--- Log opened Mon Apr 06 00:00:08 2009
--- Day changed Mon Apr 06 2009
20090406 00:00:08< Sapient> cya later, have fun.
20090406 00:00:08-!- Sapient [n=sapien-x@wesnoth/developer/sapient] has left #wesnoth-dev []
20090406 00:01:12< alink> not a crash, but still not nice
20090406 00:01:57< alink> grr I hate these WML event changing things in the middle of the code's flow
20090406 00:05:32< Crab_> alink:  Second part about : none of the attacks hit ?
20090406 00:05:44< alink> yes
20090406 00:06:51< alink> it seems that none of the attacks are even tried
20090406 00:07:24 * Crab_ fires up gdb again...
20090406 00:07:45< alink> I suppose that one of the checks we were talking about  cancels the attack when the unit change
20090406 00:09:12-!- wesbot changed the topic of #wesnoth-dev to: 1.6.1 planned for Sunday, April 12th, stringfreeze for branches/1.6 ! | Summer of Code info: http://www.wesnoth.org/wiki/SummerOfCodeIdeas | 51 bugs, 231 feature requests, 10 patches | logs: http://irclogs.wesnoth.org/ | Don't paste on IRC! Use a pastebin: http://wesnoth.pastebin.com | http://imagebin.org
20090406 00:10:36< alink> Crab_: btw one mainline example where I think the defender may change side or is_human during the attack is when the attacker has the plague ability and kill the defender
20090406 00:10:49< Crab_> alink: but this will be after the attack
20090406 00:11:21< alink> yes but we use defender_human after the attack too
20090406 00:11:49< corn> ok. I am going to go with bargraphs instead of line graphs and let the end user pick out the granularity (bar width)
20090406 00:11:50< alink> but the new undead will normally not advance with his 0xp
20090406 00:12:21< Crab_> yes
20090406 00:14:35< Reisiger> Ivanovic: I've submitted the patch.
20090406 00:20:06< Crab_> alink:  it gets to "fire_event("attack");" in src/actions.cpp:948 successfully. and fire_event("attack"); throws an exception, attack_end_exception. and so it ends.
20090406 00:20:38-!- zookeeper [n=l@wesnoth/developer/zookeeper] has quit []
20090406 00:22:27< alink> Crab_: no idea why we do that
20090406 00:23:02< alink> but we probably should call set_attack and set_movement after that exception not before
20090406 00:23:16< crimson_penguin> anyone have any ideas about this? http://www.wesnoth.org/forum/viewtopic.php?f=4&t=24587
20090406 00:23:23 * alink is not familiar with this area of code
20090406 00:25:55< alink> Crab_: maybe we should ask a WML guru if he can change this WML to avoid killing the unit just for adding the dehydrated icon
20090406 00:26:27< Crab_> alink: I'll step through the code for some more time..
20090406 00:26:55< alink> because, in fact, if the WML kill the unit that we are attacking, it's normal that we stop
20090406 00:27:08< alink> (but not normal to lose our attack and MP)
20090406 00:27:35< Crab_> alink: understood. so, wml is killing that poor defender and re-creating it ?
20090406 00:28:03< alink> Crab_: yes but no idea if it's a common WML action
20090406 00:28:50< alink> Crab_: check the savegame line 19209
20090406 00:29:48< alink> but make the code more robust against every WML is always good :)
20090406 00:30:08< cjhopman> Crab_: i think it also changes the unit type (not just add dehydrated)
20090406 00:31:06< Crab_> alink: thanks. I see ) it's good to work on the same bug from different directions )
20090406 00:31:16< Reisiger> Night everyone. I'm in need of some shut-eye ;-)
20090406 00:31:23< Crab_> the question is: what should be the *correct* behavior ?
20090406 00:31:25-!- Reisiger [n=Reisiger@adsl-84-226-16-20.adslplus.ch] has quit ["Verlassend"]
20090406 00:31:38< alink> cjhopman: i don't see that, but i am a WML noob :-/
20090406 00:32:10< cjhopman> you're right, it doesnt
20090406 00:32:43< alink> Crab_: generally I ask zookeeper what crazy WML authors want to do
20090406 00:32:51< alink> or Sapient
20090406 00:33:09< Crab_> ok ) but the end result is not what the player expects, thats for sure :)
20090406 00:35:09< Crab_> alink:  because there can be many such situations. all it requires is a 'sighted' event which kills a unit that can be attacked as part of the move that triggers that sighted event
20090406 00:36:04< alink> the lost of attack is not related to the sighted event
20090406 00:36:19-!- YogiHH [n=chatzill@c174121.adsl.hansenet.de] has left #wesnoth-dev []
20090406 00:36:30-!- Espreon [n=chatzill@wesnoth/developer/espreon] has quit ["ChatZilla 0.9.84 [Firefox 3.0.8/2009032609]"]
20090406 00:36:30< alink> that's the problem with 2 parts bugs, it's confusing :/
20090406 00:36:35< Crab_> but it is related to the fact that the enemy was killed, isn't it ?
20090406 00:37:03< alink> yes during the attack event
20090406 00:37:30< Crab_> and enemy was killed by WML as a consequence of that 'sighted' event.
20090406 00:38:20< Crab_> and that 'sighted' event fired because moves can trigger such events, and we have got 'move-and-attack' situation
20090406 00:38:50< alink> mmh yes sorry the same code is in the sighted event and in the attack event few lines below
20090406 00:39:04< alink> your patch already fix the sighted event part
20090406 00:39:29< alink> btw this seems to fix it too : http://wesnoth.pastebin.com/m5d7ac03b
20090406 00:39:39< alink> a bit more annoying but possibly safer
20090406 00:39:59-!- stikonas [n=stikonas@wesnoth/translator/stikonas] has quit [Read error: 110 (Connection timed out)]
20090406 00:40:10-!- BenUrban [n=benurban@unaffiliated/benurban] has quit ["Power failu"]
20090406 00:40:12< cjhopman> what about moving the check for human into advance_unit?
20090406 00:40:30< cjhopman> i havent actually looked at the code much, so i don't know if thats feasible
20090406 00:40:48< Crab_> alink: correct me if I am wrong. there is [event] name="sighted" on line 19135, it ends on line 19278. so,  line 19209 you're mentioned is inside it.
20090406 00:41:06< Polarina> So many lines...
20090406 00:41:21< alink> cjhopman: yes I thought doing that but that bool is not really just a is_human check, but it's for replay or something
20090406 00:41:29< cjhopman> yeah, i just saw that
20090406 00:42:27< alink> Crab_: ok it was confusing this is why i said
20090406 00:42:33< alink> [00:36:18] <alink> mmh yes sorry the same code is in the sighted event and in the attack event few lines below
20090406 00:42:44< Crab_> aa. sorry, hasn't seen that.
20090406 00:42:54< alink> line 19280
20090406 00:43:16< alink> same code, no sighted and it kill the unit too causing the lost of attack_left
20090406 00:44:03< alink> I think my previous solution make sense, fire the attack event earlier
20090406 00:44:27< alink> before removing attack_left, move etc..
20090406 00:45:01-!- thespaceinvader [n=chatzill@wesnoth/artist/thespaceinvader] has quit ["ChatZilla 0.9.84 [Firefox 3.0.8/2009032609]"]
20090406 00:46:26< Crab_> alink: it may be not what the WML authors wanted: for example, imagine a unit attacking a bad guy, who (via an event) teleports away, and it wastes moves/attacks of that unit. so, by changing the events order, it is theoretically possible to ruin some scripted sequences. 
20090406 00:46:30-!- loonycyborg [n=sergey@wesnoth/developer/loonycyborg] has quit ["Zzzzzzzzzzzzzzzzzzzzzzzzzzzz"]
20090406 00:47:04-!- loonybot [n=loonybot@wesnoth/bot/loonybot] has quit [Remote closed the connection]
20090406 00:47:24-!- BenUrban [n=benurban@c-68-49-10-243.hsd1.md.comcast.net] has joined #wesnoth-dev
20090406 00:47:34< alink> well in that case, this WML author could still manually remove the attack_left and moves of the attacker
20090406 00:48:29< Crab_> yes, he can. but he doesn't know that he needs to do it, thus, bugs may appear :)
20090406 00:49:16< alink> Crab_: btw, obviously I am speaking about changing that in 1.7, breaking crazy WML is allowed there
20090406 00:49:48< alink> I am not sure that we can fix that in 1.6 without the problem you mentionned
20090406 00:50:27< alink> unless someone change the WML of this scenario
20090406 00:51:26< Crab_> where do you want to fire the attack event ?
20090406 00:52:02< alink> maybe just before  line 929  a_.get_unit().set_attacks(a_.get_unit().attacks_left()-1);
20090406 00:52:43< alink> not sure what to do with the   // no attack weapon => stop here and don't attack
20090406 00:52:54< alink> who attack with no weapon ? 
20090406 00:52:58< Crab_> remember, that both units and both attacks modes must be known when firing attack
20090406 00:53:02< Crab_> for filters to work
20090406 00:54:00< Crab_> alink: what about adding a new event type ? "preattack" or something like that
20090406 00:54:44-!- noy [n=Noy@wesnoth/developer/noy] has quit [Connection timed out]
20090406 00:54:56< alink> it's possibly already a FR
20090406 00:54:56< Crab_> and putting it just before line 929..
20090406 00:55:01-!- Blueblaze [n=nick@c-98-199-143-139.hsd1.tx.comcast.net] has quit [Remote closed the connection]
20090406 00:55:11< alink> I remember something about "consider_attack"
20090406 00:55:34-!- Gauteamus [n=chatzill@wesnoth/translator/Gauteamus] has quit ["ChatZilla 0.9.84 [Firefox 3.0.8/2009032609]"]
20090406 00:56:05< alink> wesbot bug 12375
20090406 00:56:05< wesbot> Bug #12375  Assigned to: Alexander van Gessel <ai0867>  Status: Wont Fix  Priority: 5 - Normal
20090406 00:56:08< wesbot> Summary: 2 new events: consider attack + unconsider attack
20090406 00:56:11< wesbot> Original submission: Make an event that triggers just before a unit brings up 
20090406 00:56:14< wesbot> an attack selection screen and then another event that triggers if that screen i
20090406 00:56:17< wesbot> URL: https://gna.org/bugs/?12375
20090406 00:56:20< alink> but that's not exactly what you have in mind
20090406 00:58:34< Crab_> no, that is not what i have in mind.
20090406 00:59:59< Crab_> I am thinking about a preattack event, somewhere at line 921
20090406 01:00:03< alink> I like your "preattack" idea, but the only thing which bother me is that the current "attack" event is already fired before the first hit
20090406 01:00:24< Crab_> but it is fired after attacker has "begin" the attack..
20090406 01:01:04< alink> well in c++ code, yes
20090406 01:01:45< alink> in fact removing the attacks_left before the first hit seems also wrong
20090406 01:01:50< Crab_> basically,  a_.get_unit().set_attacks(a_.get_unit().attacks_left()-1); means that the attack is "expended', that some action is done...
20090406 01:02:25< Crab_> so, IMO, we need another event before decreasing  attacks_left
20090406 01:02:30< alink> ah yes I suppose that you could see it like that
20090406 01:03:28< alink> Crab_: unless you already write WML, i suggest talking about that with WML gurus
20090406 01:04:24< Crab_> so, I'm thinking about something like: select how to attack - event(preattack) - expend "energy" (moves, attacks_left, hidden/moved status for the attacker, resting status for the defender) - begin attacking - event(attack) - actual attacks -...
20090406 01:04:44< Crab_> I'll discuss later. for now, you'll apply your modified patch ?
20090406 01:05:35< alink> Ah yes, i suppose I should do that, since I maybe triggered the bug (by clearing the fog before the attack instead of after)
20090406 01:05:53< Crab_> ok
20090406 01:08:26-!- Blueblaze [n=nick@c-98-199-143-139.hsd1.tx.comcast.net] has joined #wesnoth-dev
20090406 01:09:13< CIA-30> alink * r34572 /branches/1.6/src/mouse_events.cpp: 
20090406 01:09:13< CIA-30> Fix the crash part of bug #13323: Attack plus dialog crashes Wesnoth
20090406 01:09:13< CIA-30> Thanks to Crab to have spotted the bad line
20090406 01:10:29-!- Chusslove [n=caslav@brsg-d9beeee5.pool.mediaWays.net] has left #wesnoth-dev ["Kopete 0.12.7 : http://kopete.kde.org"]
20090406 01:15:03< CIA-30> alink * r34573 /trunk/src/mouse_events.cpp: also port r34572 to trunk (fix crash in bug #13323)
20090406 01:15:43-!- Zen_Clark [n=user@99-136-80-191.lightspeed.rcsntx.sbcglobal.net] has joined #wesnoth-dev
20090406 01:16:09< alink> now, i have no idea what is the status of this bugreport
20090406 01:17:02< alink> "worst part is fixed" ?
20090406 01:19:04< Crab_> alink: maybe, something like : "first part (crash) is fixed, second part (no attacks hit) will be fixed in a new version of the UtBS" )
20090406 01:19:13-!- Polarina [n=polarina@wesnoth/translator/Polarina] has quit ["Leaving."]
20090406 01:19:31< Crab_> and the overall status "in progress" ..
20090406 01:19:31-!- ihsan_ [n=ihsan@88.241.16.16] has joined #wesnoth-dev
20090406 01:19:48< alink> I don't see that long line in the combo choice :)
20090406 01:19:55< Crab_> in progress :)
20090406 01:20:22< Crab_> since we need to talk to WML guru's about the 'right things to do in such a case' )
20090406 01:20:41< alink> ok, I will do that
20090406 01:21:02< Crab_> and change either their WML (btw, since fog is revealed before combat, it will be easier to rely on 'sighted' events to achieve the same behavior...)
20090406 01:21:15< Crab_> either the code. or both.
20090406 01:22:11-!- ihsan_ [n=ihsan@88.241.16.16] has left #wesnoth-dev ["Konversation terminated!"]
20090406 01:23:54< Crab_> alink: another variant is to file a new bug with the second part of this, reference it from this, and mark this as fixed.
20090406 01:24:58< alink> I will just mark it as in progress and assign it to zookeeper, to see if he has an opinion on this WML
20090406 01:25:12< Crab_> alink:  yes, that's simpler and better :)
20090406 01:25:16< alink> we will see what to do after that
20090406 01:26:11< alink> done
20090406 01:26:17< Crab_> good
20090406 01:31:13< alink> Sirp (or Crab_ since you seems to already know this area): this code seems weird/wrong  http://wesnoth.pastebin.com/m7b4172f7
20090406 01:31:31< alink> (the comment is mine)
20090406 01:31:43< alink> Or at least it's very suspicious that the target location is not involved
20090406 01:31:53< Crab_> i'll take a look now
20090406 01:32:31< alink> possibly not important because it seems that most 1-turn actions seems executed before reaching this point
20090406 01:33:17-!- JW1 [n=X@c-71-57-85-168.hsd1.il.comcast.net] has joined #wesnoth-dev
20090406 01:34:08< alink> but i believe that sometimes the code has an effect
20090406 01:34:41< alink> Crab_: thanks
20090406 01:34:44-!- AI0867__ [n=ai@wesnoth/developer/ai0867] has joined #wesnoth-dev
20090406 01:34:56-!- AI0867 [n=ai@wesnoth/developer/ai0867] has quit [Nick collision from services.]
20090406 01:35:02-!- AI0867__ is now known as AI0867
20090406 01:37:34< CIA-30> alink * r34574 /branches/1.6/src/mouse_handler_base.cpp: port r34222 to 1.6 (Fix "next unit" in context menu, bug #13251)
20090406 01:40:28< Crab_> alink:  see line 40 in the same file, for a bunch of commented code
20090406 01:40:45-!- AI0867 [n=ai@wesnoth/developer/ai0867] has quit [Remote closed the connection]
20090406 01:41:34< alink> Crab_: ah yes didn't see that before
20090406 01:41:50< alink> but at least it uses 2 different locations there
20090406 01:41:53< Crab_> yes.
20090406 01:43:20< alink> wesbot log 5343
20090406 01:43:23< wesbot> Sirp * r5343 : added Redsun's pathfinding patch
20090406 01:43:23< wesbot> URL: http://svn.gna.org/viewcvs/wesnoth?view=rev&rev=5343
20090406 01:43:41-!- karl_m [n=karlm@210.11.197.219] has joined #wesnoth-dev
20090406 01:44:16< Crab_> and they filter by destination, since dstsrc is a map of dst->src ...
20090406 01:44:40-!- PK [n=pk@r74-192-44-206.vctrcmta01.vctatx.tl.dh.suddenlink.net] has joined #wesnoth-dev
20090406 01:45:29< Crab_> alink: since multimap equal_range filters using keys, not values
20090406 01:45:54-!- gabm [n=gabriel@132.207.239.127] has left #wesnoth-dev []
20090406 01:46:04< Crab_> so,  dstsrc.equal_range(u->first);  will return "a list of possible moves having my current location as destination"
20090406 01:46:57< alink> yes, this is also how i understood it. But that seems wrong
20090406 01:47:24< Crab_> so that current code just says "all moves which start on my location and end on my location" cost nothing.
20090406 01:47:27-!- AI0867 [n=ai@wesnoth/developer/ai0867] has joined #wesnoth-dev
20090406 01:47:30< Crab_> so, at least it does nothing wrong )
20090406 01:47:58< alink> ok so it's just a misleading comment :  // if this unit can move to that location this turn
20090406 01:48:22< Crab_> need to see more. since there may be the error you've spotted
20090406 01:48:38< Crab_> maybe something else was intended (as in :40 commented out lines)
20090406 01:48:57< alink> yes indeed it's near the move_left mess
20090406 01:49:18< Crab_> for if we change it to  "dstsrc.equal_range(tg->loc);" , then the comment suddenly  becomes true :)
20090406 01:49:46< alink> yes thus my comment in my pastebin
20090406 01:49:49< Crab_> "if we can move to target in 1 turn, set the cost to 0".
20090406 01:50:10-!- PK [n=pk@r74-192-44-206.vctrcmta01.vctatx.tl.dh.suddenlink.net] has quit ["Java user signed off"]
20090406 01:51:20< Crab_> so the current code is not that wrong, it just does nothing... 
20090406 01:51:51< alink> mmh not sure there is always a "stay here" move in dstsrc
20090406 01:52:14< Crab_> and its cost is already 0, isn't it ?
20090406 01:52:14< alink> but the move_left to target may be > 0
20090406 01:52:45< alink> yes but we are currently evaluating a specific move (to a target)
20090406 01:52:59< alink> the fact that we can stay where we are seems irrelevant
20090406 01:54:39< Crab_> lets see if dstsrc contains 'stay here' moves..
20090406 01:55:07-!- Turuk [n=Turuk@wesnoth/forumsith/turuk] has quit [Read error: 60 (Operation timed out)]
20090406 01:55:51< alink> IIRC we manually insert them
20090406 01:56:18-!- Turuk [n=Turuk@wesnoth/forumsith/turuk] has joined #wesnoth-dev
20090406 01:56:38< Crab_> yes, we do
20090406 01:57:25< Crab_> so  dstsrc.equal_range(u->first); will not be empty
20090406 01:58:04< Crab_> and inside that range, once "  if (locRange.first->second == u->first) " will be true. 
20090406 01:58:10< Crab_> in that 'stay here' move
20090406 01:58:22< Crab_> so cur_route.move_left = 0;  will be executed no matter what
20090406 01:58:27< alink> a possible explanation of an error is that the difference between loc and loc_ was not noticed in the initial code 
20090406 01:58:35< Crab_> yes, quite possible
20090406 01:59:00< Crab_> so, current code always resets cur_route.move_left = 0; , if it gets to that point...
20090406 01:59:31< alink> the good news is that it mitigate the move_left bug in 1.6 :)
20090406 01:59:43< Crab_> he-he )))
20090406 02:00:04< alink> but having such useless/bad code there since 2005 is less good
20090406 02:00:27< Crab_> we can replace it with just  "cur_route.move_left = 0;" )
20090406 02:01:11< Crab_> for clarity, without changing any behavior.
20090406 02:01:46< alink> Crab_: you mean to make clear that it's stupid :)
20090406 02:02:03< Crab_> something like this )
20090406 02:02:12< Crab_> and to not introduce new bugs
20090406 02:02:22< Crab_> which may be hidden so far.
20090406 02:02:48< alink> then, better directly replace all the following cur_route.move_left by 0 in code ;)
20090406 02:02:48-!- grantwu3 [n=UserXP@67-194-3-216.wireless.umnet.umich.edu] has joined #wesnoth-dev
20090406 02:03:08< Crab_> I want to create some sort of AI test framework, to test any changes that may cripple the ai..
20090406 02:03:10< alink> std::max<int>(1,0) will really look nice
20090406 02:03:15< Crab_> hehe
20090406 02:03:25< Crab_> and only after then, play with fixing something..
20090406 02:04:08< alink> Crab_: about AI test framework, did you see the ideas of Darth_fool in the dev-ml
20090406 02:04:19< Crab_> date ?
20090406 02:04:48< alink> searching
20090406 02:06:16-!- AI0867 [n=ai@wesnoth/developer/ai0867] has quit [Remote closed the connection]
20090406 02:06:32< Crab_> thanks in advance )
20090406 02:06:44< alink> I see something at 6 mars, but there is more elsewhere
20090406 02:07:30< alink> anyways, he's not active since a long time, so he may stay inactive
20090406 02:07:39< Crab_> ok
20090406 02:08:47< karl_m> alink, Crab_: I have tried to get in touch with Darth_fool for several days, re that test framework idea on the dev ml, but no luck. I would like to work on it with someone
20090406 02:09:00< alink> https://mail.gna.org/public/wesnoth-dev/2009-03/threads.html  , he's John McNabb
20090406 02:09:39< karl_m> or something like it
20090406 02:10:28< alink> karl_m: like I said, he's not active for the moment. He seems motivated about his AI framework, but maybe wait if he commit something one day
20090406 02:10:34< CIA-30> esr * r34575 /trunk/data/ (4 files in 3 dirs): (log message trimmed)
20090406 02:10:34< CIA-30> Fix continuity issues created by Chantal becoming personally important
20090406 02:10:34< CIA-30> to Delfador - happens in Delfador's Memoirs, affects Heir top the Throne.
20090406 02:10:34< CIA-30> This patch bumps Chantal from Druid to Shyde, and makes her death in HttT a
20090406 02:10:34< CIA-30> loss condition - though one that is highly unlikely to trigger given the
20090406 02:10:36< CIA-30> usual flow of battle in the scenario and the hitpoints she has as an L3.
20090406 02:10:38< CIA-30> This patch also adds dialog to both Elves Beseiged and Siege of Elensefar
20090406 02:11:08-!- Elvish_Pillager [n=eli@66-189-40-122.dhcp.oxfr.ma.charter.com] has quit ["Hi! I'm a quit message virus vaccine. If you see a quit message virus, don't replace your quit message with it!"]
20090406 02:11:32< Crab_> karl_m, alink: I'm reading the ml...
20090406 02:12:25< karl_m> I'm an Australian student in this course http://cs.anu.edu.au/student/comp8440/ and I would like to do some work on the AI overin the next couple of months.
20090406 02:12:51< alink> karl_m: are you a GCoC student ?
20090406 02:13:04< karl_m> no, not in GSoC
20090406 02:13:48< alink> good, i was almost tired of all the GSoC student and their hunger for money ;-)
20090406 02:14:31< alink> karl_m: AI is currently a hot topic in wesnoth
20090406 02:16:18< karl_m> alink: yes I gather that from reading the forums and the ml archives
20090406 02:17:04< karl_m> alink: so I want to do some work that doesnt cause confusion etc amidst all the new ideas
20090406 02:17:36< Crab_> karl_m: and I'm a GSoC student ) for my plans, see http://www.wesnoth.org/wiki/SummerOfCodeProposal_AI_Improvement_Crab     If I will  not be accepted into GSoC, I'll scale it down and still try to do it, but it will be slower (I'll devote less time to that)
20090406 02:18:40< alink> karl_m: I am no AI expert, but I suppose that you could try using the Formula AI and find bug, suggest idea, or propose patch
20090406 02:19:39< alink> karl_m: we are always open to good patches/idea/bugreport
20090406 02:20:00< karl_m> alink: yeah that was my initial idea but (a) Sirp wants to start a new AI rather than keep working on formula AI and (b) someone else has done some patches on most of the EasyCoding ideas re formula AI
20090406 02:21:04< Crab_> karl_m: in any case, formula AI, as a domain specific language, will still be actual ) 
20090406 02:21:13< alink> mmh I see, it's true that with all these plan, it's maybe not clear where to start now
20090406 02:21:17< karl_m> alink: in general it seems like 'a new AI' will happen but exactly what is to be worked out over a longer timeline than just the next couple of months. So some sort of testing framework seems like potentially a useful contribution
20090406 02:21:46< Crab_> karl_m: yes, I will try to create some kind of testing framework over the next few weeks
20090406 02:22:17< Crab_> karl_m: to do that, I've already added the ability to hot-redeploy the AI configuration from file
20090406 02:22:43< Crab_> karl_m: and added a small scenario which can be used for 'interfactive testing'. it is very basic at the moment.
20090406 02:23:08-!- ilor [n=user@wesnoth/developer/ilor] has quit []
20090406 02:23:11< karl_m> Crab_: yes the AI_Arena, I have looked briefly at that, it seemed like a very good idea
20090406 02:23:17< Crab_> karl_m: the idea is "1) select the test that you want to do, 2) select the ai that you want to use 3) run it "
20090406 02:24:02< Crab_> karl_m: it is in a very basic shape now. I plan to add more tests to it, and to make testing/writing tests more easier
20090406 02:24:04< karl_m> Crab_: so without stepping on your (potential) GSoC toes, what sort of thing could I do to help?
20090406 02:25:14< Crab_> karl_m: well, you should pick whatever you like best and whatever you are able to do, just keep me and Dragonking (also a GSoC-candidate in the AI topic and also a wesnoth developer) informed :) 
20090406 02:25:27< Crab_> karl_m: it all depends what do you like more - WML or C++ :)
20090406 02:25:50< alink> karl_m: and you can always find bugs in his code, and propose patches to fix them
20090406 02:26:29< Crab_> karl_m: for example, if you look at AI arena, you can propose ideas to make it better, or/and implement them
20090406 02:27:37< karl_m> I would prefer C++ :)  ok I will look at the arena! thanks. any pointers on where to start?
20090406 02:28:13< Crab_> karl_m: checkout wesnoth from public svn, compile it, try current state of AI arena.
20090406 02:28:47< karl_m> yes I have done that. I guess I meant, where in the code to look
20090406 02:29:14< Crab_> well, data/ai/scenarios/scenario-AI_Arena_small.cfg  is the scenario file
20090406 02:29:33-!- grantwu3 [n=UserXP@67-194-3-216.wireless.umnet.umich.edu] has quit [Read error: 110 (Connection timed out)]
20090406 02:29:37< Crab_> you can look through it and see how it works
20090406 02:30:40< Crab_> then, you can test it in wesnoth (running wesnoth with ./wesnoth -t ai_arena_small -d )
20090406 02:30:50< Crab_> -t says "load that test scenario"
20090406 02:30:54< Crab_> -d says "debug mode"
20090406 02:31:27< karl_m> so no -m necessary?
20090406 02:31:41< Crab_> no, if you want to run it interactively
20090406 02:32:31< Crab_> it is not possible at the moment to run ai arena as a batch job, without gui
20090406 02:32:48< Crab_> karl_m: AI, at the moment, consists of "C++ core class" and "various configuration"
20090406 02:33:04< Crab_> karl_m: various configuration - links to formulas, Ai parameters, etc
20090406 02:33:22-!- oak_ [n=oakus@c02--112b.kn.vutbr.cz] has quit [Read error: 110 (Connection timed out)]
20090406 02:33:51< Crab_> there are two "good" C++ core AI classes  - one is in src/ai.cpp   (default ai, commonly called C++ ai), and the second is in src/formula_ai.cpp (commonly called formula AI)
20090406 02:33:52< karl_m> Crab_:  is the batch job thing going to be difficult to get going?
20090406 02:34:21< karl_m> Crab_: yes I have looked at those files briefly. But keep going this is very useful
20090406 02:34:49< Crab_> karl_m: no, not that difficult, but there are some things to do.
20090406 02:34:58< Crab_> launch wesnoth --help
20090406 02:36:03< Crab_> to make 'batch testing' possible, we need to :
20090406 02:36:16< Crab_> 1) launch wesnoth without gui (gui = animation = slow )
20090406 02:36:36< Crab_> 2) select the scenario (map and units on it) that we want to play
20090406 02:36:52< Crab_> 3) select the controllers of the sides (both - AI, not humans)
20090406 02:37:01< Crab_> 4) select Ai algorithms for those sides
20090406 02:37:14< Crab_> the problems are (2) and (4)
20090406 02:37:45< karl_m> so something like -t blah -d  --no-gui --controler1=ai --controller2=ai
20090406 02:37:47< Crab_> because we want "one scenario - multiple possible situations, from which one is chosen"
20090406 02:37:59< Crab_> karl_m: yes, something like this
20090406 02:38:23< karl_m> and then another argument for your problem $4
20090406 02:39:01< Crab_> current code allows something like wesnoth --nogui  --multiplayer --controller1=ai --controller2=ai --algorithm1=ai1 --algorithm2=ai2  --scenario=scenario_name
20090406 02:39:18< Crab_> but the --controllerN  specifies only the C++ core part of the AI
20090406 02:39:52< Crab_> this is 'ok' for default ai (since it good as a  C++ only ai)
20090406 02:40:05< Crab_> but this is bad for register-candidate-move based ai (formula_ai.cpp)
20090406 02:40:23< Crab_> since that AI needs additional information to work.
20090406 02:40:31< karl_m> Crab_: right, and also for Sirp's new AI which totally changes things
20090406 02:40:44< karl_m> AI-to-be I should say
20090406 02:41:19< Crab_> karl_m: Sirp's idea doesn't change much in that regard. There'll still be C++ part, and it will be able to call formulas (formula language, or lua, or whatever will be)
20090406 02:41:45< Crab_> karl_m: so, the string "formula_ai" or "default_ai" in  --controllerN   is not enough to specify the Ai fully
20090406 02:42:05< Crab_> because formula_ai needs formulas to be of any use )
20090406 02:42:17-!- Espreon [n=chatzill@wesnoth/developer/espreon] has joined #wesnoth-dev
20090406 02:42:18< karl_m> Crab: :)
20090406 02:42:36< Crab_> karl_m:  so, to make batch testing possible, we need to specify, for example, 'the path to ai configuration'
20090406 02:43:04< Crab_> karl_m: view data/ai/dev/formula_ai_poisoning.cfg
20090406 02:43:05< karl_m> Crab_: something --ai_config xxx/yyy.cfg
20090406 02:43:10< Crab_> karl_m: exactly
20090406 02:43:39< Crab_> '--ai_config1=xxx/yyy.cfg --ai_config2=xxx/zzz.cfg'
20090406 02:44:10< Crab_> these options must be somehow passed into src/team.cpp
20090406 02:45:04< Crab_> inside of class team, there is structure team_info, and when it is constructed, it calls the ai construction routines
20090406 02:45:21< Crab_> it can easily request the ai_manager to load the ai from file
20090406 02:45:30< karl_m> Crab_: via the config& parameter?
20090406 02:46:39< Crab_> karl_m: team_info gets a config& parameter on construction. this config parameter corresponds to the contents of the [side] WML tag
20090406 02:47:12-!- turin [n=turin@168.215.250.1] has joined #wesnoth-dev
20090406 02:47:40< karl_m> Crab_: ok, that makes sense
20090406 02:48:04< Crab_> karl_m: so, a good task is to make the game to understand --ai_config1=
20090406 02:49:38< karl_m> Crab_: sounds reasonable. any preference about adding something to the config& class vs adding a second parameter to the team_info constructor?
20090406 02:49:54< Crab_> 'second parameter to the team_info constructor' is not good.
20090406 02:50:28< karl_m> Crab_ :) ok then
20090406 02:51:00< Crab_> and then the team_info constructor can just call ai_manager::add_ai_for_team_from_file instead of ai_manager::add_ai_for_team
20090406 02:52:56< karl_m> right, cool
20090406 02:53:55< karl_m> :Crab_: thanks heaps for explaining all that to me
20090406 02:54:21< Crab_> karl_m: ) ask more questions if you want to :)
20090406 02:55:35< karl_m> Crab_: let me just look at this for a bit, I am sure I will have more questions later!
20090406 02:56:21-!- wildpenguin [n=chpln@ppp121-45-157-74.lns11.adl6.internode.on.net] has joined #wesnoth-dev
20090406 02:56:28< Crab_> karl_m: src/game.cpp:714 is the place where command-line parameters are parsed for a MP game
20090406 02:57:06< Crab_> it is a good place to start )
20090406 02:57:33< Crab_> and take a look at wesnoth logging system
20090406 02:57:46< Crab_> full logging is ./wesnoth-debug --log-debug=all
20090406 02:57:54< Crab_> but there are various categories
20090406 02:58:23< Crab_> ./wesnoth --log-debug=ai_manager,ai_configuration   will allow you to see the ai's being created
20090406 02:59:05< karl_m> :Crab_: ahh excellent thank you
20090406 02:59:24< Crab_> in data/ai/ais/ directory, there is a default_ai.cfg and formula_ai.cfg - empty default ai configuration and empty formula ai configuraiton
20090406 02:59:59< karl_m> Crab_: also very helpful!
20090406 03:00:34< Crab_> so --ai_config1=data/ai/ais/default_ai.cfg is expected to load default ai, and --ai_config2=data/ai/ais/formula_ai.cfg is expected to load formula ai (without any formulas :) )
20090406 03:00:48< Crab_> (1,2 - side numbers, of course)
20090406 03:01:03< karl_m> Crab: yes that makes sense
20090406 03:02:55< Crab_> karl_m: if you run ./wesnoth-debug -t --log-debug=ai_manager, you'll se how default Ai initialization looks like in the logs. if you run ./wesnoth-debug -t formula --log-debug=ai_manager then you'll see how formula Ai initialization looks like in the logs
20090406 03:03:13< Crab_> (default test scenario uses default ai, and formula test scenario uses formula ai)
20090406 03:03:53< Crab_> default ai will produce messages like: 20090406 04:00:27 debug ai_manager: [] (not initialized) for team 1 : Preparing new AI holder
20090406 03:04:10< Crab_> and formula ai will produce messages like: 20090406 04:00:48 debug ai_manager: [formula_ai] (not initialized) for team 2 : Preparing new AI holder
20090406 03:04:22< karl_m> Crab_: right
20090406 03:05:33< Crab_> in the logs, in [], there's a name of C++ part of the AI. empty name == default ai. this name is passed to a factory in ai_manager::create_transient_ai 
20090406 03:06:03-!- Zen_Clark [n=user@99-136-80-191.lightspeed.rcsntx.sbcglobal.net] has quit ["Its later then I thought... Need break for food..."]
20090406 03:06:13< Crab_> anything else you would like to know ?
20090406 03:06:59< karl_m> Crab_: that's heaps for now, thanks again. I will look and play and fiddle a bit and see what I can come up with, or what I don't understand :)
20090406 03:07:14< Crab_> ok ) good luck ) 
20090406 03:08:12< karl_m> Crab_: thanks :)
20090406 03:11:50-!- Turuk [n=Turuk@wesnoth/forumsith/turuk] has quit [Read error: 104 (Connection reset by peer)]
20090406 03:12:05-!- Turuk [n=Turuk@wesnoth/forumsith/turuk] has joined #wesnoth-dev
20090406 03:15:39< Crab_> esr: hasn't managed to get that save to "don't load " ) even loaded a few 1.5 DM saves successfully, both in 1.6 and in trunk. Maybe the problem was with DM being somehow incorrectly installed ?
20090406 03:15:48< Crab_> esr:  if the problem is still present, I'd run wesnoth with --log-debug=filesystem and see, where it searches for Mage Magister description
20090406 03:31:42-!- Crab_ [n=Crab_@wesnoth/developer/crab] has quit ["Leaving."]
20090406 03:33:41< esr> Crab: My guess is that it had something to do with fendrin's recent @-path patch.
20090406 03:34:16< esr> I've restarted the campaign and don't seem to be having trouble reloading custom units now.
20090406 03:35:34< dfranke> corn: I'm almost done refactoring the binary-invocation code, btw.
20090406 04:05:41< CIA-30> dfranke * r34576 /trunk/src/ (5 files): Refactor the logic for invoking wesnoth helper-programs, e.g. wesnothd
20090406 04:05:47< dfranke> corn: ^^^
20090406 04:07:10< dfranke> corn: now you can just call get_program_invocation("wesupload").
20090406 04:10:22-!- Ivanovic_ [n=ivanovic@dtmd-4db2ba9c.pool.einsundeins.de] has joined #wesnoth-dev
20090406 04:11:34-!- Espreon [n=chatzill@wesnoth/developer/espreon] has quit ["ChatZilla 0.9.84 [Firefox 3.0.8/2009032609]"]
20090406 04:23:49-!- alink [n=alink@wesnoth/developer/alink] has quit [Remote closed the connection]
20090406 04:26:23-!- Ivanovic [n=ivanovic@wesnoth/developer/ivanovic] has quit [Read error: 110 (Connection timed out)]
20090406 04:28:20-!- Ivanovic_ is now known as Ivanovic
20090406 04:29:51-!- noy [n=Noy@d75-157-52-251.bchsia.telus.net] has joined #wesnoth-dev
20090406 04:36:33-!- Crab_ [n=Crab_@wesnoth/developer/crab] has joined #wesnoth-dev
20090406 04:38:56< Crab_> esr: around ? question about your FR 13321 (command to level-up units). How this level-up command should be named ? It is easiest to add this to unit command, to make it look like  "unit levelup" or "unit addlevels=3" or "unit promote", etc.
20090406 04:39:36< Crab_> esr: another solution is to make "experience=N" command auto-check for levelup
20090406 04:39:45< Crab_> and open the levelup dialog
20090406 04:40:33< esr> Crab_: True.  But that way the person doing the debuggiong would have to keep track of the next experience threshold, which would be kind of a pain.
20090406 04:40:53< Crab_> esr: so,  "unit levelup" or "unit addlevels=N" ?
20090406 04:41:08< esr> Don't care much.  
20090406 04:42:08< esr> I guess I might like "unit advance" since we tend to use that term heavily in the ganme help and manual.
20090406 04:42:09< CIA-30> dfranke * r34577 / (2 files in 2 dirs): Fixed non-sequitir reference to Gweddry in a comment in HttT. Noticed while looking at r34575.
20090406 04:42:40< Crab_> "unit advance" or "unit advances=N" or both ?
20090406 04:43:19< esr> For my purposes "unit advance" would be enough.  Whatever is easiest to implement.
20090406 04:43:59< Crab_> esr: both are easy, "unit advances=N" is less lines.
20090406 04:44:11< Crab_> so, I'll add "unit advances=N"
20090406 04:44:16< esr> OK.  Whichever works.
20090406 04:45:05< dfranke> esr: are you planning on mentioning Chantal in more places in HttT?  Right now the comment you added about "We cannot go to Wesmere" seems a bit out of place.
20090406 04:46:13< esr> dfranke: Wasn't really planning on bringing in Chantal. further.  You knbow she originated as an NPC in HttT scenario 1, right?
20090406 04:46:34< dfranke> esr: yes
20090406 04:47:14< esr> And, out of place how?  The player gas just seen her speaking, and being asked to gfleee to the Northern Elves, in the previous scenario.
20090406 04:47:32< dfranke> I'm talking about the reference in Siege of Elsenefar.
20090406 04:47:34< dfranke> ""We cannot go to Wesmere, for Asheviere's orcs have the approaches ringed about with steel; Kalenz and I barely escaped, and Chantal cannot get out. Until we are stronger, we must go where the orcs are not."
20090406 04:47:50< esr> So, er, what's the problem?
20090406 04:47:54< dfranke> I can't figure out why you added that line since nobody ever mentions Wesmere elsewhere in that scenario.
20090406 04:48:29< esr> Oh.  It's because Wesmere is where the Ka'lian meets -- that's well-established canon.
20090406 04:49:07< esr> I added that line to explain, post-DM, why Chantal isn't coming out to play with Delf again.
20090406 04:49:50< dfranke> ah, ok.
20090406 04:50:04 * Crab_ just tried unit advances=100    got a 359-hp Delfador :)
20090406 04:55:48< CIA-30> crab * r34578 /trunk/ (changelog src/menu_events.cpp): 
20090406 04:55:48< CIA-30> Implemented FR #13321: added "unit advances=N" command to debug console
20090406 04:55:48< CIA-30> to immediately level-up active unit N times
20090406 04:55:50< Crab_> esr: ^^
20090406 04:56:17< esr> Thanks. I'll mark it Fixed.
20090406 04:56:38< Crab_> ok
20090406 04:57:25-!- Crab_ [n=Crab_@wesnoth/developer/crab] has quit ["Leaving."]
20090406 05:16:52-!- Turuk_ [n=Turuk@wesnoth/forumsith/turuk] has joined #wesnoth-dev
20090406 05:23:23< CIA-30> dfranke * r34579 / (2 files in 2 dirs): UtBS: Add some [allow_undo]s, fixing bug #13326
20090406 05:25:14-!- Turuk [n=Turuk@wesnoth/forumsith/turuk] has quit [Read error: 110 (Connection timed out)]
20090406 05:27:50< JW1> hello all.
20090406 05:28:31< JW1> how do I go about getting the Era of Myths units on units.wesnoth.org for 1.6?
20090406 05:48:49-!- Blueblaze [n=nick@c-98-199-143-139.hsd1.tx.comcast.net] has quit [Remote closed the connection]
20090406 05:49:29-!- [Relic] [n=[Relic]@adsl-76-229-202-137.dsl.milwwi.sbcglobal.net] has joined #wesnoth-dev
20090406 05:50:44< [Relic]> Hello :)
20090406 06:09:12-!- wesbot changed the topic of #wesnoth-dev to: 1.6.1 planned for Sunday, April 12th, stringfreeze for branches/1.6 ! | Summer of Code info: http://www.wesnoth.org/wiki/SummerOfCodeIdeas | 51 bugs, 230 feature requests, 11 patches | logs: http://irclogs.wesnoth.org/ | Don't paste on IRC! Use a pastebin: http://wesnoth.pastebin.com | http://imagebin.org
20090406 06:31:45-!- crimson_penguin [n=ben@wesnoth/developer/crimsonpenguin] has quit ["Leaving"]
20090406 06:32:24-!- Turuk_ [n=Turuk@wesnoth/forumsith/turuk] has quit [Success]
20090406 06:32:39-!- Turuk [n=Turuk@wesnoth/forumsith/turuk] has joined #wesnoth-dev
20090406 06:37:53< dfranke> Soliton: have you looked at bug #13238?  I've encountered this a couple times and strongly suspect r32713 is to blame.
20090406 06:53:01-!- JW1 [n=X@c-71-57-85-168.hsd1.il.comcast.net] has quit []
20090406 07:05:59-!- ilor [n=user@wesnoth/developer/ilor] has joined #wesnoth-dev
20090406 07:07:39-!- xonev [n=chatzill@59.92.74.6] has joined #wesnoth-dev
20090406 07:22:42-!- Morian_ [i=romain@bezut.info] has joined #wesnoth-dev
20090406 07:26:04-!- karl_m [n=karlm@210.11.197.219] has quit ["Leaving."]
20090406 07:35:56-!- Morian [i=romain@bezut.info] has quit [Read error: 110 (Connection timed out)]
20090406 07:42:28-!- Sirp [n=me@wesnoth/developer/dave] has quit ["leaving"]
20090406 07:45:16-!- ilor [n=user@wesnoth/developer/ilor] has quit []
20090406 07:47:16-!- turin [n=turin@168.215.250.1] has quit [Read error: 110 (Connection timed out)]
20090406 07:48:30-!- silene [n=plouf@ASte-Genev-Bois-152-1-55-154.w82-121.abo.wanadoo.fr] has joined #wesnoth-dev
20090406 07:52:12< silene> hi
20090406 08:02:18-!- [Relic] [n=[Relic]@adsl-76-229-202-137.dsl.milwwi.sbcglobal.net] has quit ["Leaving"]
20090406 08:18:00< dfranke> gah... the SDL mutex API sucks.
20090406 08:18:08< dfranke> it doesn't even have a try_lock function.
20090406 08:19:05< dfranke> oh, aha... but it has a separate semaphore API.
20090406 08:19:10< dfranke> which does.
20090406 08:22:32< corn> dfranke: thanks for the function. I'll incorporate it into my patch and get rid of the hardcoded path today :)
20090406 08:24:31< dfranke> corn: k, cool.
20090406 08:24:46< dfranke> I hate working with concurrency at this level of abstraction.
20090406 08:24:55< corn> http://cornmander.com:8080/breakdown/
20090406 08:24:57< dfranke> Why can't Wesnoth be written in Haskell so I could use STM? :-)
20090406 08:25:18< corn> http://cornmander.com:8080/breakdown/?campaigns=TUTORIAL
20090406 08:25:24< corn> 31% of tutorial games end in defeat
20090406 08:25:39< dfranke> that's... depressing.
20090406 08:25:50< corn> lol
20090406 08:26:12< dfranke> actually, losing the second scenario is excusable.
20090406 08:26:45< corn> wait till I get a view for the time and gold stuff up
20090406 08:27:05< corn> http://cornmander.com:8080/gold/
20090406 08:27:11< corn> gold is currently mostly broken
20090406 08:27:14< corn> ranges are all wrong
20090406 09:08:47-!- YogiHH [i=d4ca9d15@wesnoth/developer/yogihh] has joined #wesnoth-dev
20090406 09:09:01< YogiHH> hello everyone
20090406 09:15:10< cjhopman> hello
20090406 09:16:13-!- grzywacz [n=grzywacz@wesnoth/developer/grzywacz] has quit [Remote closed the connection]
20090406 09:36:10-!- fabi [n=fabi@wesnoth/developer/fendrin] has quit [Remote closed the connection]
20090406 09:45:51-!- silene [n=plouf@wesnoth/developer/silene] has quit ["Leaving."]
20090406 09:47:16-!- Turuk_ [n=Turuk@FUSE-WS1-208-102-202-188.fuse.net] has joined #wesnoth-dev
20090406 09:54:50< dfranke> has anyone else recently seen dirty build issues with scons?
20090406 09:55:14< dfranke> I made a change to play_controller.cpp and it's not recompiling it.
20090406 09:55:44-!- Turuk [n=Turuk@wesnoth/forumsith/turuk] has quit [Read error: 110 (Connection timed out)]
20090406 10:05:42-!- noy [n=Noy@wesnoth/developer/noy] has quit []
20090406 10:06:11-!- euschn [n=chatzill@tmp9.logic.tuwien.ac.at] has joined #wesnoth-dev
20090406 10:06:22< euschn> good morning
20090406 10:09:07-!- loonybot [n=loonybot@79.139.246.137] has joined #wesnoth-dev
20090406 10:09:45-!- loonycyborg [n=sergey@79.139.246.137] has joined #wesnoth-dev
20090406 10:12:37< YogiHH> hi euschn
20090406 10:14:46-!- noy [n=Noy@d75-157-52-251.bchsia.telus.net] has joined #wesnoth-dev
20090406 10:19:55-!- oak_ [i=oakus@c02--112b.kn.vutbr.cz] has joined #wesnoth-dev
20090406 10:21:23-!- henning__ [n=jp0186@77-56-56-38.dclient.hispeed.ch] has joined #wesnoth-dev
20090406 10:21:46< dfranke> nevermind wrt scons... PEBKAC.  I was editing 1.6 files and trying to compile trunk.
20090406 10:31:25-!- noy [n=Noy@wesnoth/developer/noy] has quit []
20090406 10:34:20-!- Baufo [n=thomas@wesnoth/developer/baufo] has joined #wesnoth-dev
20090406 10:34:44< Baufo> morning
20090406 10:53:23< CIA-30> dfranke * r34580 /trunk/src/ (4 files): 
20090406 10:53:23< CIA-30> Implement a new class, save_blocker, analogous to input_blocker and
20090406 10:53:23< CIA-30> update_locker, which effectively masks requests to save the game and
20090406 10:53:23< CIA-30> processes them after the save_blocker has been destructed. Wrap one
20090406 10:53:23< CIA-30> around playsingle_controller::init_side(), fixing bug #9496. See also
20090406 10:53:23< CIA-30> bug #13298.
20090406 10:54:44< CIA-30> ivanovic * r34581 /branches/1.6/po/ (wesnoth/de.po wesnoth-httt/de.po): updated German translation (fixed bug #13328 and #13327 )
20090406 10:54:54< CIA-30> ivanovic * r34582 /trunk/po/ (wesnoth/de.po wesnoth-httt/de.po): updated German translation (fixed bug #13328 and #13327 )
20090406 10:55:44-!- busfahrer [n=busfahre@unixboard/user/busfahrer] has joined #wesnoth-dev
20090406 11:01:04< Ivanovic> moin
20090406 11:02:46< Smar> mion.
20090406 11:02:46< YogiHH> Ivanovic: moin moin :-)
20090406 11:04:25-!- Crab_ [i=crab@wesnoth/developer/crab] has joined #wesnoth-dev
20090406 11:04:27-!- Crab_ [i=crab@wesnoth/developer/crab] has left #wesnoth-dev []
20090406 11:04:39-!- Crab_ [i=crab@wesnoth/developer/crab] has joined #wesnoth-dev
20090406 11:05:03-!- Crab_ [i=crab@wesnoth/developer/crab] has quit [Client Quit]
20090406 11:09:24< CIA-30> ivanovic * r34583 /trunk/po/ (6 files in 6 dirs): updated Icelandic translation
20090406 11:09:27< CIA-30> ivanovic * r34584 /branches/1.6/po/ (5 files in 5 dirs): updated Icelandic translation
20090406 11:13:44-!- xonev_ [n=chatzill@59.92.10.208] has joined #wesnoth-dev
20090406 11:16:15-!- Polarina [n=polarina@wesnoth/translator/Polarina] has joined #wesnoth-dev
20090406 11:26:18-!- zookeeper [n=l@wesnoth/developer/zookeeper] has joined #wesnoth-dev
20090406 11:31:26-!- xonev [n=chatzill@59.92.74.6] has quit [Read error: 110 (Connection timed out)]
20090406 11:35:27-!- busfahrer [n=busfahre@unixboard/user/busfahrer] has quit ["leaving"]
20090406 11:36:04< dfranke> heh... funny side-effect of my save_blocker patch: If you mash Ctrl+Space, Ctrl+S, Ctrl+Q really fast in that order, then it'll end the turn, then prompt you to quit, *then* prompt you to save, and then correctly save prior to quitting.
20090406 11:38:04-!- Baufo [n=thomas@wesnoth/developer/baufo] has quit [Remote closed the connection]
20090406 11:40:20-!- xonev [n=chatzill@59.92.84.81] has joined #wesnoth-dev
20090406 11:55:39-!- molgrum [n=molgrum@c83-249-53-163.bredband.comhem.se] has joined #wesnoth-dev
20090406 11:56:35-!- xonev_ [n=chatzill@59.92.10.208] has quit [Read error: 110 (Connection timed out)]
20090406 12:01:21-!- henning__ [n=jp0186@77-56-56-38.dclient.hispeed.ch] has quit []
20090406 12:06:10-!- xonev [n=chatzill@59.92.84.81] has quit [Read error: 110 (Connection timed out)]
20090406 12:30:49-!- Turuk_ [n=Turuk@wesnoth/forumsith/turuk] has quit [Read error: 60 (Operation timed out)]
20090406 12:31:28-!- Turuk [n=Turuk@wesnoth/forumsith/turuk] has joined #wesnoth-dev
20090406 12:35:24-!- res|laptop [n=res@pdpc/supporter/student/res2k] has joined #wesnoth-dev
20090406 12:37:19< res|laptop> Ivanovic: that nice fellow wesbot just said you were looking for me?
20090406 12:37:50< Ivanovic> sure, all the application stuff
20090406 12:38:00< Ivanovic> i think i/we left messages in your application
20090406 12:39:51< res|laptop> k
20090406 12:40:16-!- henning__ [n=henning_@ini-firewall.unizh.ch] has joined #wesnoth-dev
20090406 12:44:13-!- Gauteamus [n=chatzill@c43C4BF51.dhcp.bluecom.no] has joined #wesnoth-dev
20090406 12:47:14-!- mordante [i=57d7c91a@gateway/web/ajax/mibbit.com/x-dbfc48de10c10e72] has joined #wesnoth-dev
20090406 12:47:28< mordante> hi, only here for a minute
20090406 12:48:03< mordante> alink the limit of 81 is the auto level limit when unstoring
20090406 12:48:50< mordante> this in order to avoid a long delay if some WML author gives a huge amount of XP which causes the unit to level a lot of times
20090406 12:49:22-!- mordante [i=57d7c91a@gateway/web/ajax/mibbit.com/x-dbfc48de10c10e72] has quit [Client Quit]
20090406 13:17:04-!- Gauteamus [n=chatzill@wesnoth/translator/Gauteamus] has quit ["ChatZilla 0.9.84 [Firefox 3.0.8/2009032609]"]
20090406 13:44:48-!- Turuk [n=Turuk@wesnoth/forumsith/turuk] has quit [Read error: 60 (Operation timed out)]
20090406 13:45:51-!- Turuk [n=Turuk@wesnoth/forumsith/turuk] has joined #wesnoth-dev
20090406 13:59:22-!- molgrum [n=molgrum@c83-249-53-163.bredband.comhem.se] has quit [Read error: 104 (Connection reset by peer)]
20090406 13:59:25-!- molgrum [n=molgrum@c83-249-53-163.bredband.comhem.se] has joined #wesnoth-dev
20090406 13:59:48-!- voris [n=voris@c-24-16-107-149.hsd1.wa.comcast.net] has quit [Remote closed the connection]
20090406 14:05:50-!- KingofCamelot [n=davidsan@ip98-182-16-78.sb.sd.cox.net] has joined #wesnoth-dev
20090406 14:13:03 * Polarina is working on useless stuff. :)
20090406 14:13:59< YogiHH> Polarina: Better skip it and do something for wesnoth, then ;-)
20090406 14:14:47< Polarina> That useless stuff might be useful for Wesnoth, and most likely too. The intended audience is people who work with music, audio and sound effects.
20090406 14:15:16< Polarina> For example: this script http://rafb.net/p/b8hn5v37.html would produce the following output http://simnet.is/gabrielp/output.ogg
20090406 14:16:18< YogiHH> i see
20090406 14:16:57< Shadow_Master> esr: pling
20090406 14:16:59< Polarina> The idea is to have something similar to SVG, so you can "raster" it at any sampling rate without loss of quality.
20090406 14:18:30-!- IneQuation [i=1286@2001:470:1f09:409:0:0:0:1286] has joined #wesnoth-dev
20090406 14:18:32< Polarina> YogiHH: Useless, right?
20090406 14:27:52< Ivanovic> an interesting read for our SoC students: http://www.overcomingbias.com/2007/09/planning-fallac.html
20090406 14:28:25< Ivanovic> that is: dfranke posted it to the google mentor list and wanted to post it to wesnoth-dev@gna, too, though he screwed up the email addy for the wesnoth-dev ml...
20090406 14:28:28< Ivanovic> ;)
20090406 14:29:08< Ivanovic> that is: yeah, we want a detailed plan, too
20090406 14:29:28< Ivanovic> but also have a look at it with such a very rough planning to get an idea if your timelines make sense
20090406 14:29:49< Ivanovic> (basically this is what we use to determin if your timeline is possible or not, some experience on "shit won't work as plenned" scenarios)
20090406 15:02:37-!- Crab_ [n=Crab_@wesnoth/developer/crab] has joined #wesnoth-dev
20090406 15:02:41< Crab_> hi
20090406 15:14:50-!- happygrue [n=George@wesnoth/developer/wintermute] has joined #wesnoth-dev
20090406 15:14:55-!- KingofCamelot [n=davidsan@ip98-182-16-78.sb.sd.cox.net] has left #wesnoth-dev []
20090406 15:16:21-!- mjs-de [n=mjs-de@vpw.wh.uni-dortmund.de] has joined #wesnoth-dev
20090406 15:18:01-!- xonev [n=chatzill@59.92.66.48] has joined #wesnoth-dev
20090406 15:29:54-!- ilor [n=user@wesnoth/developer/ilor] has joined #wesnoth-dev
20090406 15:32:15-!- Reisiger [n=Reisiger@adsl-84-226-135-18.adslplus.ch] has joined #wesnoth-dev
20090406 15:32:17-!- ABCD [n=ABCD@wikipedia/ABCD] has quit [Read error: 104 (Connection reset by peer)]
20090406 15:32:34-!- ABCD [n=ABCD@wikipedia/ABCD] has joined #wesnoth-dev
20090406 15:33:09< Reisiger> Good afternoon everyone.
20090406 15:34:26< Polarina> Good morning.
20090406 15:40:33-!- BenUrban [n=benurban@unaffiliated/benurban] has quit ["Power failu"]
20090406 15:43:31-!- elias [n=allefant@allegro/developer/allefant] has joined #wesnoth-dev
20090406 15:55:54-!- euschn [n=chatzill@tmp9.logic.tuwien.ac.at] has quit [Read error: 110 (Connection timed out)]
20090406 15:59:05-!- ikarius [n=ross@216.27.182.3] has quit []
20090406 16:01:48-!- AndreLuiz [n=AndreLui@unaffiliated/andreluiz] has joined #wesnoth-dev
20090406 16:26:45-!- IneQuation [i=1286@2001:470:1f09:409:0:0:0:1286] has quit [Read error: 110 (Connection timed out)]
20090406 16:31:35-!- wildpenguin [n=chpln@ppp121-45-157-74.lns11.adl6.internode.on.net] has quit ["Leaving"]
20090406 16:31:50-!- molgrum [n=molgrum@c83-249-53-163.bredband.comhem.se] has quit ["Lämnar"]
20090406 16:32:08-!- molgrum [n=molgrum@c83-249-53-163.bredband.comhem.se] has joined #wesnoth-dev
20090406 16:32:32-!- crimson_p [n=irchon@64.201.60.216] has joined #wesnoth-dev
20090406 16:34:24-!- crimson_p [n=irchon@64.201.60.216] has quit [Remote closed the connection]
20090406 16:44:35< CIA-30> crab * r34585 /trunk/ (changelog src/game_events.cpp src/team.cpp src/team.hpp): 
20090406 16:44:35< CIA-30> Added ability to change the share_view side attribute using [modify_side] tag.
20090406 16:44:35< CIA-30> Be sure to use fog=yes for that side and have it as an ally (bug #13320)
20090406 16:45:30< Crab_> esr: ^^, and see http://wesnoth.pastebin.com/m140b11bd
20090406 16:53:16-!- ikarius [n=ross@smtp.gridironsystems.com] has joined #wesnoth-dev
20090406 16:54:17-!- Noyga [n=noyga@wesnoth/developer/noyga] has joined #wesnoth-dev
20090406 16:56:49-!- busfahrer [n=busfahre@unixboard/user/busfahrer] has joined #wesnoth-dev
20090406 16:57:59< Crab_> esr: and you will also need to change the mages wake-up a bit (since they will be already 'sighted' when the view is first shared, so their 'sighted' event may not fire)
20090406 17:00:52-!- ABCD [n=ABCD@wikipedia/ABCD] has quit [Read error: 104 (Connection reset by peer)]
20090406 17:01:01-!- ABCD [n=ABCD@wikipedia/ABCD] has joined #wesnoth-dev
20090406 17:03:43-!- Noyga [n=noyga@wesnoth/developer/noyga] has left #wesnoth-dev ["Quitte"]
20090406 17:05:44< Reisiger> ilor: You there?
20090406 17:11:44-!- grantwu [n=Zenneth@westquad-147025.reshall.umich.edu] has joined #wesnoth-dev
20090406 17:18:37-!- Blueblaze [n=nick@c-98-199-143-139.hsd1.tx.comcast.net] has joined #wesnoth-dev
20090406 17:21:10-!- henning__ [n=henning_@ini-firewall.unizh.ch] has quit []
20090406 17:21:50-!- crimson_penguin [n=ben@64.201.60.216] has joined #wesnoth-dev
20090406 17:23:03-!- grantwu [n=Zenneth@westquad-147025.reshall.umich.edu] has left #wesnoth-dev []
20090406 17:27:05-!- Noyga [n=lame-z@wesnoth/developer/noyga] has joined #wesnoth-dev
20090406 17:27:22-!- Noyga [n=lame-z@wesnoth/developer/noyga] has left #wesnoth-dev ["Quitte"]
20090406 17:29:48-!- gabm [n=gabriel@132.207.234.109] has joined #wesnoth-dev
20090406 17:30:08-!- fendrin [n=fabi@f051016201.adsl.alicedsl.de] has joined #wesnoth-dev
20090406 17:30:43-!- YogiHH [i=d4ca9d15@wesnoth/developer/yogihh] has left #wesnoth-dev []
20090406 17:43:55-!- cjhopman [n=chris@wesnoth/developer/cjhopman] has quit ["Konversation terminated!"]
20090406 17:49:36-!- cjhopman [n=chris@wesnoth/developer/cjhopman] has joined #wesnoth-dev
20090406 17:50:39-!- AI0867_ is now known as AI0867
20090406 17:52:07-!- turin [n=turin@168.215.250.1] has joined #wesnoth-dev
20090406 17:52:55-!- thespaceinvader [n=chatzill@wesnoth/artist/thespaceinvader] has joined #wesnoth-dev
20090406 17:57:23-!- Turuk [n=Turuk@wesnoth/forumsith/turuk] has quit [Connection reset by peer]
20090406 17:57:38-!- Turuk [n=Turuk@wesnoth/forumsith/turuk] has joined #wesnoth-dev
20090406 18:02:23-!- Velory [n=GProcess@88.227.112.18] has joined #wesnoth-dev
20090406 18:15:46-!- noy [n=Noy@d75-157-52-251.bchsia.telus.net] has joined #wesnoth-dev
20090406 18:17:32-!- ABCD [n=ABCD@wikipedia/ABCD] has quit ["No Ping reply in 30 seconds."]
20090406 18:17:52-!- ABCD [n=ABCD@wikipedia/ABCD] has joined #wesnoth-dev
20090406 18:22:13< CIA-30> ivanovic * r34586 /trunk/po/ (5 files in 5 dirs): updated Russian translation
20090406 18:22:18< CIA-30> ivanovic * r34587 /branches/1.6/po/ (5 files in 5 dirs): updated Russian translation
20090406 18:23:09-!- Turuk [n=Turuk@wesnoth/forumsith/turuk] has quit [Read error: 104 (Connection reset by peer)]
20090406 18:23:24-!- Turuk [n=Turuk@wesnoth/forumsith/turuk] has joined #wesnoth-dev
20090406 18:26:25-!- Sutnarova_Lucie [n=chatzill@247.79.broadband5.iol.cz] has joined #wesnoth-dev
20090406 18:26:38-!- Sutnarova_Lucie is now known as Tokiko
20090406 18:26:54-!- Tokiko is now known as Lairaku
20090406 18:32:12-!- busfahrer [n=busfahre@unixboard/user/busfahrer] has quit ["leaving"]
20090406 18:40:06-!- Gnutoo [n=gnutoo@host155-29-dynamic.17-79-r.retail.telecomitalia.it] has joined #wesnoth-dev
20090406 18:40:41< Gnutoo> helo, is there an easy way to remove all assertion about the screen size? because wesnoth crash because of that...last version(1.4.7) worked better
20090406 18:40:45< Gnutoo> and didn't crash
20090406 18:40:56< Gnutoo> and I've a 320x240 screen on armv6
20090406 18:41:40-!- crimson_penguin [n=ben@wesnoth/developer/crimsonpenguin] has quit [Success]
20090406 18:47:09< zookeeper> Gnutoo, i suggest waiting for mordante
20090406 18:47:55< Gnutoo> ok thanks
20090406 18:55:05< Gnutoo> zookeeper, is he usually online during the week? if so is he used to come the afternoon/night
20090406 19:06:51< zookeeper> he is, maybe check irclogs.wesnoth.org for the times..
20090406 19:08:29< Ivanovic> Gnutoo: most likely: no
20090406 19:08:48< Ivanovic> that is: the cases where it now asserts were cases of "undefined behaviour" before
20090406 19:09:57< Gnutoo> ah so how can it be fixed? I've tinygui and CMakeCache.txt says //Set for GUI reductions for resolutions down to 320x240 (PDAs)
20090406 19:10:18< freim> zookeeper: I have two more snow mountains I will commit soon, can you do the wml?
20090406 19:10:48< freim> zookeeper: they will most likely need a tweak on the probability so they don't show up to often since the ridges are missing yet
20090406 19:10:52< Ivanovic> Gnutoo: basically the "problems" have to be fixed by hand each
20090406 19:11:03< Ivanovic> that is noone checked if 320x240 is working anymore *at all*
20090406 19:11:27< Gnutoo> I checked it...last release(1.4.7)
20090406 19:11:46< Gnutoo> I also use 480x640
20090406 19:11:57< Ivanovic> (when i did some work on the theme i made it work nicely for 400x300 as min res and just left the crude stuff as it was in 320x240, that was shortly before 1.4.x
20090406 19:12:15< Ivanovic> but over the last months *noone* tested how tinygui really worked with 1.5.x
20090406 19:12:33< Gnutoo> but I've to write a part in arm assembly to make it compile on the device with 480x640 (armv4)
20090406 19:12:54< Ivanovic> 480x640 should work nicely with tinygui
20090406 19:12:56-!- Turuk [n=Turuk@wesnoth/forumsith/turuk] has quit [Read error: 104 (Connection reset by peer)]
20090406 19:12:57< Shadow_Master> loonycyborg: what are the steps to create  a new build configuration for scons?
20090406 19:13:03< zookeeper> freim, sure, i think the WML is already there, just commented out
20090406 19:13:10-!- Turuk [n=Turuk@wesnoth/forumsith/turuk] has joined #wesnoth-dev
20090406 19:13:12< Gnutoo> Ivanovic, ok so it'w worth writing this assembly part
20090406 19:13:39< zookeeper> freim, so just commit the images whenever you want
20090406 19:13:39< Shadow_Master> loonycyborg: that is  wan tot be able to build a tinygui binary _aside_ from my release, debug, glibcxx-debug and profiled builds, preferably without resorting to cmake or autfoo
20090406 19:13:43< Shadow_Master> autofoo
20090406 19:14:17< Ivanovic> Gnutoo: the question is: why do you have to write a part in assembler?
20090406 19:14:19< freim> zookeeper: ok
20090406 19:15:06< Gnutoo> Ivanovic, because of boost...but I've just had a mail on how to do it because there was a part I didn't understand
20090406 19:16:46< Gnutoo> because of the lack of __sync_fetch_and_add function
20090406 19:17:21-!- boucman [n=rosen@wesnoth/developer/boucman] has joined #wesnoth-dev
20090406 19:17:52< Ivanovic> ah, okay
20090406 19:18:14< boucman> hey all
20090406 19:18:21< Crab_> hi boucman
20090406 19:19:12-!- crimson_penguin [n=ben@64.201.60.216] has joined #wesnoth-dev
20090406 19:25:49< CIA-30> crab * r34588 /trunk/ (changelog src/ai.cpp src/formula_ai.cpp): 
20090406 19:25:49< CIA-30> Made default AI fallback to formula recruitment, if "recruitment"
20090406 19:25:49< CIA-30> config option is set in AI config.
20090406 19:25:49< CIA-30> Per IRC conversation with Dragonking.
20090406 19:30:49-!- Velory [n=GProcess@88.227.112.18] has quit [Read error: 110 (Connection timed out)]
20090406 19:37:01-!- cjhopman [n=chris@wesnoth/developer/cjhopman] has quit ["Konversation terminated!"]
20090406 19:41:13< boucman> Crab_: about your last commit... could you sum the discussion you had with DK, 
20090406 19:41:40< boucman> I don't really understand why adding an AI config option instead of simply including a formula= calling DK's script
20090406 19:42:00< freim> map editor no longer a seperat executable?
20090406 19:42:20< zookeeper> freim, that's right. "map editor" in the title screen.
20090406 19:43:00< loonycyborg> Shadow_Master: SConstruct line 374.
20090406 19:43:42< Crab_> boucman: a long time ago, there was an idea to replace just one part of default AI with formula. namely, recruiting. So, Dragonking has written a recruitment.fai, and made formula_ai to run it, given a team_formula.
20090406 19:43:46< Crab_>  It required a parameter, called "recruit", to work (not that 'recruit' in [side], but another, undocumented 'recruit' in [ai]).
20090406 19:44:41< loonycyborg> Shadow_Master: Also, line 40 must be updated and another extra_flags_* option added.
20090406 19:44:42< Crab_> then I came, and changed AI lifecycle management, extracting it all to a single class. and I've spotted a number of places where formula_ai called default_ai, and default_ai called formula_ai
20090406 19:45:30< Crab_> boucman: so, I firstly committed those places as-is (you've even asked once 'am I sure?' when you've reviewed that my big  patch)
20090406 19:46:03< Crab_> boucman: and then, I asked: how it should really work?
20090406 19:47:00< Crab_> boucman: and, we've come to understanding that there should be a way, in the current code, to run default_ai, BUT with formula_ai recruitment phase.
20090406 19:47:10< ilor> Reisiger: here
20090406 19:47:46< boucman> wouldn't simply specifying recruiting.fai in the formula= part work ?
20090406 19:47:51< Crab_> boucman: so, I've reworked the code to create an instance of formula ai only when it is necessary (when that specific parameter, "recruitment", is set), and then cache it for the lifetime of default ai
20090406 19:48:02< Crab_> boucman: only with ai_algorithm=formula_ai
20090406 19:49:13< Crab_> boucman: and I've asked to rename that config  parameter (in [ai]) "recruit" to "recruitment', to avoid confusion (for there is already "recruit" in side).
20090406 19:49:21< boucman> I'm still not sure why you want to do a new special case for "recruitement only" instead of making a FAI which would fallback to c++ except for recruitement, it would avoid having a new AI parameter for a particular case
20090406 19:52:11< boucman> Crab_: could you just point me to the time of your discussion, I'll look in irclog
20090406 19:53:22-!- Blueblaze [n=nick@c-98-199-143-139.hsd1.tx.comcast.net] has quit [Remote closed the connection]
20090406 19:57:01< Dragonking> boucman: It was supposed to work like that all the time.
20090406 19:57:30< Crab_> boucman: mostly - 2009-04-04.log , but there were some talks before that
20090406 19:57:53< boucman> recruitement is a bit special since it's the only part that can be easily separated, but ok
20090406 19:58:10< Dragonking> boucman: It can be separated, sure, if you have formula to control leader.
20090406 20:00:39< Dragonking> boucman: Anyway, code that did something like that (or was ssuposed to) was in codebase since like forever, just for recruit= formula. We are just fixing it and making work better.
20090406 20:01:38< Crab_> boucman: current change makes it *lazy init* - do not initialize formula_ai if formula_ai recruitment is not needed.
20090406 20:01:41< boucman> ok, I'll  trust your jugement on that, but please try to keep the whole "how AI formulas are called" as simple as possible, we don't want to go in a special casing mess, and with all the proposals floating around it's pretty easy to go that way
20090406 20:02:40< Dragonking> boucman: Recruting is only one special case "mess" AFAIK.
20090406 20:02:44< Crab_> boucman: recruitment is a kind of formula that requires parameters. for example, default_ai recruitment has parameter "recruitment_pattern" 
20090406 20:03:09< boucman> we need to think a bit about FAI and AI parameters interactions
20090406 20:03:35< boucman> should they be global variables, a unique hash-type global variable, something else...
20090406 20:03:49< boucman> but yes, we need to have that
20090406 20:04:03< Crab_> boucman: FAI runs *in context* of AI, so fai should have access to ai parameters
20090406 20:04:14< boucman> actually, having a generic WML<->variant conversion to have easily all AI parameters would be nifty :P
20090406 20:04:15< Crab_> boucman: and AI parameters is a mess that needs discussion...
20090406 20:04:36< Dragonking> boucman: Basically allAI parameters are evil.
20090406 20:04:50< boucman> Dragonking: evil, but needed unfortunately
20090406 20:04:54< Dragonking> Look in what mess we are now.
20090406 20:04:58< Dragonking> boucman: No
20090406 20:05:00< Dragonking> Not needed
20090406 20:05:02< Crab_> boucman: if you have time atm, look at void team::set_time_of_day(int turn, const time_of_day& tod) in team.cpp
20090406 20:05:03< boucman> hmm
20090406 20:05:41< Crab_> boucman: IMO, ai parameters are needed, but we need to make them actually *affect something*, and be understandable to content creators )
20090406 20:05:48-!- molgrum_ [n=molgrum@c83-249-53-163.bredband.comhem.se] has joined #wesnoth-dev
20090406 20:05:53< Dragonking> boucman: Ok. Seriously ;p
20090406 20:06:07< boucman> Dragonking: not having them would force WML authors to have to know FAI to tweak the AI (assuming no AI perception tweaking, which is a different idea)
20090406 20:06:19< Dragonking> For sure parameters are needed, just in case of recruitment I doubt that just "paramters" are needed and desired.
20090406 20:06:22-!- molgrum [n=molgrum@c83-249-53-163.bredband.comhem.se] has quit [Read error: 104 (Connection reset by peer)]
20090406 20:06:28< Dragonking> But for general AI setup, they may be.
20090406 20:06:36< boucman> Dragonking: I'm not sure what you mean
20090406 20:06:46< Dragonking> boucman: Did you read my proposal?
20090406 20:06:46< Crab_> Dragonking: isn't a .fai script location a special case of 'parameter' ?
20090406 20:06:57< boucman> I am currently reading it
20090406 20:07:25< Dragonking> Crab_: Um.. no.
20090406 20:07:26< freim> I would like some opinions on a screenshot. I'm making a new multihex snow mountain and it doesn't quite match the single hex one atm
20090406 20:07:46< freim> do you guys prefer the more blue hue of the single hex one or the darker look of the larger one:
20090406 20:07:48< Dragonking> Crab_: By that logic all C++ code is a parameter and you just need to recompile wesnoth to affect it with such parameters.
20090406 20:07:51< freim> http://feathers.os-tr.net/~hogne/snowtest.png
20090406 20:07:58< freim> I would like to adjust them to be similar
20090406 20:08:12< Dragonking> Crab_: Sorry :)
20090406 20:08:16-!- molgrum_ [n=molgrum@c83-249-53-163.bredband.comhem.se] has quit [Client Quit]
20090406 20:08:20< Dragonking> Crab_: Did not see "location" 
20090406 20:08:22< freim> I might be partial to making the smaller bluish one a bit darker to make it stand out from the snow hills better
20090406 20:08:27< Dragonking> Crab_: You are sort of right.
20090406 20:08:29-!- molgrum [n=molgrum@c83-249-53-163.bredband.comhem.se] has joined #wesnoth-dev
20090406 20:08:38< Reisiger> ilor: Sorry, had dinner. Wanted to ask if you had a chance to look at the README.txt inside the patch
20090406 20:09:10< zookeeper> freim, well, yeah, looks like quite a difference
20090406 20:10:08< Dragonking> boucman: Tu sum it, if we want to give "parameters" for recruting, we would have cover *load* of paramters, and still could not ocver all usages, I have idea of recruitment fai lib that one could use order to customize recruitment 
20090406 20:10:36< ilor> Reisiger: sorry, no, but will look now ;)
20090406 20:10:41< boucman> freim: it doesn't look good to me, I subconciously expect high mountain to have a darker, more impressive look, not sure if that's the case IRL but it doesn't look strange to me
20090406 20:11:23< freim> the point wasn't that it should look like that
20090406 20:11:27< freim> but which way to adjust
20090406 20:11:33< boucman> well, I tried to imagine a recruitement parameter we would want to tweak in a scenario, but couldn't think of one (except the actual recruitement list, but that's a bit special)
20090406 20:11:40< freim> I think I want to adjust the smaller ones to be less bluish
20090406 20:12:07< boucman> freim: oops, just proofread myslef, I meant "it doesn't look bad to me" my sentence makes more sense that way
20090406 20:12:18< freim> boucman: alright :)
20090406 20:12:30< freim> boucman: yeah, I don't think I want to make them so light and blue
20090406 20:12:33< Dragonking> boucman: Like recruit units that are good in forest in north keep, but good in mountains in south one
20090406 20:12:35< zookeeper> freim, i wouldn't mind those and the snow hills being slightly less blue indeed
20090406 20:12:37-!- voris [n=voris@c-24-16-107-149.hsd1.wa.comcast.net] has joined #wesnoth-dev
20090406 20:12:40< freim> I think I want to adjust the smaller ones a bit
20090406 20:12:47< boucman> Dragonking: ok
20090406 20:12:54< freim> zookeeper: yeah, that is my thoughts too
20090406 20:13:10< boucman> freim: I think that what's important with high mountain is that they are VERY affected by light direction
20090406 20:13:39< boucman> on your big mountains the dark side looks natural, but the light side is not bright enough, I would expect ti to be more like the lower mountains
20090406 20:13:39< freim> yeah
20090406 20:13:54< freim> I can adjust that
20090406 20:14:20-!- YogiHH [n=chatzill@d052201.adsl.hansenet.de] has joined #wesnoth-dev
20090406 20:15:30< Dragonking> boucman: Also various different, say we have attack-only scenario where AI defends all the time i caves, amount of villages is limited and we want it to recruit only up to 10 units, to not overrecruit <- it is made up idea, but there can be dozens of such
20090406 20:16:00< boucman> ok, I get it now
20090406 20:16:17< Dragonking> boucman: I did my homework when proposing this idea ;p ;p
20090406 20:17:34< Crab_> boucman: and there is, in my opinion, no problem in having as many AI parameters as needed, including parameters which make sence only for specific AI implementation (as long as those parameters are actually useful), since we are storing them in the config object, not in individual variables.
20090406 20:18:44< boucman> Crab_: yes, I never questionned that
20090406 20:19:19< Dragonking> And my idea is to avoid load of parameters for recruting, just make is tweakable easily with formula library. :)
20090406 20:19:32< boucman> my remark was not about "parameters are bad" but more about "assuming a perfect recruitement AI, I don't see any typical cases where an author would want to tweak the recruitement AI for scenaristic reasons"
20090406 20:19:48< Crab_> boucman: tweak for difficulty levels )
20090406 20:19:58< Crab_> boucman: to limit AI's options and imagination:)
20090406 20:20:09< boucman> Crab_: that's usually done with recruitement list, not by tweaking the AI 
20090406 20:20:22< Ivanovic> boucman: tweaking the *AI* might be a good thing, too
20090406 20:20:25< Crab_> boucman: that's because there was the only way to do that
20090406 20:20:44< boucman> Crab_: you've got a point :P
20090406 20:20:51< Ivanovic> i could imagine an ai that on hard does consider the enemy recruits by far more than on easy
20090406 20:21:10< Crab_> boucman: and, if any *good* alternative will be present, they will try to make the recruitment do things we'd never imagine here :)
20090406 20:21:27< boucman> don't worry, I trust them for that
20090406 20:21:32< Dragonking> I agree with what Crab_ writes here, comletely.
20090406 20:21:45 * Ivanovic can imagine great and strange things regarding recruitment but does not know any real algprithms for this stuff
20090406 20:21:49< boucman> I've seen our users do ANL/RPG/Wesband, I'm sure they'll do more crazy stuff
20090406 20:22:36< Crab_> boucman: so, (from reading current mainline campaigns), I'd say that they will want to change recruitment based on WML events..
20090406 20:23:14< Crab_> boucman: to make it 'more dynamic' during the scenario
20090406 20:23:38< Crab_> boucman: so, if we use WML events to set ai parameters, and read them from formula..
20090406 20:23:46< Crab_> boucman: we'd have one nice way to do that.
20090406 20:24:08< Dragonking> Sure.
20090406 20:25:19< boucman> ok, makes sense
20090406 20:25:49< Crab_> I (personally) do not want to tinker with AI recruitment in any way, since I know that Dragonking is working on that :) So, I just want to make it easier to integrate any of good recruitment ideas into the AI behavior :)
20090406 20:25:58< YogiHH> *sigh*, is there anything someone does not want to do with WML?
20090406 20:26:16 * zookeeper thinks
20090406 20:26:26< Ivanovic> YogiHH: flip burgers
20090406 20:26:27< Crab_> YogiHH: Wonder Markup Language :)
20090406 20:26:43< Dragonking> Crab_: And your ideas are good. Actually that's sopmething I've been thinking about too. :)
20090406 20:26:57< boucman> YogiHH: cofee
20090406 20:27:15 * zookeeper thinks those are good candidates
20090406 20:27:20< Dragonking> boucman: Not *yet*.
20090406 20:27:21< YogiHH> well, i am glad coffee won't be an issue for savegames ;-P
20090406 20:27:27< ilor> [brew_coffee]type=black sugar=no[/brew_coffee]
20090406 20:28:05< Dragonking> Once we have good AI, we can add AI paramters to alter it's personality, and what cofee it likes.
20090406 20:28:23< Polarina> [brew_tea]type=strawberries sugar="3 spoons"[/brew_tea]
20090406 20:29:06< YogiHH> Begone, shadows of the dark! The master of replays is not going to tolerate you anymore!!
20090406 20:29:33< Ivanovic> Polarina: i'd take: type="Earl Grea" sugar="at least three spoon" milk=yes
20090406 20:29:58< Polarina> <brew_tea type="strawberries" sugar="3 spoons" milk="true" />
20090406 20:30:04< Polarina> Ivanovic: There you go.
20090406 20:30:07-!- cjhopman [n=chris@wesnoth/developer/cjhopman] has joined #wesnoth-dev
20090406 20:30:43< Ivanovic> Polarina: ah, good
20090406 20:31:04 * YogiHH feels like the last solid rock in a world of madness :-)
20090406 20:31:54< Ivanovic> though, instead of stawberrie tea i prefer a (good) black tea
20090406 20:32:20 * Polarina prefers lots of sugar.
20090406 20:33:02< Ivanovic> Polarina: lots of sugar is fine in earl grey, too
20090406 20:33:15< Polarina> Earl gray?
20090406 20:33:22< Polarina> s/ay/ey/
20090406 20:33:47< Ivanovic> http://en.wikipedia.org/wiki/Earl_Grey_tea
20090406 20:34:00< Polarina> Never heard of it.
20090406 20:34:33< Ivanovic> really a great tea
20090406 20:35:33< Polarina> I'll keep it in mind if I'll ever come across it.
20090406 20:36:27< Ivanovic> i think this is about the best known kind of black tea
20090406 20:36:32< zookeeper> i thought earl gray was like coca cola, everyone knows it.
20090406 20:36:56< Ivanovic> zookeeper: thought so, too
20090406 20:37:18< boucman> Polarina: afaik earl grey is probably the most famous tea...
20090406 20:38:22< esr> Ivanovic: Around?
20090406 20:38:26< Ivanovic> sure
20090406 20:38:40< Ivanovic> esr: and you maybe want to login at nickserv again
20090406 20:38:42< Ivanovic> ;)
20090406 20:39:50< esr> Ivanovic: Logged in.
20090406 20:40:05< Ivanovic> so what can i do for you?
20090406 20:41:06< esr> Ivanovic: Last week I said to Sirp that I may actually have a functional need for one of those "Director" titles he tosses around occasionally, and he aid it's actually you doing the tossing these days.  So:
20090406 20:41:40< boucman> Dragonking: just commented on your google application
20090406 20:41:57< Ivanovic> what?
20090406 20:42:00< Ivanovic> director titles?
20090406 20:42:13< esr> It looks like we're about to grow a writing department. I've been PMed twice by a guy who wants to do wordsmithing for us.
20090406 20:42:16< Ivanovic> those are not cheap nowaddays, i think they are at "lots of good chocolate" by now
20090406 20:42:18< Ivanovic> ;)
20090406 20:42:54< Crab_> esr: and see https://gna.org/bugs/?13320 - I've got share_view to work in that DM scenario (but you need to make further changes to that scenario, since 'sighted' event to give mages to player will not work if those mages are already uncovered by share_view)
20090406 20:43:15< esr> So I think I now need a title something like "Continuity Director" or "Chief Story Doctor", or something like that.
20090406 20:44:06< esr> I don't care much but it should be something junior wordsmiths think is impressive.
20090406 20:45:06< crimson_penguin> Ivanovic: I'm in luck! I make chocolate! :D
20090406 20:45:13-!- Sapient [n=patrickp@wesnoth/developer/sapient] has joined #wesnoth-dev
20090406 20:45:46< crimson_penguin> Of course, that costs a certain amount of anti-laziness to make, and "good" depends on what you like
20090406 20:45:47< crimson_penguin> Hi Sapient 
20090406 20:45:57< Sapient> yo crim
20090406 20:46:24 * esr likes good chocolate, too
20090406 20:46:38< Ivanovic> hi Sapient
20090406 20:46:55< Ivanovic> esr: tell me the name of the title, i am not good with coming up with names
20090406 20:47:02< esr> Crab_: febdrin is already on it.  I expect to test shortly.
20090406 20:47:11< Sapient> hi Ivanovic
20090406 20:47:13< esr> Ivanovic: I'll meditate on it.
20090406 20:47:47< Crab_> esr: good
20090406 20:48:44< zookeeper> maybe we need a new prose/storyline subforum. let's call it "writers' block".
20090406 20:50:36< zookeeper> (not funny? oh darn)
20090406 20:50:55< boucman> esr: do you want it related to THoT, you toolsmith work, or your english reviewing work ?
20090406 20:51:17 * boucman loves to come up with names :)
20090406 20:51:48< dfranke> esr: "Lorekeeper of Irdya".
20090406 20:52:01< boucman> not bad...
20090406 20:52:51< zookeeper> why not just "Lord of Prose" in line with several other titles?
20090406 20:52:51< esr> Actually that isn't bad at all, but for this purpose I need something more focused on the nglish reviewing.
20090406 20:53:07< esr> zookeeper: That could work, yes.
20090406 20:53:52< Dragonking> boucman: Ok, thanks. :)  Will answer once I have some time (yea.. this if.... been thinking about itmplementing it as such but apparently forgot.. )
20090406 20:54:07< Crab_> esr: "High Loremaster of Wesnoth"
20090406 20:54:15< boucman> please avoid "director" that word is currently only used for "chief of an artist community on the forum" and it would be nice to keep it that way
20090406 20:54:17< esr> :-)
20090406 20:54:24< esr> Noted.
20090406 20:54:51< esr> That's a good argument for "Lord of Prose".
20090406 20:56:11-!- xonev [n=chatzill@59.92.66.48] has quit ["ChatZilla 0.9.84 [Firefox 3.0.8/2009032609]"]
20090406 20:58:51< Sapient> "Grammar Nazi" ?
20090406 20:59:11< Sapient> just kidding... :p
20090406 20:59:32< dfranke> Sapient: not really what you want to be for Wesnoth prose anyway.
20090406 20:59:34< dfranke> Linguists shoot grammar nazis.
20090406 20:59:38< Ivanovic> Turuk: so please give esr the title "Lord of Prose" (has to be created first)
20090406 21:00:08< Turuk> Sure.
20090406 21:00:20< Sapient> or "Annalist"
20090406 21:00:26< esr> :-)
20090406 21:00:36< zookeeper> (and maybe rename your website "guns and proses")
20090406 21:00:40< Ivanovic> esr: yeah, i am currently busy practising the concept of "delegating work"
20090406 21:00:40< esr> That's continuity-focused again.
20090406 21:00:41< Ivanovic> ;)
20090406 21:00:43< zookeeper> (i'm so clever tonight)
20090406 21:01:20 * esr squirts zookeeper with a water pistol
20090406 21:01:59< esr> dfranke: "Linguists shoot grammar nazis." - well put.
20090406 21:02:18< dfranke> esr: I stole that from somewhere; I forget where.
20090406 21:02:26< dfranke> esr: old Reddit comment or something.
20090406 21:04:54< Turuk> esr: done :)
20090406 21:05:08< esr> Thanks.
20090406 21:11:22-!- silene [n=plouf@wesnoth/developer/silene] has joined #wesnoth-dev
20090406 21:11:23< Sapient> ah, I thought Supreme Annalist would have been better
20090406 21:11:33< silene> hi
20090406 21:11:46< Sapient> yo silene
20090406 21:13:33< freim> zookeeper: made a bunch of tweaks: http://feathers.os-tr.net/~hogne/snowtest2.png 
20090406 21:13:42< freim> think it is getting closer to where I want it
20090406 21:14:10< zookeeper> yeah, much better
20090406 21:14:15< boucman> yes, better
20090406 21:14:44< CIA-30> fendrin * r34589 /trunk/data/campaigns/Legend_of_Wesmere/scenarios/02_Hostile_mountains.cfg: LoW 02: Made Olurf not attacking Elves on water tiles.
20090406 21:16:57< YogiHH> zookeeper: did you find the cause for the missing Konrad graphic when loading from a savegame?
20090406 21:17:12-!- cjhopman [n=chris@wesnoth/developer/cjhopman] has quit ["Konversation terminated!"]
20090406 21:18:09-!- elias [n=allefant@allegro/developer/allefant] has quit ["Leaving"]
20090406 21:20:06< zookeeper> YogiHH, nope
20090406 21:21:02< CIA-30> fendrin * r34590 /trunk/data/campaigns/Legend_of_Wesmere/scenarios/07_Elves_last_stand.cfg: LoW 7: Added useless eye candy.
20090406 21:23:02-!- Elvish_Pillager [n=eli@66-189-40-122.dhcp.oxfr.ma.charter.com] has joined #wesnoth-dev
20090406 21:24:34-!- cjhopman [n=chris@wesnoth/developer/cjhopman] has joined #wesnoth-dev
20090406 21:28:44-!- grzywacz [n=grzywacz@wesnoth/developer/grzywacz] has joined #wesnoth-dev
20090406 21:29:15< YogiHH> boucman: If loading the start-of-scenario save, the image for Konrad is missing (only hitpoint/experience bar, ellipsis and crown are shown). What is the WML value i would have to check to see which is the supposed image?
20090406 21:30:03< boucman> no idea :P
20090406 21:30:23< boucman> more seriously, though, is this unit special, animation wise, in any way ?
20090406 21:30:35 * YogiHH checks
20090406 21:30:54< boucman> I usually take the image from the animation in unit_type, not unit, so if unit_type is not properly found or incorrect, it might cause problems
20090406 21:31:46< YogiHH> boucman: only attack and defense animations, it seems
20090406 21:31:56< boucman> hmm
20090406 21:32:03< boucman> maybe that's it...
20090406 21:32:15< YogiHH> boucman: which animation in unit_type are you searching for?
20090406 21:32:39< boucman> standing_anim if any, and if not, I'm doing a default using the image= attribute
20090406 21:32:50< boucman> (btw, is the image correct in the side panel ? )
20090406 21:33:01< YogiHH> boucman: oops, sorry, attack and defense are from the savegame and variations
20090406 21:33:07 * YogiHH checks the unit type
20090406 21:33:38< YogiHH> boucman: image in the side panel is missing as well, portrait is there, though
20090406 21:34:09< boucman> ok, then it's the image= attribute that is not correct/not loaded properly
20090406 21:34:22< boucman> it's not an animation problem, so to speak...
20090406 21:34:45< YogiHH> boucman: yes, the unit type doesn't have a standing animation
20090406 21:35:24< YogiHH> boucman: image="units/konrad-fighter.png" looks ok to me
20090406 21:35:35< CIA-30> hogne * r34591 /trunk/data/core/ (10 files in 2 dirs): 
20090406 21:35:35< CIA-30> Added two variants of 2x2 snow mountains
20090406 21:35:35< CIA-30> Some color tweaks to the single tile snow mountains
20090406 21:35:35< CIA-30> Reduced probability for 2x2 variants to offset for still missing 2x4
20090406 21:35:35< CIA-30> variants
20090406 21:35:49< YogiHH> and the file is there, so maybe the code doesn't search in the right directory?
20090406 21:36:00< boucman> hmm
20090406 21:36:17< boucman> what suprises me is that it's a bug that appears only when loading from savegame...
20090406 21:36:25< YogiHH> does it search for core and campaign images respectively?
20090406 21:36:31< boucman> maybe the image= attribute is not saved properly/at all
20090406 21:36:55< YogiHH> that attribute is from the savegame
20090406 21:37:04< freim> zookeeper: I reduced the probabilitiy somewhat on the 2x2, you might want to adjust it further
20090406 21:37:19< zookeeper> freim, roger
20090406 21:38:47< boucman> hmm
20090406 21:38:58< boucman> YogiHH: well, I can tell you it doesn't find the file :)
20090406 21:39:11< boucman> what's suprising is that the bug is konrad only...
20090406 21:39:21< boucman> could it be not checking campaign specific folders ?
20090406 21:39:24< YogiHH> hmm, maybe it doesn't know the campaign for some reason. I don't know why but it would match the bug
20090406 21:39:40< YogiHH> boucman: where is the drawing code for the unit to be found?
20090406 21:40:31< YogiHH> boucman: or better even, where is the place it searches for the image?
20090406 21:40:37< boucman> unit_animation::fill_initial_animations
20090406 21:40:43< boucman> should be a good starting point
20090406 21:40:47< Crab_> YogiHH: maybe --log-debug=filesystem will be of any help ?
20090406 21:40:57< boucman> yes, that too
20090406 21:41:07< YogiHH> yeah, i will try that first
20090406 21:44:58< YogiHH> now that is interesting, it looks for HttT images within the UTBS folder...
20090406 21:45:09< boucman> wow
20090406 21:45:12< boucman> hmm
20090406 21:45:14< YogiHH> no wonder it doesn't find anything
20090406 21:45:33< boucman> grep -r Under data/campaigns/Heir_To_The_Throne time, I guess
20090406 21:46:06< YogiHH> boucman: ?
20090406 21:46:12< zookeeper> that's exotic since the UtBS dir shouldn't even be [binary_path]ed unless you actually run it
20090406 21:46:22-!- Rhuvaen [n=chatzill@dslb-088-072-203-062.pools.arcor-ip.net] has joined #wesnoth-dev
20090406 21:46:44< YogiHH> hmm, it happens to be the last campaign in an alphabetical list...
20090406 21:46:56 * YogiHH smells blood :-)
20090406 21:47:34< boucman> hmm
20090406 21:49:06< YogiHH> boucman: where are the pathes set for image searching?
20090406 21:49:26< boucman> no idea :/
20090406 21:49:33< boucman> I'm only a user of the filesystem API
20090406 21:51:16< boucman> zookeeper might know
20090406 21:51:29< boucman> what's weird is the diff normal game/save game, though
20090406 21:52:12< YogiHH> boucman: well, i restructured some of the savegame code, it is possible that i overlooked something for start-of-scenario saves
20090406 21:52:33< boucman> oh, ok
20090406 21:54:47< zookeeper> from WML image search pathss are set via [binary_path]s, but the UtBS one is inside the campaign #ifdef
20090406 22:03:28-!- Lairaku [n=chatzill@247.79.broadband5.iol.cz] has quit ["ChatZilla 0.9.83 [Firefox 2.0.0.20/2008121709]"]
20090406 22:04:42-!- Sapient [n=patrickp@wesnoth/developer/sapient] has left #wesnoth-dev []
20090406 22:07:25< Crab_> YogiHH: see get_binary_file_location in filesystem.cpp
20090406 22:07:49< YogiHH> Crab_: ok, thanks
20090406 22:11:23-!- fendrin [n=fabi@wesnoth/developer/fendrin] has quit [Read error: 104 (Connection reset by peer)]
20090406 22:11:49-!- fendrin [n=fabi@f051019205.adsl.alicedsl.de] has joined #wesnoth-dev
20090406 22:12:04-!- happygrue [n=George@wesnoth/developer/wintermute] has quit ["http://xkcd.com/91/"]
20090406 22:18:47-!- Rhuvaen [n=chatzill@dslb-088-072-203-062.pools.arcor-ip.net] has quit ["later"]
20090406 22:20:17< Reisiger> Q: what kind of events should end up logged in stdout.txt?
20090406 22:20:38-!- Chusslove [n=caslav@brsg-d9befa68.pool.mediaWays.net] has joined #wesnoth-dev
20090406 22:22:44< boucman> Reisiger: no hard rules, but usually non-critical abnormal conditions that the user might want to report
20090406 22:23:18< Reisiger> Any way for me to verify stdout works? So far I got an empty txt file
20090406 22:23:49< zookeeper> i don't recall ever getting anything in stdout.txt
20090406 22:23:52< boucman> actually let me correct what I said above
20090406 22:23:56< zookeeper> stderr.txt is where all goes
20090406 22:24:00< boucman> I don't think we print on stdout
20090406 22:24:04< Reisiger> Had to build SDL_main.dll to get that and I need to know if I have to build SDL.dll as well.
20090406 22:24:06< boucman> it was stderr I was thinking of
20090406 22:24:18< Reisiger> Ok. That one has information :)
20090406 22:24:57< grzywacz> hi...
20090406 22:25:17 * Reisiger goes of updating the VC9 readme to reflect that
20090406 22:28:27-!- crimson_penguin [n=ben@wesnoth/developer/crimsonpenguin] has quit ["Leaving"]
20090406 22:29:35< Crab_> YogiHH: http://wesnoth.pastebin.com/m39b6cfff
20090406 22:31:36< YogiHH> Crab_: yeah, for some reason it is reading the UtBS path from the cache
20090406 22:31:38< Ivanovic> corn: i am currently reading that you are a high school senior
20090406 22:31:53< Ivanovic> corn: uhm, are you 100% sure that you are eligible for summer of code?
20090406 22:31:57< YogiHH> Crab_: i am searching at the moment where exactly that happens
20090406 22:32:31< Ivanovic> (me just wanting to be 100% sure that, if you are accepted, everything really works out as it should)
20090406 22:33:45< Crab_> YogiHH: it gets into the startup game config somehow, at startup...
20090406 22:33:57< YogiHH> Crab_: exactly
20090406 22:34:25< Ivanovic> corn: cf http://socghop.appspot.com/document/show/program/google/gsoc2009/faqs#eligibility
20090406 22:34:32< Ivanovic> "In order to participate in the program, you must be a student. Google defines a student as an individual enrolled in or accepted into an accredited institution including (but not necessarily limited to) colleges, universities, masters programs, PhD programs and undergraduate programs. You are eligible to apply if you are enrolled in an accredited pre-university educational program provided you meet all of the other eligibility 
20090406 22:34:34< Ivanovic> requirements."
20090406 22:34:53< Ivanovic> me not knowing the things for sure i prefer asking once too often instead of facing the problems later on
20090406 22:44:38< corn> I have been accepted into several universities. I can provide the acceptances letters 
20090406 22:44:50< Crab_> YogiHH: game.cpp:1531
20090406 22:45:00< Crab_> // Append the $user_campaign_dir/*/_main.cfg files to addons_to_load.
20090406 22:45:17< corn> Ivanovic: noy and I asked about this in the gsoc IRC channel when I first expressed interest in participating
20090406 22:45:26< Ivanovic> corn: ah, okay
20090406 22:45:32< noy> Ivanovic: dave and I sorted it out
20090406 22:45:44< Ivanovic> corn: just wanted to be 100% sure since "done with high school" is really one of the bordercases
20090406 22:45:52< corn> sure, completely understandable
20090406 22:45:57< noy> he needs to show his acceptance into an accredited university
20090406 22:46:06< noy> which he has
20090406 22:46:08< noy> so he's fine
20090406 22:46:23< Ivanovic> perfect
20090406 22:46:52< Ivanovic> (so in germany it would probably be a "the student has his a-level exams done" for most universities and subjects"...
20090406 22:47:11< YogiHH> Crab_: It's actually happening before that, where cache_.get_config() is called. Can't tell you the linenumber, since i don't have latest svn here.
20090406 22:47:32< Crab_> ok )
20090406 22:47:34< Ivanovic> (yes, only a limited number of studies is regulated on the amount of people getting in, most regulate themselves by being really strict in the first terms and thus failing those that are not suited)
20090406 22:48:16< Crab_> cache_.get_config is a bit after that )
20090406 22:50:27< YogiHH> oops, not over here...
20090406 22:51:19-!- Velory [n=GProcess@85.106.139.45] has joined #wesnoth-dev
20090406 22:51:37< Velory> Hi
20090406 22:51:48< noy> Ivanovic: I think it has to show he will be going to university...
20090406 22:52:04< cjhopman> not even that, just that he's accepted
20090406 22:57:41-!- Espreon [n=chatzill@wesnoth/developer/espreon] has joined #wesnoth-dev
20090406 22:58:16< noy> cjhopman: yeah... thats right
20090406 22:58:48< cjhopman> a broad interpretation of student I guess
20090406 22:59:45< Ivanovic> and in germany it is more like "you are eligible, so you are almost automatically accepted" (at least as long as it is not one of those really strange areas of studies like psychology or such where they got some crude entry barriers because of "by far too many people applying")
20090406 23:09:10< silene> zookeeper: i just saw your post about the lessluck era; it feels like an impressively broken concept to me; you could just have changed all the defense values to 50% and you would have obtained a similar result
20090406 23:10:11< zookeeper> silene, how come? if all defense values were 50%, then you could still "easily" have long streaks of hits and misses
20090406 23:10:23< corn> Ivanovic: that's a pretty nice system. in the US it is _extremely_ hard to get into a university. I was rejected by my top choices (MIT, stanford) - they both have acceptance rates of 10% or less
20090406 23:10:45< cjhopman> for example, if a unit does only magic attacks, they will end up hitting about 50% of them
20090406 23:10:57< silene> zookeeper: what i mean is that, from a statistical point of view, what you are trying to do is bayes' formula; and looking at your description, this is not at all bayes' formula
20090406 23:10:57< Ivanovic> corn: on the other hand it is: they wipe the floor with your ass in the first terms to make the weak ones leave themselves
20090406 23:11:01< Ivanovic> ;)
20090406 23:11:19< corn> sure. but atleast you'll know why you were screwed
20090406 23:11:43< zookeeper> silene, sorry, i understand very little of mathematics and statistics, especially the english vocabulary ;)
20090406 23:11:44< Ivanovic> that is right
20090406 23:11:45< Elvish_Pillager> corn, well, it's extremely hard to get into _some_ universities
20090406 23:11:58< zookeeper> cjhopman, how?
20090406 23:12:39< zookeeper> if you got 70% to hit, and you... A) miss -> 80% CtH on next strike B) hit -> 60% CtH on next strike
20090406 23:12:51< silene> zookeeper: think about it, the only case when you have as much chance to get +10% and -10% is the case where you have as much chance to hit or to miss; as a consequence, you are making the chance to hit tend toward 50%, that is the real defense becomes irrelevant
20090406 23:12:51< cjhopman> but you are more likely to hit
20090406 23:13:34< Ivanovic> Velory: got some mins?
20090406 23:13:58< YogiHH> Crab_: ok, i found where UTBS paths are coming from, finally
20090406 23:14:11< YogiHH> Crab_: However, i am not sure what to do about it
20090406 23:14:25< Crab_> YogiHH: good :) from reading all campaigns _main.cfg's in a row ?
20090406 23:14:45< YogiHH> Crab_: they are defined in scenario-test.cfg ;-)
20090406 23:14:48< Crab_> hehe
20090406 23:15:01< YogiHH> Kaleh is the bad boy :-)
20090406 23:15:04< dfranke> corn: I didn't even bother applying to MIT for undergrad because I knew my grades weren't high enough...
20090406 23:15:14< Crab_> change him for someone else )
20090406 23:15:17< Crab_> or let Dragonking do it
20090406 23:15:24< Velory> Ivanovic:  yes
20090406 23:15:35< Ivanovic> Velory: please join #wesnoth-mentor for a moment
20090406 23:15:35< dfranke> corn: but for grad school, I had perfect GREs and two very strong recommendations, so I gave it a shot, and *still* got rejected.
20090406 23:15:41< YogiHH> Crab_: I wonder, if the problem is lying deeper, actually.
20090406 23:15:50< Velory> Ivanovic:  ok
20090406 23:16:00< zookeeper> silene, well, can't say i'm succeeding in figuring out how that would happen
20090406 23:16:03< YogiHH> Crab_: If i load a HttT save, shouldn't the binary paths be adjusted?
20090406 23:16:22< Crab_> YogiHH: yes, HttTs config should be loaded..
20090406 23:16:35< YogiHH> Crab_: And that doesn't seem to happen
20090406 23:16:43< zookeeper> silene, if you don't mind, i could use an example case ;)
20090406 23:16:49< Crab_> Is is a specific save ?
20090406 23:16:55< Crab_> or any save will be ok ?
20090406 23:17:13< Crab_> s/Is is/is it
20090406 23:17:21< YogiHH> Crab_: only start-of-scenario save are doing this
20090406 23:17:33< silene> zookeeper: imagine you have 60% chance to hit; it means that you 60% chance that your next chance to hit is reduced to 50%; does that make sense?
20090406 23:17:42< cjhopman> zookeeper: http://codepad.org/O50V1mBD
20090406 23:17:54< zookeeper> silene, such as, if i have an adept in water (20% def) with infinite HP, how does the CtH of an ulf attacking it end up shifting towards 50%?
20090406 23:17:58-!- Turuk_ [n=Turuk@wesnoth/forumsith/turuk] has joined #wesnoth-dev
20090406 23:18:19< zookeeper> silene, sure
20090406 23:18:37< Crab_> YogiHH: what problems can additional loading a campaign config cause ?
20090406 23:18:54< Crab_> s/a /of a
20090406 23:19:12< YogiHH> Crab_: countless problems ;-)
20090406 23:19:15< silene> zookeeper: similarly, if you have 40% chance to hit; then you have 60% chance your next chance to it is increased to 50%
20090406 23:19:25< Crab_> and what about loading only binary_path ?
20090406 23:19:27< dfranke> corn: don't know if you heard the same line, but when I toured MIT (the legitimate kind of tour, but I've been on the other :-), the admissions officer bragged that they could reject every student they accepted, accept the next batch, and nobody would notice anything odd.
20090406 23:19:45< corn> yeah, that's definitely true
20090406 23:19:51< Crab_> YogiHH: or it should be stored in the save itself ?
20090406 23:19:52< silene> zookeeper: so whether you are above or below 50%, your next chance to hit has 60% chance being 50%
20090406 23:20:14< corn> anyway, I am going to finish off/improve my prototype today. I should have sourcecode packaged for it too as well
20090406 23:20:17< zookeeper> silene, hmm, ok..
20090406 23:20:21< YogiHH> Crab_: I suspect that something in the save is triggering it and that this information is missing
20090406 23:20:21< corn> and the timeline that I was supposed to update yesterday :)
20090406 23:20:49< cjhopman> zookeeper: with better output http://codepad.org/zwCnEdIS
20090406 23:20:57< boucman> how come we load test-scenario if game is not launched with the -t param ?
20090406 23:21:01< silene> zookeeper: but look at cjhopman's figure, it's quite clear; it starts from 70%, and after an "infinite" amount of hit, it stabilizes at 50%
20090406 23:21:09< Crab_> boucman: it's referenced from _main.cfg
20090406 23:21:27< boucman> should it be ?
20090406 23:21:54< Crab_> boucman: no. we only need to be able to translate from scenario id to scenario path
20090406 23:22:13< boucman> k, probably something to fix here, then
20090406 23:22:29< Crab_> test -> scenario-test.cfg,  ai_arena_small -> ai/scenarios/scenario-AI_Arena_small.cfg , ...
20090406 23:23:33< Crab_> YogiHH: so, *some* saves trigger that behavior ? or *all* saves ?
20090406 23:23:36< zookeeper> cjhopman, silene, ok, so it seems...
20090406 23:23:51< YogiHH> Crab_: no, just some
20090406 23:24:01< Crab_> YogiHH: can I have one ?
20090406 23:24:18< YogiHH> Crab_: i think i just found it, one minute
20090406 23:24:21< Crab_> ok
20090406 23:24:28< silene> zookeeper: the correct way to do it is to not have the same variation for both winning and losing
20090406 23:25:26-!- Turuk [n=Turuk@wesnoth/forumsith/turuk] has quit [Connection timed out]
20090406 23:31:06< zookeeper> silene, cjhopman, i get the point, but are you sure the fact that you're going to be attacking units with very different defense values won't somehow...uh, magically alleviate the problem?
20090406 23:34:54-!- noy [n=Noy@wesnoth/developer/noy] has quit []
20090406 23:35:18-!- Velory [n=GProcess@85.106.139.45] has left #wesnoth-dev []
20090406 23:35:38-!- Turuk_ [n=Turuk@wesnoth/forumsith/turuk] has left #wesnoth-dev []
20090406 23:35:50< Crab_> zookeeper: can you give a link to that forum post ?
20090406 23:35:56< YogiHH> Crab_: no, false alarm. You want me to send you the file?
20090406 23:36:24< Crab_> YogiHH: if there's something *reproducibly wrong* with it, yes ( Crab on forum or by email )
20090406 23:36:26< silene> zookeeper: i'm sure it won't; on the long run, you will tend to 50%; and on the short run, you will probably see some large variations, possibly worse than with an unbiased rng (try to imagine what happens if you first attack with a mage and then attack against an orcish assassin)
20090406 23:37:18< YogiHH> Crab_: trying DCC...
20090406 23:37:21< Crab_> ok
20090406 23:37:39< YogiHH> Crab_: got it?
20090406 23:37:44< Crab_> yes
20090406 23:38:21< Crab_> what is wrong with that save ?
20090406 23:38:41< YogiHH> Crab_: It misses the image for konrad
20090406 23:39:02< Crab_> yes, so it doesn't load...
20090406 23:39:19< Crab_> even segfaulted for me )
20090406 23:39:30< zookeeper> silene, oh, this is fascinating :) i'll see if i can get a feel for that with some ulfserkers...
20090406 23:39:48< YogiHH> Crab_: uh, yeah, that uses a slightly different savegame format :(
20090406 23:40:01< YogiHH> Crab_: didn't think of that
20090406 23:40:09< Crab_> but it still has "could not open image 'units/konrad-fighter.png'"
20090406 23:40:19< zookeeper> Crab_, http://www.wesnoth.org/forum/viewtopic.php?f=19&t=24869 ?
20090406 23:40:19< Crab_> so, it segfaults *after* don't finding images )
20090406 23:40:24< Crab_> zookeeper: thanks
20090406 23:40:28< YogiHH> yes, that's true
20090406 23:43:09-!- boucman [n=rosen@wesnoth/developer/boucman] has quit [Remote closed the connection]
20090406 23:43:35< zookeeper> wow, my ulfs are really hitting that 70% unit half the time ;)
20090406 23:43:46< silene> :-)
20090406 23:46:36< zookeeper> now, the problem is that dynamic increments (not in static x% steps) would be really hard to make
20090406 23:46:49< cjhopman> yeah, that code i posted i had a huge bug... here it's fixed http://codepad.org/YTQQgSKp
20090406 23:47:09< cjhopman> stabilizes around 50% quickly
20090406 23:47:11< YogiHH> Crab_: binary paths for HttT are set, but too late, obviously
20090406 23:47:58-!- cjhopman [n=chris@wesnoth/developer/cjhopman] has quit ["Konversation terminated!"]
20090406 23:48:58< CIA-30> ivanovic * r34592 /trunk/po/ (wesnoth/zh_CN.po wesnoth-manpages/zh_CN.po): updated Chinese (Simplified) translation
20090406 23:49:04< CIA-30> ivanovic * r34593 /branches/1.6/po/ (wesnoth/zh_CN.po wesnoth-manpages/zh_CN.po): updated Chinese (Simplified) translation
20090406 23:49:33< silene> zookeeper: how hard is it? hard as in impossible? (just trying to get a clue at what wml allows or not)
20090406 23:50:26< zookeeper> silene, not impossible, i think, but the approach would have to be totally different than the one i have now and somewhat more complicated
20090406 23:51:04< zookeeper> how i currently do is like this:
20090406 23:51:45-!- molgrum [n=molgrum@c83-249-53-163.bredband.comhem.se] has quit ["Lämnar"]
20090406 23:53:52< zookeeper> each attack gets a bunch of invisible weapon specials, each giving a certain CtH modifier (-10, -20, 10, 20, 30, etc) and only one is active at any given time, depending on the value of a certain variable of the unit. then i just have some attack and hit and miss events, which bump that variable of the currently fighting unit(s) so that the right special is active to give the CtH modifier.
20090406 23:54:29< YogiHH> Crab_: Correction, it is using the wrong path, it looks for HttT within userdata/... O-O
20090406 23:54:30< zookeeper> with dynamic modifiers, i'd have to actually constantly add and remove new specials into the units mid-fight
20090406 23:56:08< silene> zookeeper: an intermediate solution would perhaps be to have a bit more granularity -15 -10 -5 +5 +10 +15... and to choose the one the closest to the mathematical one
20090406 23:57:13< silene> that way, if you miss a 70%, you get +10%, but if you hit, you get -5% (just making up the value, i have no idea if it would change that much on the first attack)
20090406 23:57:40< Polarina> Why not just keep it 70% at all times?
20090406 23:58:39< Polarina> silene: -5% means it will be 65% or -5% of 70% (thus 66.5%)?
20090406 23:59:17< silene> Polarina: 65%, the wml system is additive
20090406 23:59:33< Polarina> Ok.
20090406 23:59:51< Polarina> WML is evil, in my opinion.
--- Log closed Tue Apr 07 00:00:29 2009