:55:26] <fuzzie> this is not so smart, it is recreating the entire treeview if i use that :P
[20:55:51] <Avenger> yep, it needs a major rewrite, i just rarely edit dialogs
[20:58:37] <fuzzie> meh, dialogs not so simple
[21:05:28] <fuzzie> so, you can transition to/from banter dialogs as much as you want
[21:05:50] <lynxlynxlynx> http://pastebin.com/8JknBpgR <-- how does this look?
[21:06:27] <fuzzie> either the original stashes the interdia entries somewhere i didn't look, or the dialog code looks them up on-the-fly as tomprince's patch does
[21:08:29] <fuzzie> that looks fine at a glance, only looking at the diff (no gemrb here)
[21:09:17] <fuzzie> but, well, easy to check if it works, i hope
[21:09:33] <lynxlynxlynx> it works for me as far as saving goes, but doesn't prevent that pre-cutscene hang
[21:09:43] <fuzzie> i can't reproduce that here, i think
[21:10:01] <lynxlynxlynx> the fact that we didn't save saved locations wasn't a problem since they're used only in cutscenes, i guess
[21:10:05] <fuzzie> unless it's right after a dialog
[21:13:28] <lynxlynxlynx> the last dialog was before the teleport to the pp
[21:16:25] <lynxlynxlynx> oh, now i write 4 bytes more of the filling, not less, but this is the length of the header in the iesdp
[21:16:41] <fuzzie> just look at the read code?
[21:17:52] <lynxlynxlynx> it does even more :)
[21:18:12] <lynxlynxlynx> oh, that's pst, the others match
[21:19:39] <CIA-26> GemRB: 03lynxlupodian * r82a48daae790 10gemrb/gemrb/plugins/GAMImporter/GAMImporter.cpp: save the saved (normal and pp) locations
[21:20:03] <Avenger> omg, that was not finished?
[21:21:06] <lynxlynxlynx> apparently
[21:21:34] <fuzzie> we should go through and fix everything else we don't bother saving (except iwd2 creature stuff, which is a lot of work)
[21:25:08] <lynxlynxlynx> bah, why do things always start working when i set a bunch of breakpoints
[21:25:57] <fuzzie> if that actually happens (not just coincidence) then do check valgrind :)
[21:26:18] <lynxlynxlynx> it usually is a coincidence
[22:03:41] <-- lynxlynxlynx has left IRC (Remote host closed the connection)
[22:13:39] <fuzzie> > ReactionGT(Player1,NEUTRAL_UPPER))
[22:13:41] <fuzzie> ^- more meh
[22:28:00] <-- edheldil_ has left IRC (Ping timeout: 245 seconds)
[22:30:25] <-- Avenger has left IRC (Quit: bye!)
[22:48:40] --> SiENcE_ has joined #GemRb
[22:50:49] <-- SiENcE has left IRC (Ping timeout: 252 seconds)
[23:30:57] <-- Maighstir has left #GemRb
[23:39:18] <-- SiENcE_ has left IRC (Quit: cya @all)
[23:58:15] --> |Cable| has joined #GemRb