#gemrb@irc.freenode.net logs for 26 Dec 2009 (GMT)

Archive Today Yesterday Tomorrow
GemRB homepage


[00:04:12] <tomprince> Avenger: I'd like some one to try out (http://pastebin.ca/1727092) on windows. There are already a number of cross-module alloc/free, so I don't think that they are a problem in practice.
[00:20:35] --> Forgetful_Lion has joined #GemRB
[02:07:11] --> barra__out has joined #gemrb
[02:24:32] <-- barra_away has left IRC (Read error: 110 (Connection timed out))
[03:45:23] --> Gekz has joined #GemRB
[03:45:35] <Gekz> noo
[04:10:14] <-- Gekz has left IRC (Read error: 60 (Operation timed out))
[06:49:41] <-- Forgetful_Lion has left IRC (" HydraIRC -> http://www.hydrairc.com <- Like it? Visit #hydrairc on EFNet")
[07:43:05] <-- tomprince has left IRC (Read error: 54 (Connection reset by peer))
[07:48:16] --> raevol has joined #GemRB
[10:23:41] <-- raevol has left IRC ("Leaving.")
[10:59:53] <-- |Cable| has left IRC (Remote closed the connection)
[11:06:19] --> tomprince has joined #gemrb
[11:23:31] <-- tombhad-AC has left IRC (Read error: 60 (Operation timed out))
[11:48:05] <-- Nomad010 has left IRC (Remote closed the connection)
[11:53:09] <-- Lightkey has left IRC (Remote closed the connection)
[12:29:02] --> |Cable| has joined #gemrb
[12:29:14] <-- |Cable| has left IRC (Client Quit)
[12:29:22] --> |Cable| has joined #gemrb
[14:09:35] --> Maighstir has joined #gemrb
[14:09:47] <-- |Cable| has left IRC ("Leaving")
[14:09:50] --> |Cable| has joined #gemrb
[14:43:09] --> tombhadAC has joined #gemrb
[14:48:20] <-- Maighstir has left IRC (Read error: 110 (Connection timed out))
[18:34:44] --> Nomad010 has joined #GemRB
[18:44:06] --- barra__out is now known as barra_desktop
[18:46:42] --> lynxlynxlynx has joined #gemrb
[18:46:42] --- ChanServ gives channel operator status to lynxlynxlynx
[19:10:52] <-- tomprince has left IRC ()
[19:11:11] --> tomprince has joined #gemrb
[19:28:47] <-- tomprince has left IRC (Read error: 60 (Operation timed out))
[20:54:08] --> tomprince has joined #gemrb
[22:54:26] <-- lynxlynxlynx has left IRC (Remote closed the connection)
[23:27:31] --> Avenger has joined #gemrb
[23:27:35] --- ChanServ gives channel operator status to Avenger
[23:27:40] <Avenger> hey
[23:27:57] <wjp> hi
[23:28:41] <Avenger> fuzzie: about approaching traps: you simply have to stop BEFORE you walk into one. It should be fairly easy, by simply checking the walk path, as soon as your next step is inside the trap, your current step should be 'close enough'
[23:29:06] <Avenger> i don't think there is any high level math needed :)
[23:29:12] <fuzzie> i wrote that, and i think it gave worse results than bg1, even
[23:29:20] <Avenger> how ?
[23:29:22] <fuzzie> i don't remember why though, maybe i just messed it up
[23:29:47] <Avenger> maybe you didn't calculate with personal space, while the trap triggering code did?
[23:29:48] <fuzzie> it was pathfinding around obstacles, i guess
[23:29:59] <fuzzie> so you had very silly paths which led you into enemies, etc
[23:30:14] <Avenger> well, i just talk about not stepping into traps
[23:30:21] <fuzzie> i guess it just needs some pathfinder magic - i was trying to do a path first and *then* check it
[23:30:22] <Avenger> while approaching them
[23:30:48] <Avenger> the pathfinder shouldn't know the trap polygons when approaching enemies, if you walk in them it is your fault :)
[23:31:36] <Avenger> you just click on the trap, it gets the deactivation point, starts walking towards it whatever path is it, and stops before stepping right into it
[23:31:37] <fuzzie> oh, sure, i just think of the pathfinder logic for when you click 'disarm trap'
[23:31:50] <fuzzie> but the problem is that if you path to the deactivation point, it finds the wrong paths
[23:32:01] <fuzzie> because you have enemies in the way or whatever
[23:32:20] <Avenger> well this isn't a trap specific problem them
[23:32:24] <Avenger> then
[23:32:36] <fuzzie> so we have to do "path to nearest place on trap", but that could be done in pathfinder
[23:32:38] <Avenger> enemies are simply obstacles
[23:33:01] <Avenger> i don't see why there should be any special code?
[23:33:04] <fuzzie> sure, but if you have [you] [trap polygon start] [enemy] [trap deactivation point]..
[23:33:17] <fuzzie> you want to path to the polygon start, not to somewhere else because the pathfinder avoided the enemy
[23:33:44] <Avenger> how often your enemy stands inside the trap?
[23:33:59] <fuzzie> well, i think they were neutrals when i found them in bg2
[23:34:05] <fuzzie> maybe in spellhold somewhere? it was months ago
[23:34:22] <Avenger> so, it cannot walk to it, i guess?
[23:34:27] <Avenger> and in the original, it can?
[23:34:51] <fuzzie> and i just thought it 'nicer' to have some clever math way, rather than adding more pathfinder code
[23:35:46] <Avenger> so, you say, the original doesn't walk to the deactivation point at all?
[23:36:12] <Avenger> it picks the nearest available edge of the trap???
[23:36:17] <fuzzie> i think sometimes the deactivation point is inaccessible
[23:36:18] <Avenger> i doubt that
[23:36:21] <fuzzie> but that is a red herring
[23:36:26] <fuzzie> i think gemrb sets
[23:36:31] <fuzzie> Pos to the centre point specifically to avoid that
[23:36:47] <fuzzie> our code for that is not at all perfect, there are still problems with inaccessible points in iwd/bg2 places, i think?
[23:36:54] <Avenger> maybe we use the wrong field for the deactivation point
[23:37:10] <fuzzie> the others are usually (0,0)
[23:37:10] <Avenger> there are more than one fields
[23:37:24] <Avenger> maybe 0,0 means default anyway
[23:37:31] <fuzzie> maybe so
[23:37:34] <Avenger> so it could be hacked on load time to a sane value
[23:38:13] <Avenger> anyway,i don't think there is any special needed to be done in the pathfinder itself. You might need to pick a better target point, but i think the key is simply stop before stepping in the trap
[23:38:31] <fuzzie> well, the math would just be for a better target point :)
[23:38:46] <Avenger> middle of the nearest edge is good enough?
[23:39:07] <fuzzie> i think the original does something like that
[23:39:11] <fuzzie> i will load bg2 tomorrow and look
[23:39:29] <fuzzie> or someone else can
[23:39:41] <Avenger> it would still stop at a vertex, if it reaches it :) because next step would trigger the trap
[23:40:44] <fuzzie> well, the original game does just trigger the trap, i think :P
[23:40:48] <fuzzie> but we can improve on that
[23:40:51] <Avenger> we could do better
[23:43:09] <fuzzie> i have to go to bed, anyway
[23:43:14] <Avenger> well, i just came in because i saw your talk with nomad
[23:43:19] <Avenger> see you later!
[23:44:23] <Avenger> for tom: yes, there could be cross module memory allocations, but they might cause problems. I rarely check with msvc7
[23:44:50] <Avenger> i run only bg2 at my workplace where the msvc7 environment is
[23:45:45] <Avenger> bye for now!
[23:45:46] <-- Avenger has left IRC ("bye!")
[23:51:41] <-- tomprince has left IRC ()
[23:51:46] --> tomprince has joined #gemrb