#nuvie@irc.freenode.net logs for 14 Aug 2005 (GMT)

Archive Today Yesterday Tomorrow
Nuvie homepage

[03:54:13] <-- Kirben has left IRC (Read error: 104 (Connection reset by peer))
[04:01:35] --> Kirben has joined #nuvie
[07:01:46] --> SB-X has joined #nuvie
[07:34:33] <-- servus has left IRC (Read error: 104 (Connection reset by peer))
[07:35:43] --> servus has joined #nuvie
[09:07:03] --> luteijn has joined #nuvie
[09:21:03] <SB-X> hi lute
[09:21:10] <SB-X> luteijn
[09:22:43] <SB-X> this party following text description i have doesn't seem to match the behavior of walking through doorways in u6 :\
[09:22:53] <SB-X> but I might be misreading it
[09:29:04] <luteijn> hi sbx.
[09:30:08] <SB-X> hows it going?
[09:31:19] <luteijn> okay; busy with real work, didn't do much hobby-work ;)
[09:33:15] <SB-X> oh that's too bad
[09:33:17] <luteijn> Re-reading the following algo. from the forum..
[09:33:26] <SB-X> did your u6o character get restored?
[09:34:24] <luteijn> it wasn't restored yet last time I looked. It doesn't matter that much, but I can't give away any of the stuff it still has on the henchmen unti lit is..
[09:34:45] <SB-X> the stuff will stay won't it?
[09:34:48] <luteijn> the forum does say that the author isn't sure it was used exactly in this way.
[09:35:00] <SB-X> yep
[09:35:05] <luteijn> so far it has ;)
[09:35:37] <SB-X> cool
[09:36:13] <luteijn> It might be a good idea to add this partyfollow.txt to the documentation in cvs (and all the other info from the same person)
[09:37:28] <SB-X> yeah, we could do that but not sure if we want to
[09:38:05] <SB-X> among those forum posts, there is one much longer description of party following that I assume is older and the description i sent you is closer to what is actually used
[09:38:48] <SB-X> the other one has a diagram of links showing how everyone is connected when walking around obstacles
[09:40:18] <SB-X> the other one is more like "this is what we are trying to do" and partyfollow.txt is "this is how we do it"
[09:41:08] <luteijn> the way I m reading this, if you walk through a door, while you are in the doorway, everyone would be in the diamond configuration
[09:42:00] <SB-X> isn't that the only configuration?
[09:44:29] <luteijn> well they have to break that up to go through the door orderly.
[09:45:01] <luteijn> so the next step, you are through the door, positons 1 and 2 are in the walls next to the door, so can't be reached.
[09:45:32] <luteijn> Since you're one step away, this is not considered 'contiguos'
[09:45:38] <luteijn> (misspelled that)
[09:45:59] <luteijn> so they would 'try moving in target direction' (which can't be done
[09:46:39] <luteijn> so then they try all other directions in circular order, towards leader first; so number 1 will move to inside the door, number 2 will fall in line behind him.
[09:47:32] <luteijn> you wrote "Doesn't match U6 behavior, when moving through doorways." directly below 'If adjacent to target, and target is blocked, don't move.'
[09:48:40] <luteijn> but that line wouldn't be valid, as 0 is not 'contiguous' with 1.
[09:49:53] <SB-X> i read it as "contiguous" being relative to the party member's current location, not the target location
[09:50:12] <SB-X> oh, so did you
[09:50:45] * SB-X looks again.
[09:52:39] <SB-X> if leader is zero, after 1 moves inside the door, 2 is contiguous
[09:54:09] <luteijn> hmm yes. Also final configuration depends on DEX as the order of making the move doesn't have to be 1 2 3 4 5 6 7..
[09:54:27] <SB-X> i think dex is only used during combat
[09:55:29] <SB-X> maybe it records the position of everyone at the start of the pass, and uses that position even if they move
[09:56:54] <luteijn> Maybe I should bring out my weichi board and play out the algorhythm ;)
[09:58:04] <SB-X> what is that?
[09:58:11] <servus> Ouija?
[09:58:41] <luteijn> no, oriental game with black and whit stones to place on board of crossed lines.
[09:58:53] <luteijn> aka 'go'
[09:59:06] <SB-X> oh, i've played that
[09:59:36] <luteijn> a text editor works too ;)
[10:00:05] <SB-X> or you can load u6 and push people where you want them to go
[10:02:33] <SB-X> hikaru no go
[10:03:54] <luteijn> okay, so 2 decides not to move as he's already contiguous with 1, 3 will move one step north, as that is not blocked.
[10:05:27] <SB-X> yeah I've got that charted graphically
[10:05:46] <luteijn> hmm but maybe 3 wouldn't do that because the target square is not 'contguous'?
[10:05:46] <SB-X> but 3 and 4 need to switch positions then
[10:06:18] <luteijn> oh sorry, disregard my last line.
[10:09:22] <luteijn> after 3's move there is an L shape of 0 1 3 and 2, 4's target would one north of him, and not blocked at all, so he'll move up one square
[10:09:31] <luteijn> probably now second pass will kick in?
[10:11:17] <SB-X> i didn't think there'd be a 4
[10:11:33] <SB-X> that changes everything!
[10:11:56] <luteijn> moving forward not allowed(blocked by wall) so will try in circular direction, which will make it step next to 3, forming an upside down T.
[10:12:59] <SB-X> remember 2 can displace 3
[10:13:11] <luteijn> But does he need to?
[10:14:19] <luteijn> second pass for 2 wouldn't do anything as 2 is contiguous, and moving forward is not an option and leader didn't move diagonally so there are no options left for a second pass move
[10:15:50] <luteijn> 5,6,7 all can reach their target in first pass, and are then contiguous, not behind target and leader moved in straight line, so no second pass
[10:17:16] <SB-X> where did you get the upside down T?
[10:17:40] <SB-X> i dont see how 2 can move at all there
[10:17:55] <SB-X> since he's already right next to 3
[10:18:21] <luteijn> I made a little drawing as I went along.. I'll send it over.
[10:18:44] <SB-X> oh, good
[10:19:02] <SB-X> my drawings are on paper
[10:19:55] <luteijn> I probably made some mistakes in it ;)
[10:20:06] <SB-X> it's still really nice
[10:20:28] <SB-X> you didn't need a box of #'s, but you included one anyway
[10:21:09] <luteijn> just to make sure they wouldn't pathfind around it or something ;) and makes it a little clearer too
[10:21:13] <SB-X> and the text to the right is clear enough
[10:21:18] <SB-X> ok
[10:21:28] <luteijn> Do all the steps make sense?
[10:21:38] <luteijn> and is this what happens in U6 ;)
[10:21:52] <SB-X> you can load it up and check
[10:22:12] <SB-X> will have to get a party of 8 though
[10:23:25] <SB-X> 4's target isn't contiguous
[10:23:54] <SB-X> they won't move in that case
[10:24:34] <luteijn> I read it as ' target ' is the position relative to the leader they should be in, and marked with letter
[10:24:49] <SB-X> yeah
[10:25:04] <luteijn> first pass just looks if 4 is already contiguous (no) then follows directions in that part of the 'if'
[10:25:29] <luteijn> first step is to try moving in target direction, which works, so done with pass 1
[10:25:37] <SB-X> but it won't move to the target because it wouldn't be contiguous then
[10:25:54] <SB-X> oh i see
[10:26:06] <SB-X> it's already not contiguous, so who cares
[10:26:13] <luteijn> hmm well maybe you are right, as 'try a move' needds the target to be contiguous?
[10:26:34] <SB-X> yeah, but there would be an exception in that case
[10:26:47] <luteijn> lets see what happens if it doesn't have the exception..
[10:27:02] <SB-X> the other following doc gives priorities to moves
[10:27:08] <luteijn> it would then 'try all other directions..' so probably move north-east
[10:27:38] <luteijn> so ends up in the same place I made it end up after pass 2
[10:28:06] <luteijn> and pass 2 would be a no-move, as already contiguous, not behinf target etc.
[10:28:12] <SB-X> great
[10:28:43] <SB-X> you repeated the last step
[10:29:15] <SB-X> nm, it is to show them moved
[10:30:12] <SB-X> so does each member get a first then second pass, or does it give them all a first pass, then give them all a second pass?
[10:31:00] <luteijn> I interpreted it to be 'depth first', but already was doubting if it might be width-first..
[10:31:25] <luteijn> in this case it wouldn't matter as no-one needs to move in their second phase.
[10:32:01] <SB-X> but 2 and 3 are still in the wrong positions
[10:32:27] <luteijn> for which definition of 'wrong'?
[10:32:37] <SB-X> eachother's
[10:34:09] <luteijn> hmm after the next step, 2 will just step into the doorway from where he is, so it's not that wrong.
[10:35:52] <SB-X> it's not wrong after the next step
[10:36:50] <luteijn> would be interesting to check if Iolo (3) is behind dupre or shamino is behind dupre in U6 after avatar just through the door.
[10:41:41] <luteijn> What would make more sence; everyone first gets a 1st pass, then everyone gets 2nd pass or give everyone two steps, then move on to next character?
[10:45:55] <SB-X> No idea about which one of those makes more sense (but I like the first better). Shamino is behind Dupre and Iolo is to the right of him. That's why it says "Does'nt match U6 behavior, when moving through doorways."
[10:46:05] <SB-X> Doesn't*
[10:46:48] <SB-X> although I wrote that in the wrong location
[10:48:47] <SB-X> then if you Pass, Dupre moves to the left of the Avatar, Shamino moves up, and Iolo moves left
[10:49:06] <SB-X> then if you Pass again, Dupre moves back into the doorway and Shamino moves to the right of the avatar
[10:49:30] <SB-X> hold spacebar and they keep switching positions
[10:49:33] <SB-X> it's really amusing
[10:49:43] <luteijn> So they actually even try to overshoot their 'target' in the formation
[10:51:52] <SB-X> they behave a little differently when passing than when moving, but it could still be the exact same code
[10:55:22] <luteijn> Did you get anywhere with that magic-stuff, or I've made such a mess of it that you've decided to look at the path-finding instead?
[10:55:55] <SB-X> heh
[10:56:09] <SB-X> I was actually planning to finish pathfinding first, but got tied up with other things.
[10:56:25] <SB-X> there shouldn't be any problem with the magic, should there?
[10:56:32] <SB-X> just merge it in
[10:56:42] <SB-X> ...is what I'll do
[10:56:58] <SB-X> merge it with current cvs that is
[10:57:05] <SB-X> and resolve conflicts
[10:58:59] <luteijn> That should work. For path finding, I think you might want to just implement the given algorythm, then see if there's anything disastrously wrong with it..
[11:00:13] <SB-X> that's what has been done so far, but it's crashing mysteriously
[11:02:36] <luteijn> mysteriously as in 'no usuable info in the core dump'?
[11:02:47] <luteijn> or 'non reproducable'
[11:05:28] <SB-X> as in I don't know why it's crashing, probably not that mysterious after all
[11:06:03] <SB-X> the core dump might be useful
[11:06:30] <SB-X> PathFinder::check_loc (this=0x18, loc=@0x18)
[11:06:33] <SB-X> hehe
[11:06:54] <SB-X> this is due to a design flaw :p
[11:07:01] <luteijn> hmm looks like some sort of index instead of pointer is used
[11:07:30] <SB-X> for some reason my corefile was numbered
[11:07:40] <SB-X> or nuvie's corefile
[11:07:47] <SB-X> is that normal?
[11:08:03] <SB-X> probably the pid
[11:08:39] <luteijn> hmm never seen it; maybe feature of your shell?
[11:09:35] <SB-X> must be new because in the past it was always named "core"
[11:10:19] <SB-X> havn't upgraded bash or glibc though
[11:10:27] <SB-X> ah well
[11:11:29] <SB-X> also loc in check_loc is supposed to be a MapCoord (not a pointer)
[11:12:00] <luteijn> http://www.ussg.iu.edu/hypermail/linux/kernel/9603.0/0045.html
[11:12:29] <SB-X> hmm, except it's also a reference
[11:15:05] <SB-X> heh, that proc file doesn't exist anymore
[11:15:19] <SB-X> but i do have /proc/sys/kernel/core_pattern
[11:15:26] <SB-X> and it just says "core"
[11:16:25] <SB-X> and /proc/sys/kernel/core_uses_pid = 0
[11:18:06] <luteijn> strange..
[11:27:04] <SB-X> do you think u6o will add the rest of the party members?
[11:27:33] <SB-X> and should they, will anyone care
[11:28:35] <luteijn> I don't think they will
[11:29:41] <SB-X> only a few of them have importance to the quest
[11:30:39] <luteijn> They don't have much use except as cannonfodder and torch bearer. Too easy to lose their equipment if you lose connection etc.
[11:30:51] <luteijn> bbl, have to cook lunch)
[11:30:56] <SB-X> ok
[11:31:00] <-- luteijn has left IRC ("[BX] They killed Kenny! THOSE BASTARDS!")
[13:58:43] <-- Kirben has left IRC ("System Meltdown")
[17:01:51] <SB-X> oops
[17:01:55] <SB-X> void set_pathfinder(PathFinder *pathfinder) { pathfinder = pf; }
[17:01:59] <SB-X> ...as I said, design flaw :)
[17:02:09] <SB-X> close enough
[17:49:45] <-- SB-X has left IRC ("casts gate travel")