--- Log opened Sun Apr 12 00:00:31 2009
20090412 00:07:21< mordante> gnutoo_, Crab_ the formula parser + engine is written by Sirp that's why he might have an idea
20090412 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 | 52 bugs, 232 feature requests, 11 patches | logs: http://irclogs.wesnoth.org/ | Don't paste on IRC! Use a pastebin: http://wesnoth.pastebin.com | http://imagebin.org
20090412 00:10:13< Crab_> mordante: well, it looks like a boost bug to me. since when you've told gnutoo_ to evaluate i1, it showed up ok. and all the formula engine does is a call to boost::regex_search, which segfaults...
20090412 00:11:19< mordante> Crab_, I think so too, but Sirp might have a clue
20090412 00:11:35< mordante> and I also asked gnutoo_ to test with the latest boost
20090412 00:11:48< Crab_> mordante: so did I...
20090412 00:12:28< Crab_> mordante:  std::string::const_iterator i1 = str.begin(), i2 = str.end();
20090412 00:12:39< Crab_> so, i1 and i2 start sane
20090412 00:13:02< mordante> yup
20090412 00:13:33-!- BenUrban_ [n=benurban@unaffiliated/benurban] has joined #wesnoth-dev
20090412 00:13:52< Crab_> and i2 is not moved at all
20090412 00:14:26< Crab_> while i1 is moved as follows:     i1 = res.end = i1 + match.length();
20090412 00:15:18< Crab_> mordante: where match.length() is, by common sense and definition, is less-or-equal to the length of the string from i1 to i2.
20090412 00:15:31< Crab_> since it's a matched *substring*
20090412 00:15:38< Crab_> so, boost -> evil.
20090412 00:15:44< Ivanovic> off to bed, n8
20090412 00:15:48< mordante> night Ivanovic 
20090412 00:15:53< Crab_> night, Ivanovic
20090412 00:16:17< mordante> Crab_, I also saw that and expect the problem to be with boost
20090412 00:16:36< Crab_> mordante: something platform-specific, maybe..
20090412 00:18:29-!- gnutoo_ [n=gnutoo@host216-163-dynamic.51-79-r.retail.telecomitalia.it] has quit [Read error: 113 (No route to host)]
20090412 00:18:51< mordante> might be 
20090412 00:20:27< CIA-30> silene * r34740 /trunk/data/campaigns/Sceptre_of_Fire/scenarios/8_The_Dragon.cfg: Fixed a typo that broke all the translations of this scenario. Nice...
20090412 00:24:06< dfranke> O_O
20090412 00:27:16< Polarina> silene: Thanks.
20090412 00:27:54< dfranke> ok, looks like the typo was just from a few days ago.
20090412 00:28:11< dfranke> I was wondering how long that had been around.
20090412 00:28:32-!- BenUrban [n=benurban@unaffiliated/benurban] has quit [Read error: 110 (Connection timed out)]
20090412 00:28:46-!- BenUrban_ is now known as BenUrban
20090412 00:30:27< Ivanovic> added some more markers, though only one (in trunk) makes a difference, the rest is only for completion
20090412 00:34:57< Polarina> The string in wesnoth.po "stoned: ", shouldn't it be "petrified: "?
20090412 00:35:49< Polarina>  1. To convert, as any animal or vegetable matter, into stone
20090412 00:35:49< Polarina>  or stony substance; as, petrified wood.
20090412 00:35:59 * dfranke wonders how many translators have translated that as "high".
20090412 00:36:25-!- molgrum [n=molgrum@c83-249-53-163.bredband.comhem.se] has quit ["Lämnar"]
20090412 00:37:05< Polarina> It should be petrified.
20090412 00:39:11 * esr adds a translation note.
20090412 00:39:23< Polarina> What note?
20090412 00:40:07-!- cjhopman [n=chris@wesnoth/developer/cjhopman] has quit ["Konversation terminated!"]
20090412 00:41:03< CIA-30> ivanovic * r34746 /trunk/data/ (9 files in 6 dirs): add some textdomain markers to files that had none, should not introduce any strings, just for completions sake (every .cfg should have a marker, have done the job util utils/change_textdomain -t )
20090412 00:41:04< CIA-30> ivanovic * r34745 /branches/1.6/data/gui/default/macros/ (horizontal_scrollbar.cfg vertical_scrollbar.cfg): add some textdomain markers just for "completion sake" (every .cfg file should have a textdomain marker!)
20090412 00:41:06< CIA-30> ivanovic * r34744 /trunk/data/campaigns/Northern_Rebirth/units/Cave_Serpent.cfg: add another missing textdomain marker
20090412 00:41:08< CIA-30> silene * r34743 /trunk/data/campaigns/Legend_of_Wesmere/scenarios/ (08_Council_of_hard_choices.cfg 16_The_Chief_must_die.cfg): Fixed markup.
20090412 00:41:09< CIA-30> silene * r34742 /trunk/po/ (5 files in 5 dirs): Updated French translation.
20090412 00:42:23< Polarina> Because "stoned: " doesn't exactly describe what is meant (according to the manual).
20090412 00:44:51< dfranke> silene: heh, we just hit a race condition of sorts in the bug tracker.
20090412 00:44:51-!- zookeeper [n=l@wesnoth/developer/zookeeper] has quit []
20090412 00:45:05-!- ryochan7 [n=ryochan7@173.9.238.161] has quit ["Leaving"]
20090412 00:45:15< dfranke> silene: you closed #13356 it as invalid at the same time I was assigning it to myself.
20090412 00:46:44< silene> dfranke: there may still be a bug (in particular i wonder how hits too strong wrt to remaining life are accounted for), but the report itself is not a bug imo
20090412 00:48:20< dfranke> silene: well, we haven't released the fix of #13277 yet
20090412 00:48:54< dfranke> silene: but I think your math is wrong too.  .4*.4*36 + 2*.4*.6*30 = 17.76
20090412 00:48:56< dfranke> silene: not 20.
20090412 00:49:49< Polarina> dfranke: It's 20.16
20090412 00:49:50< silene> dfranke: here it's 20.16; seems like my pocket calculator is broken then
20090412 00:51:48< dfranke> silene: nm... you have it right.
20090412 00:51:51< dfranke> silene: it was bc being broken.
20090412 00:52:21< dfranke> so yeah... bug is invalid.
20090412 00:57:38< _ping_> it is not invalid (at least i dont think so)
20090412 00:58:02< dfranke> _ping_: why?
20090412 00:58:45< _ping_> first i got replays that "should not happen" i should attach thoses
20090412 00:59:31< _ping_> second after the explained attack the stats of the attacker hitting once says 30 of 20 when it easy to see that it is 30 of 24
20090412 00:59:59< dfranke> _ping_: no, it's 20.
20090412 01:00:08< _ping_> try doing 100000 simulations
20090412 01:00:33< _ping_> and i bet you that you will be of with the a %
20090412 01:01:03< dfranke> Possible outcomes are 36 (with 16% probability), 30 (with 48% probability), and 0 (with 36% probability)
20090412 01:01:37< _ping_> i know what you are trying to do
20090412 01:01:51< Crab_> _ping_: as the horseman has only 36 hp, there's no way to deal 60 dmg to him
20090412 01:02:14< _ping_> but you got the math wrong, you should not take into account that he ca die when he doesnt
20090412 01:02:29< _ping_> the stats are right calc when he dies
20090412 01:02:36< _ping_> but wrong when he lives
20090412 01:02:49< Crab_> _ping_: how do you calculate 24 ?
20090412 01:03:11-!- Elvish_Pillage2 [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!"]
20090412 01:03:14< dfranke> are you saying that if he dies, it reports 36/20?
20090412 01:03:43< dfranke> because that's what it should be.
20090412 01:05:24< _ping_> true
20090412 01:05:51< _ping_> but as long as he dont die you should not downgrade the ev dealt
20090412 01:06:00< _ping_> it will lead to a wrong result
20090412 01:06:21< dfranke> so you're saying the EV should be dependent on the result?
20090412 01:07:03< _ping_> you can easy either lean statistics or run a 100000 simulations that will show that the ev as it is now is screwed
20090412 01:07:11< _ping_> learn*
20090412 01:07:35< dfranke> answer my question, please.
20090412 01:14:02< mordante> night
20090412 01:14:15< Crab_> night, mordante
20090412 01:14:16< silene> _ping_: here is program that does 100000 simulations: http://wesnoth.pastebin.com/d65dbd14b and guess what the answer is? 20...
20090412 01:14:38-!- mordante [n=mordante@wesnoth/developer/mordante] has quit ["Leaving"]
20090412 01:14:50< _ping_> i'm playing so i answer as soon as i can, but yes since there is a possible outcome that depend on a unit being dead and therefore can not take full damage the result will be reflect in ev, but as it is now the ev will appear to be wors then it is for the assasin in the above case, okay let me attach an file show the stats in game
20090412 01:15:18< _ping_> silene you need to use the game enigne ofcause:)
20090412 01:16:07< _ping_> i do not disupte the calc done i dispute the presumptions
20090412 01:16:22-!- noy [n=Noy@wesnoth/developer/noy] has quit [Connection timed out]
20090412 01:16:41< Crab_> _ping_  and here's php script that does the simulations - http://crab.terraninfo.net/data/misc/trand.php - and it's source is  http://wesnoth.pastebin.com/m71e036ae   
20090412 01:17:06< Crab_> _ping_: for example, let's take a unit with 1 hp, and 100% chance to hit it.
20090412 01:17:21< Crab_> and lets take a horseman with 2x 30 strikes
20090412 01:17:35< Crab_> what should be the evaluation  ?
20090412 01:17:55< Crab_> current logic will return 1 hp 
20090412 01:18:51< CIA-30> silene * r34741 /trunk/data/campaigns/tutorial/scenarios/2_Speaking.cfg: Fixed incorrect Pango markup.
20090412 01:20:05< _ping_> it was not the example it only when you hit a unit but does not kill it (but where there is still a chance to kill it) that the calc are wrong
20090412 01:20:26-!- gnutoo_ [n=gnutoo@host216-163-dynamic.51-79-r.retail.telecomitalia.it] has joined #wesnoth-dev
20090412 01:20:54< esr> Heads up, everyone: I'm about to change "stoned" to "petrified" everywhere.  "Stoned" is not correcxt the way we use it.
20090412 01:21:23< dfranke> _ping_: you still haven't given a clear explanation of what you want or an answer to my question.  Should or should not the EV be dependent on the outcome?
20090412 01:21:54< Crab_> esr: do not forget about ".stones" )
20090412 01:22:04< esr> I will not.
20090412 01:22:10< Crab_> good )
20090412 01:25:32< Crab_>  _ping_: if we take 10000 36hp-assassins, and charge a horseman into each of them, on average, we will have ~ 1600 dead assassins, 4800 6-hp assassins, and 3600 36-hp assassins
20090412 01:25:51< Crab_> so, how do you propose we should calculate ?
20090412 01:25:57< _ping_> dfranke: okay a horse hits 1 time for 30 hit a 26 hp assain 
20090412 01:27:04< _ping_> the stats read 30/20 that the attack stats are being pushined for the defender not having engouh hit incase it was hit twice
20090412 01:27:21< Crab_> _ping_: it simply means that some of the attacker's power is wasted..
20090412 01:27:21< _ping_> but it is infact only hit once and is not dead
20090412 01:27:26< dfranke> of course they're punished, why wouldn't they be?
20090412 01:27:42< dfranke> if you hit a 1hp unit with a 30hp attack, you're still only doing 1hp of damage.
20090412 01:27:50< _ping_> well since it did not die the stat case is a normal case
20090412 01:28:04< _ping_> and the stats should show 30/ 24
20090412 01:28:30< _ping_> mokay i'm busy now i will return with a more understandable arguement later
20090412 01:28:30< Crab_> _ping_: yes, but the horseman didn't knew that assassin won't die, when he was attacking...
20090412 01:28:57< Crab_>  _ping_, so, he *expected* ~20 avg damage, and, if he hit, he got 30.
20090412 01:29:01< Crab_> so, 30/20
20090412 01:30:06< Crab_> _ping_: maybe you want to see another number - simply "average damage, such as 30*2*0.4" ?
20090412 01:31:00-!- gnutoo_ [n=gnutoo@host216-163-dynamic.51-79-r.retail.telecomitalia.it] has quit [Read error: 60 (Operation timed out)]
20090412 01:31:02< Crab_> this number(24) will have no relation with "expected damage", it is simply "average damage of attacks made, ignoring the defenders hitpoints"
20090412 01:32:04< dfranke> Crab_: before 1.6, though, we did compute it something like that.
20090412 01:32:11< dfranke> Crab_: not exactly though.
20090412 01:32:28< dfranke> Crab_: we calculated EV for each attack rather than for each battle.
20090412 01:32:34< Crab_> dfranke: current way is smarter, imo
20090412 01:32:44< dfranke> Crab_: I agree.
20090412 01:34:42< dfranke> Crab_: one perverse result of the old way is that hitting and then missing the assassin would get a different result than missing and then hitting.
20090412 01:35:32< Crab_> dfranke: truly perverse )
20090412 01:36:24-!- silene [n=plouf@wesnoth/developer/silene] has quit [Read error: 110 (Connection timed out)]
20090412 01:36:37< Crab_> dfranke: but, there's one more way to interpret _ping_'s proposal - make dmg statistics dependent on the result of the battle.
20090412 01:37:08< Crab_> basically, it is 'ignore the defender's hitpoints' IF the defender lives.
20090412 01:37:20< dfranke> Crab_: yes, that's what I've been trying to pin him down on.
20090412 01:37:35< dfranke> brb, getting caffeine.
20090412 01:41:27< Crab_> but, imo, since the purpose of such statistics is to measure the decisions and their outcomes (real vs.expected), we need to compute the expected outcome at the time when final decision to attack is made (just before the attack) - just as it is done now.
20090412 01:47:44< _ping_> (still busy just wanted to add you all really done a great job making wesnoth even more awesome)
20090412 01:48:46< dfranke> _ping_: apropos nothing.... are you the same ping as the server moderator, or is the nick a coincidence?
20090412 01:49:21< _ping_> the same
20090412 01:49:27< Crab_> _ping_, so, just answer: what do you propose 1)  'ignore the defender's hitpoints' IF the defender lives.  2) 'ignore the defender's hitpoints' all the time  3) something else 
20090412 01:49:48< _ping_> 1
20090412 01:50:28-!- loonycyborg [n=sergey@wesnoth/developer/loonycyborg] has quit ["Zzzzzzzzzzzzzzzzzzzzzzzzzzzzzz"]
20090412 01:50:58< Crab_> so, this will give 30/24 if one attacks hit, this will give 0/24 if none of attacks hit, and this will give 36/20 if two attacks hit (and the defender dies) ?
20090412 01:51:11-!- loonybot [n=loonybot@wesnoth/bot/loonybot] has quit [Remote closed the connection]
20090412 01:52:01< dfranke> _ping_: have you read this thread? http://www.wesnoth.org/forum/viewtopic.php?f=12&t=21350&p=298788
20090412 01:52:36-!- cjhopman [n=chris@wesnoth/developer/cjhopman] has joined #wesnoth-dev
20090412 01:54:08< _ping_> yes i talked about that with maboul before he posted
20090412 01:54:43< dfranke> That idea is basically what got implemented somewhere during 1.5.
20090412 01:54:56< dfranke> Do you prefer the old calculations?
20090412 01:56:48-!- Chusslove [n=caslav@brsg-d9bee567.pool.mediaWays.net] has quit [Read error: 110 (Connection timed out)]
20090412 01:57:45-!- PK [n=pk@r74-192-44-206.vctrcmta01.vctatx.tl.dh.suddenlink.net] has joined #wesnoth-dev
20090412 01:58:07< Crab_> dfranke: basically, a problem with current calculations, is that they are rather counter-intuitive - with time, x/y, X will grow to be much greater than Y. so the statistics will not be a intuitive measure of how 'lucky' you are
20090412 01:58:14< _ping_> well i like part of it, the shaman should ofcause have a ev against the horse even that it do get killed first strike, also that you can hit for more than max is okay as well, so my only concern is the case i described
20090412 01:58:26-!- PK [n=pk@r74-192-44-206.vctrcmta01.vctatx.tl.dh.suddenlink.net] has quit [Client Quit]
20090412 01:59:04< dfranke> _ping_: anyway, your bug is invalid since it's working as designed.  But you should probably start a forum thread about it if you have a suggested improvement to the current system.
20090412 01:59:22< _ping_> okay point taken
20090412 01:59:44< _ping_> still maybe see the replay i attach (as soon as i have time)
20090412 02:00:23< _ping_> because as i said there may be another problem (again that could be caused by the bug you linked too)
20090412 02:02:50< dfranke> there's actually one other strange-yet-completely-correct result that I noticed from the current system.
20090412 02:03:07< Crab_> dfranke: which one ?
20090412 02:03:08< dfranke> due to drain, your EV can be negative...
20090412 02:03:28< Crab_> dfranke: it can..
20090412 02:04:16< dfranke> if you end up doing 2 damge with an EV of -4, that gets calculated as -50% damage.
20090412 02:05:04< dfranke> or -150%?  I forget.
20090412 02:05:12< dfranke> anyway, the sign is negative, which looks very strange.
20090412 02:05:23< dfranke> but it's right.
20090412 02:06:00< Crab_> dfranke, "you expected to heal your enemy by attacking but, actually,  you've hurt it"
20090412 02:06:13< Crab_> so the sign is negative)
20090412 02:07:55< Crab_> dfranke: and what with ev 0 ?
20090412 02:08:39< dfranke> Crab_: ooooh...
20090412 02:08:43< dfranke> Crab_: I'd better test that :-)
20090412 02:09:14< dfranke> 0/0 is special-cased, I remember seeing that.
20090412 02:09:22< Crab_> but it can be X/0 0
20090412 02:09:36< Crab_> btw, if "if you end up doing -2 damage with an EV of 4,", that will be -50% damage, too )
20090412 02:09:55< Crab_> but it will be  "you expected to hurt your enemy by attacking but, actually,  you've healed it"
20090412 02:10:46< Crab_> dfranke: x/0 is ok
20090412 02:11:00< Crab_> there's  "(expected == 0 ? 0 : 100 * dsa / expected)"
20090412 02:11:11< dfranke> alright.
20090412 02:11:57< Crab_> ( in statistics_dialog.cpp:198 )
20090412 02:11:57< Crab_> it is just the problem that the stats cannot determine the players intentions )
20090412 02:12:19< Crab_> so, it cannot distinguish  "you expected to heal your enemy by attacking but, actually,  you've hurt it" and "you expected to hurt your enemy by attacking but, actually,  you've healed it"
20090412 02:12:28< Crab_> when calculating %
20090412 02:13:21< Crab_> it can only say that "the player got an opposite of what he expected", hence "minus" sign.
20090412 02:16:27< Polarina> What does "ten-fold" mean?
20090412 02:16:53-!- alink [n=alink@wesnoth/developer/alink] has joined #wesnoth-dev
20090412 02:17:01< Crab_> "multiplied by 10"
20090412 02:17:15< Crab_> Polarina: http://www.thefreedictionary.com/tenfold
20090412 02:17:46< Polarina> Thank you.
20090412 02:17:53< Crab_> np
20090412 02:26:54< Polarina> What does YW mean?
20090412 02:27:08< Crab_> Polarina: in what context/ where ?
20090412 02:27:24< Polarina> Crab_: In the "hints" part of the tile screen.
20090412 02:27:47-!- _ping_ [n=ping_c@56341227.rev.stofanet.dk] has quit ["Leaving"]
20090412 02:28:57< Blueblaze> YW=Years Wesnoth
20090412 02:29:09< Polarina> Thanks.
20090412 02:29:13< Crab_> its a date..
20090412 02:30:02< Blueblaze> YW is like AD and BW is like BC(E)
20090412 02:41:07< esr> The big stoned -> petrified change is committed.
20090412 02:46:24< Crab_> esr: looks good ) the only instance of 'stoning' that remains, hides inside comment )
20090412 02:46:55-!- _ping_ [n=ping_c@56341227.rev.stofanet.dk] has joined #wesnoth-dev
20090412 02:47:10< _ping_> okay dfranke i mailed you two replays
20090412 02:47:36< dfranke> _ping_: k.
20090412 02:59:38< _ping_> okay a very rough calcuation of the stats showing 399 done out of 266 using appromited figures in binormial proberabilities gives me that chance of that happen is less than 0.0000001
20090412 02:59:57< _ping_> anyway i'm off to bed cya
20090412 03:05:11< JW1> _ping_: I love you
20090412 03:05:20< JW1> :)
20090412 03:11:02< CIA-30> esr * r34747 /trunk/ (34 files in 14 dirs): (log message trimmed)
20090412 03:11:02< CIA-30> Change the trannstive verb "stone" (as in "turn to stone") to "petrify", and
20090412 03:11:02< CIA-30> all related words such as "unstone", "stoned", etc. Former usage was not
20090412 03:11:02< CIA-30> really correct and led to unhelpful ambigiuity with both "to stone" is in to hit
20090412 03:11:02< CIA-30> with thrown stones, and "stoned" meaning intoxicated.
20090412 03:11:03< CIA-30> All C++ code, WML, and tags are changed. New wmllint rules will handle all
20090412 03:11:07< CIA-30> lifting for UMC. The wiki has been updated.
20090412 03:14:46-!- gabm [n=Gabriel@72.0.215.2] has joined #wesnoth-dev
20090412 03:18:02< Shadow_Master> what the hell
20090412 03:18:20< Shadow_Master> can't I rest for a single day, without people hilighting me for whatever reasons?
20090412 03:18:57< JW1> Shadow_Master: 
20090412 03:19:15-!- gabm [n=Gabriel@72.0.215.2] has left #wesnoth-dev []
20090412 03:19:45< Shadow_Master> don't test me
20090412 03:20:06< JW1> :)
20090412 03:20:25< JW1> you know you like being popular :p
20090412 03:21:35-!- mode/#wesnoth-dev [+o Shadow_Master] by ChanServ
20090412 03:21:46-!- mode/#wesnoth-dev [-o Shadow_Master] by ChanServ
20090412 03:23:25< Shadow_Master> Mordante: um... why was it inecessary to revert the whole commit?
20090412 03:23:51-!- Sapient [n=sapien-x@c-71-204-10-166.hsd1.ga.comcast.net] has joined #wesnoth-dev
20090412 03:24:11< Shadow_Master> Mordante: ah, you placed #include <iostream> elsewhere... right... sorry about that.
20090412 03:25:31< Sapient> ah, so units can no longer get stoned
20090412 03:25:40< Sapient> esr is the buzz killer
20090412 03:25:57< Shadow_Master> wut?
20090412 03:26:10< Blueblaze> Sapient: Great on so many levels :p
20090412 03:26:33-!- PK [n=pk@r74-192-44-206.vctrcmta01.vctatx.tl.dh.suddenlink.net] has joined #wesnoth-dev
20090412 03:26:40< PK> yo Sapient
20090412 03:26:49< PK> awesome question/idea =o
20090412 03:26:56< Sapient> yo PK
20090412 03:27:14< PK> is it possible to save a units stats/objects/variables into a .cfg in the userdata that could be remotely loaded into a game?
20090412 03:27:31< PK> and be automatically edited by the game.
20090412 03:27:49< PK> via events, of course.
20090412 03:28:13-!- Shadow_Master [n=ignacio@wesnoth/developer/shadowmaster] has left #wesnoth-dev ["need a longer rest"]
20090412 03:28:41< Sapient> so you want to know if WML actions can write arbitrary files on your computer?
20090412 03:28:47< PK> sortof
20090412 03:28:50< Sapient> no, they cannot
20090412 03:29:20< PK> so its not possible to have a unit in one game, moved to a unit in another game when you want it to be?
20090412 03:29:36< PK> like with a "load ..." function
20090412 03:30:15< Sapient> no, but there was someone talking about that idea on the forum earlier
20090412 03:30:21< Sapient> Mr. Markm
20090412 03:30:45< PK> darn, that makes my idea faulty.
20090412 03:31:20< PK> so theres no way to personally store a units data? at all?
20090412 03:31:50< Sapient> you can save it in a variable alright
20090412 03:31:59< Sapient> into a savefile
20090412 03:32:01< PK> but wouldn't that disappear after you exit one game?
20090412 03:32:14< PK> or am I delusional
20090412 03:32:36< Sapient> it will be there when save the game
20090412 03:32:44< Sapient> in the userdata/saves
20090412 03:32:59< PK> I mean, like you quit one game, and load the unit into a NEW game.
20090412 03:33:08< PK> so that you could keep using a unit you've been upgrading for an rpg type thing
20090412 03:33:11< Sapient> technically, you could even load it into another game... (if you know the name of the save)
20090412 03:34:13< Crab_> Sapient: or if you ask the player about that name ?
20090412 03:34:44< Sapient> no, Crab_, you'd have to know it in advance
20090412 03:34:55< Sapient> since the preprocessor is used to insert file contents
20090412 03:35:06< PK> hrm.
20090412 03:35:14< Crab_> Sapient: ok
20090412 03:35:16< PK> could you pull the unit out of that and have it put into a list?
20090412 03:35:36< Sapient> PK: yes, it's possible
20090412 03:36:03< Sapient> but in a networked game, it would only work for files saved on the hosts computer
20090412 03:36:21< PK> on the hosts only?
20090412 03:36:31< Sapient> yes
20090412 03:36:31< PK> hmm
20090412 03:36:39< PK> what about if the other player could send his savefiles?
20090412 03:37:21< Sapient> you mean outside of the game?
20090412 03:37:29< PK> inside.
20090412 03:37:33< Sapient> no
20090412 03:38:03< Sapient> they could host a game, which you enter, then save with a certain name
20090412 03:38:22< Sapient> however, you'd still need to restart wesnoth for the preprocessor to parse it
20090412 03:39:02< PK> okay so, if you played and saved as: CharacterSave1 and it checked for that save specifically, you'd have to restart wesnoth to load that character?
20090412 03:39:44< Sapient> let me give you a hint: this is not a good idea
20090412 03:39:45-!- BenUrban [n=benurban@unaffiliated/benurban] has quit ["Power failu"]
20090412 03:40:09< PK> still be awesome if it worked >>
20090412 03:40:14< Sapient> what you need is support in the game engine for what you are trying to do
20090412 03:40:22< PK> thats unlikely though.
20090412 03:40:30< Sapient> (not an ugly hack)
20090412 03:40:42 * PK likes ugly hacks 8D
20090412 03:42:15< PK> so I'd have to post this in the ideas forum, and hope I don't get shot down by turuk or somebody =D
20090412 03:42:23< Sapient> right
20090412 03:42:29< voris> heh
20090412 03:42:34< PK> but then everybody could do it 3=
20090412 03:42:39< PK> and it wouldn't be as awesome.
20090412 03:44:16-!- Shadow_Master [n=ignacio@wesnoth/developer/shadowmaster] has joined #wesnoth-dev
20090412 03:45:02-!- Shadow_Master [n=ignacio@wesnoth/developer/shadowmaster] has left #wesnoth-dev []
20090412 03:45:26-!- irmorelle [n=ignacio@wesnoth/developer/shadowmaster] has joined #wesnoth-dev
20090412 03:46:01< Sapient> Crab_: you've already started AI refactoring?
20090412 03:46:08< Crab_> Sapient: yes
20090412 03:46:21< Sapient> Crab_: I meant to tell you this before you started, but...
20090412 03:46:29< Crab_> tell now )
20090412 03:46:38< Sapient> please familiarize yourself with all the different unit map iterators
20090412 03:46:53< Sapient> and use the most appropriate one for each situation
20090412 03:47:08< PK> well, I posted >> hope someone agrees with me.
20090412 03:47:50-!- gabm [n=gabriel@72.0.215.2] has joined #wesnoth-dev
20090412 03:48:46< Crab_> Sapient: thanks for the hint
20090412 03:49:03< Sapient> no problem
20090412 03:49:37< Sapient> the unit_map iterators were introduced last year by cjhopman and have not been widely adopted by all parts of the code
20090412 03:49:59< Sapient> in particular, the AI and actions.cpp related code
20090412 03:50:45< Sapient> but in theory, they could help greatly with reducing and identifying errors due to invalid unit iterators
20090412 03:50:56< Sapient> which have been a constant source of error in the past
20090412 03:51:49< Crab_> Sapient: is there any documentation/wiki on those iterators ? or "only source" ?
20090412 03:52:02< Sapient> there is some discussion on the forum
20090412 03:52:13< Sapient> of the motivation behind them
20090412 03:52:22-!- Shadow_Master [n=ignacio@wesnoth/developer/shadowmaster] has joined #wesnoth-dev
20090412 03:52:34-!- Shadow_Master [n=ignacio@wesnoth/developer/shadowmaster] has left #wesnoth-dev []
20090412 03:52:48< Sapient> http://www.wesnoth.org/forum/viewtopic.php?p=236079#p236079
20090412 03:52:56-!- irmorelle [n=ignacio@wesnoth/developer/shadowmaster] has quit []
20090412 03:53:04< Crab_> Sapient: thanks again
20090412 03:53:25< Sapient> the most uptodate documentation should be in unit_map.hpp, though
20090412 03:53:53< Crab_> yes, I've already looked through unit_map.?pp
20090412 03:55:49< dfranke> _ping_: your replay does indeed look really odd.
20090412 03:56:06< Sapient> cya later, have fun.
20090412 03:56:06-!- Sapient [n=sapien-x@wesnoth/developer/sapient] has left #wesnoth-dev []
20090412 03:56:33< dfranke> I want to ask Krimzon what client version he's running but he's not around.  Is it possible to get that info from the server logs?
20090412 04:02:34< nagbot> Hi GSoC Applicants!
20090412 04:02:34< nagbot> SoC decision is near
20090412 04:02:34< nagbot> We are doing some last checks before working on our final "ranking", which we will do *soon*
20090412 04:02:38< nagbot> So, if you got anything to add, post a comment in your proposal
20090412 04:03:12< Crab_> night
20090412 04:03:23-!- Crab_ [n=Crab_@wesnoth/developer/crab] has quit ["Leaving."]
20090412 04:04:58-!- BenUrban [n=benurban@c-68-49-10-243.hsd1.md.comcast.net] has joined #wesnoth-dev
20090412 04:05:52-!- voris [n=voris@c-24-16-107-149.hsd1.wa.comcast.net] has quit [Remote closed the connection]
20090412 04:17:11-!- Tigge_ [n=tigge@bacchus.olf.sgsnet.se] has joined #wesnoth-dev
20090412 04:20:53-!- Netsplit brown.freenode.net <-> irc.freenode.net quits: wesbot, Tigge
20090412 04:23:37-!- voris [n=voris@c-24-16-107-149.hsd1.wa.comcast.net] has joined #wesnoth-dev
20090412 04:25:30-!- Netsplit over, joins: Tigge
20090412 04:25:39-!- wesbot [n=wesbot@86.59.118.153] has joined #wesnoth-dev
20090412 04:26:16-!- Tigge [n=tigge@bacchus.olf.sgsnet.se] has quit [Connection reset by peer]
20090412 04:31:20-!- Netsplit brown.freenode.net <-> irc.freenode.net quits: Mellar, Polarina, crimson_penguin, gabm, johani, corn, nagbot, dfranke, Smar, _ping_,  (+16 more, use /NETSPLIT to show all of them)
20090412 04:35:46-!- lobby_ [n=wesnoth@wesnoth.wesnoth.org] has joined #wesnoth-dev
20090412 04:35:46-!- Topic for #wesnoth-dev: 1.6.1 planned for Sunday, April 12th, stringfreeze for branches/1.6 ! | Summer of Code info: http://www.wesnoth.org/wiki/SummerOfCodeIdeas | 52 bugs, 232 feature requests, 11 patches | logs: http://irclogs.wesnoth.org/ | Don't paste on IRC! Use a pastebin: http://wesnoth.pastebin.com | http://imagebin.org
20090412 04:35:46-!- Topic set by wesbot [] [Sun Apr 12 00:09:12 2009]
20090412 04:35:46[Users #wesnoth-dev]
20090412 04:35:46[ _ping_   ] [ crimson_penguin] [ isaac    ] [ nital     ] [ Smar   ] 
20090412 04:35:46[ AI0867   ] [ dfranke        ] [ Ivanovic ] [ oak_      ] [ Soliton] 
20090412 04:35:46[ alink    ] [ Dragonking     ] [ jmunro[a]] [ Patterner ] [ Tigge_ ] 
20090412 04:35:46[ benetnash] [ erl            ] [ johani   ] [ PK        ] [ voris  ] 
20090412 04:35:46[ BenUrban ] [ esr            ] [ JW1      ] [ Polarina  ] [ wesbot ] 
20090412 04:35:46[ Blueblaze] [ ettin          ] [ lobby    ] [ res|laptop] [ yann   ] 
20090412 04:35:46[ CIA-30   ] [ fendrin        ] [ lobby_   ] [ Rhonda    ] 
20090412 04:35:46[ cjhopman ] [ freim          ] [ Mellar   ] [ shikadibot] 
20090412 04:35:46[ corn     ] [ gabm           ] [ nagbot   ] [ Sirp      ] 
20090412 04:35:46-!- Irssi: #wesnoth-dev: Total of 42 nicks [0 ops, 0 halfops, 0 voices, 42 normal]
20090412 04:35:56-!- Channel #wesnoth-dev created Tue Jan 27 06:28:41 2009
20090412 04:36:08-!- Netsplit grisham.freenode.net <-> irc.freenode.net quits: lobby, jmunro[a], res|laptop
20090412 04:36:36-!- Irssi: Join to #wesnoth-dev was synced in 61 secs
20090412 04:37:24-!- jmunro[a] [i=james@217.112.83.12] has joined #wesnoth-dev
20090412 04:43:14-!- Netsplit over, joins: lobby, res|laptop
20090412 04:43:22-!- lobby [n=wesnoth@wesnoth/bot/lobby] has quit [Read error: 104 (Connection reset by peer)]
20090412 04:44:28-!- AnMaster [n=AnMaster@unaffiliated/anmaster] has joined #wesnoth-dev
20090412 04:47:21-!- Ivanovic_ [n=ivanovic@dtmd-4db2a1d4.pool.einsundeins.de] has joined #wesnoth-dev
20090412 04:50:20< dfranke> hmm, expected damage reported in StatisticsWML is totally out of whack.
20090412 04:50:30< dfranke> This is from the save that _ping_ sent me:
20090412 04:50:53< dfranke> damage_inflicted="902"
20090412 04:50:53< dfranke> damage_taken="560"
20090412 04:50:53< dfranke> expected_damage_inflicted="894124"
20090412 04:50:53< dfranke> expected_damage_taken="444889"
20090412 04:50:57< dfranke> all sides are like that.
20090412 04:51:51< Blueblaze> lol
20090412 04:52:17-!- cjhopman_ [n=chris@68-118-159-7.dhcp.mdsn.wi.charter.com] has joined #wesnoth-dev
20090412 04:54:18-!- Netsplit grisham.freenode.net <-> irc.freenode.net quits: esr, cjhopman, CIA-30, isaac, crimson_penguin
20090412 04:56:55-!- isaac [n=isaac@debian/developer/isaac] has joined #wesnoth-dev
20090412 04:56:55-!- esr [n=chatzill@wesnoth/developer/esr] has joined #wesnoth-dev
20090412 04:56:55-!- CIA-30 [n=CIA@208.69.182.149.simpli.biz] has joined #wesnoth-dev
20090412 04:56:57-!- crimson_penguin [n=ben@wesnoth/developer/crimsonpenguin] has joined #wesnoth-dev
20090412 05:03:42-!- Ivanovic [n=ivanovic@wesnoth/developer/ivanovic] has quit [Read error: 110 (Connection timed out)]
20090412 05:05:20-!- Ivanovic_ is now known as Ivanovic
20090412 05:19:03-!- oak__ [n=oakus@252.108.broadband12.iol.cz] has joined #wesnoth-dev
20090412 05:27:27-!- res|laptop [n=res@pdpc/supporter/student/res2k] has quit [Nick collision from services.]
20090412 05:27:28-!- res2k3 [n=res@erft-4d07da5a.pool.mediaWays.net] has joined #wesnoth-dev
20090412 05:27:48-!- res2k3 is now known as res|laptop
20090412 05:35:05-!- nerwa [n=nerwa@59.78.26.75] has joined #wesnoth-dev
20090412 05:35:45< dfranke> Oh, looks like that's actually normal.  It just multiplies it by 1000.
20090412 05:45:00-!- oak_ [n=oakus@252.108.broadband12.iol.cz] has quit [Read error: 110 (Connection timed out)]
20090412 05:45:43< CIA-30> esr * r34749 /trunk/data/campaigns/ (4 files in 4 dirs): wmllint-inspired cleanups.
20090412 05:47:31-!- nerwa [n=nerwa@59.78.26.75] has quit [Read error: 104 (Connection reset by peer)]
20090412 05:49:20-!- nerwa [n=nerwa@59.78.26.75] has joined #wesnoth-dev
20090412 06:10:07-!- ryochan7 [n=ryochan7@c-98-214-119-19.hsd1.il.comcast.net] has joined #wesnoth-dev
20090412 06:19:33-!- nerwa_ [n=nerwa@59.78.26.75] has joined #wesnoth-dev
20090412 06:28:09< CIA-30> alink * r34750 /branches/1.6/src/ (multiplayer_connect.cpp multiplayer_wait.cpp): 
20090412 06:28:09< CIA-30> Fix bug #13243 (missing faction name in MP wait UI) in 1.6
20090412 06:28:09< CIA-30> but only when host and client are 1.6.1 (1.6.0 does not send faction's name)
20090412 06:28:09< CIA-30> This introduce a new key 'faction_name'. Not sure if we must hide it for UMC,
20090412 06:28:09< CIA-30> so I add the hiding part in the next commit for the moment.
20090412 06:28:18< Blueblaze> dfranke: Why would it multiply it by 1000?
20090412 06:29:55-!- Lord_Aether [n=castle@206.170.190.49] has joined #wesnoth-dev
20090412 06:32:52< CIA-30> alink * r34751 /branches/1.6/src/ (multiplayer_connect.cpp multiplayer_connect.hpp): 
20090412 06:32:53< CIA-30> Hide the new 'faction_name' key (now only used in the choose-faction part)
20090412 06:32:53< CIA-30> This will prevent UMC to start using it (and thus create addon incompatible
20090412 06:32:53< CIA-30> with 1.6.0). Small change but a bit hackish, maybe safer to replace this by
20090412 06:32:53< CIA-30> a simple warning in the wiki ?
20090412 06:33:38< alink> ^really not sure if needed and I would feel slightly safer without adding this code
20090412 06:38:33< alink> does replace that by writing in wiki: "faction_name : name of the faction, blank if host uses 1.6.0"  is too confusing ?
20090412 06:39:08< alink> btw this will also restore a similar WML feature killed by the same bug
20090412 06:41:20-!- nerwa [n=nerwa@59.78.26.75] has quit [Read error: 110 (Connection timed out)]
20090412 06:42:20< alink> Ivanovic: not sure when you will tag, but I would appreciate some other opinion about this
20090412 06:42:39-!- crimson_penguin [n=ben@wesnoth/developer/crimsonpenguin] has quit ["Leaving"]
20090412 06:43:34< alink> Ivanovic: btw it seems that boucman didn't commit yet the framerate optimization, so also need to wait that
20090412 06:44:38< dfranke> _ping_: ok, so, I tested your replays...
20090412 06:45:18< dfranke> _ping_: when I loaded the savegames, using 1.6+svn, the statistics were out of whack as you say.
20090412 06:46:18< dfranke> _ping_: however, when I play through them from the beginning as replays, this causes the statistics to be recomputed rather than loaded from the stored values in the save.  Doing this, I got wrong results in 1.5.14, but in 1.6+svn they were fine.
20090412 06:46:35< dfranke> _ping_: so whatever bug caused this has already been fixed.
20090412 06:48:59< dfranke> While I was at it I also extracted each attack from the save file and did a χ² regression analysis on what hit and what didn't.  That came back unremarkable.
20090412 06:49:14< dfranke> (So your RNG was fine; just the statskeeping was broken)
20090412 06:55:55-!- nerwa_ [n=nerwa@59.78.26.75] has quit [Read error: 110 (Connection timed out)]
20090412 06:58:31-!- gabm [n=gabriel@72.0.215.2] has left #wesnoth-dev []
20090412 06:59:19-!- nerwa [n=nerwa@59.78.26.75] has joined #wesnoth-dev
20090412 06:59:47< CIA-30> alink * r34752 /branches/1.6/ (changelog players_changelog): update changelogs
20090412 07:03:32-!- boucman [n=rosen@wesnoth/developer/boucman] has joined #wesnoth-dev
20090412 07:12:40-!- Lord_Aether [n=castle@206.170.190.49] has quit []
20090412 07:13:55-!- PK [n=pk@r74-192-44-206.vctrcmta01.vctatx.tl.dh.suddenlink.net] has left #wesnoth-dev []
20090412 07:18:40-!- ikarius [n=ross@216.27.182.3] has joined #wesnoth-dev
20090412 07:32:00< alink> I still hesitate if we must port 34430 to 1.6. I know that it sometimes affect rating of distant target, but I can't tell the impact on the AI
20090412 07:32:10< alink> wesbot log 34430
20090412 07:32:13< wesbot> alink * r34430 : Fix a possibly serious breakage of the AI:
20090412 07:32:19< wesbot> URL: http://svn.gna.org/viewcvs/wesnoth?view=rev&rev=34430
20090412 07:39:01-!- lobby [n=wesnoth@wesnoth/bot/lobby] has joined #wesnoth-dev
20090412 07:39:01-!- Topic for #wesnoth-dev: 1.6.1 planned for Sunday, April 12th, stringfreeze for branches/1.6 ! | Summer of Code info: http://www.wesnoth.org/wiki/SummerOfCodeIdeas | 52 bugs, 232 feature requests, 11 patches | logs: http://irclogs.wesnoth.org/ | Don't paste on IRC! Use a pastebin: http://wesnoth.pastebin.com | http://imagebin.org
20090412 07:39:01-!- Topic set by wesbot [] [Sun Apr 12 00:09:12 2009]
20090412 07:39:01[Users #wesnoth-dev]
20090412 07:39:01[ _ping_   ] [ cjhopman_ ] [ isaac    ] [ nital     ] [ Smar   ] 
20090412 07:39:01[ AI0867   ] [ corn      ] [ Ivanovic ] [ oak__     ] [ Soliton] 
20090412 07:39:01[ alink    ] [ dfranke   ] [ jmunro[a]] [ Patterner ] [ Tigge_ ] 
20090412 07:39:01[ AnMaster ] [ Dragonking] [ johani   ] [ Polarina  ] [ voris  ] 
20090412 07:39:01[ benetnash] [ erl       ] [ JW1      ] [ res|laptop] [ wesbot ] 
20090412 07:39:01[ BenUrban ] [ esr       ] [ lobby    ] [ Rhonda    ] [ yann   ] 
20090412 07:39:01[ Blueblaze] [ ettin     ] [ Mellar   ] [ ryochan7  ] 
20090412 07:39:01[ boucman  ] [ fendrin   ] [ nagbot   ] [ shikadibot] 
20090412 07:39:01[ CIA-30   ] [ freim     ] [ nerwa    ] [ Sirp      ] 
20090412 07:39:01-!- Irssi: #wesnoth-dev: Total of 42 nicks [0 ops, 0 halfops, 0 voices, 42 normal]
20090412 07:39:04< alink> there is also this bad "cur_route.move_left = 0;" in ai_move which mitigate the problem, but it's not always set to 0, so sometimes there is a difference
20090412 07:39:08-!- Channel #wesnoth-dev created Tue Jan 27 06:28:41 2009
20090412 07:39:41-!- Irssi: Join to #wesnoth-dev was synced in 48 secs
20090412 07:45:21< alink> of course I could blindly commit it in 1.6 because i am confident that it will fix some AI errors, but the fact that i don't know what it fixes make me hesitate
20090412 07:47:16< boucman> alink: if it fixes stuff, that's fine, the question is about the risk of breaking new stuff...
20090412 07:48:10< alink> boucman: yeah, and the fact that i don't know how to test AI things doesn't help
20090412 07:48:26< boucman> ask crab :P
20090412 07:48:35< boucman> more seriously, though
20090412 07:48:58< boucman> we know that we have no way to really test the AI yet, 
20090412 07:49:09< boucman> so we have to use our best guess now
20090412 07:49:12< alink> well, crab didn't worked on the late 1.6 AI bugs
20090412 07:49:48-!- Netsplit niven.freenode.net <-> irc.freenode.net quits: erl, shikadibot, ettin, corn, cjhopman_, nagbot, Blueblaze, dfranke, Patterner, Mellar
20090412 07:50:04-!- Netsplit over, joins: Blueblaze, cjhopman_, Patterner, nagbot, ettin, corn, shikadibot, erl, dfranke, Mellar
20090412 07:51:31< alink> Sirp: maybe I can interest you to look if we must port r34430 to 1.6, IIRC you fixed some AI bugs in 1.6 and maybe it could explain some odd behavior
20090412 07:56:02< alink> boucman: btw, you will still commit the in_hex stuff to 1.6 ?
20090412 07:56:17< boucman> I already did
20090412 07:57:05< alink> ah ok sorry, didn't saw it
20090412 07:57:08-!- nerwa [n=nerwa@59.78.26.75] has quit []
20090412 07:57:16< boucman> don't worry, I had forgoten too :)
20090412 08:00:19< alink> [07:53:37] <boucman> I already did    <- hum, i don't see in 1.6
20090412 08:00:44< boucman> alink: ok, I hadn't commited
20090412 08:00:47< boucman> sorry about that
20090412 08:01:31< boucman> done
20090412 08:01:40< CIA-30> boucman * r34753 /branches/1.6/ (9 files in 2 dirs): add the in_hex optimization to 1.6, I thought I had already commited, sorry about that
20090412 08:02:50< alink> boucman: thanks
20090412 08:03:45-!- Synergizer [n=synergiz@c-75-73-225-141.hsd1.mn.comcast.net] has joined #wesnoth-dev
20090412 08:12:57-!- ryochan7 [n=ryochan7@c-98-214-119-19.hsd1.il.comcast.net] has quit ["Leaving"]
20090412 08:14:24-!- mordante [n=mordante@wesnoth/developer/mordante] has joined #wesnoth-dev
20090412 08:14:36< mordante> hi
20090412 08:14:43< alink> hello mordante
20090412 08:14:52< mordante> hi alink 
20090412 08:15:43< CIA-30> esr * r34754 /trunk/data/campaigns/Delfadors_Memoirs/scenarios/12_terror_at_the_ford_of_parthyn.cfg: Fix a broken wakeup event.
20090412 08:18:15< alink> boucman: about r34753 the difference with trunk are intentional ? (for example trunk doesn't use hex_need_update anymore, which seems to make sense)
20090412 08:24:23< boucman> alink: i'm a bit in a hurry, could you point me to the line ?
20090412 08:24:25< mordante> esr, regarding your stone commit, I don't like it breaks all old savegames. 
20090412 08:24:40< mordante> esr, 1 the reason to break it is a trivial change
20090412 08:24:52< mordante> esr, 2 wmllint can't convert savegames
20090412 08:25:16< mordante> esr, 3 can break debugging, eg caves of the basilik will be broken
20090412 08:25:38< mordante> esr, 4 it breaks with the 2 version compatible rule
20090412 08:25:40< alink> boucman: grep hex_need_update 
20090412 08:25:55< alink> or in trunk unit_frame.cpp ~522
20090412 08:26:34< mordante> esr, so please add a fix for the load code to read stoned as petrified and give a deprecated warning about it
20090412 08:26:58< esr> mordante: I guess I can do that.
20090412 08:27:01< mordante> esr, and IMO this warning should stay > 2 releases,
20090412 08:27:22< mordante> esr, and when the code gets removed, the min savegame version should be increased as well
20090412 08:27:26< alink> boucman: mainly visible at 507 in 1.6
20090412 08:27:28< boucman> alink: I need to run, i'll let you look into it, sorry about that
20090412 08:27:58< alink> boucman: ok i check what is different, seems not big deal, I was just asking
20090412 08:29:06< mordante> thanks esr 
20090412 08:30:25< alink> boucman: ah i see you moved up image_fix_hex to avoid the image search
20090412 08:31:48< alink> wesbot log 34212
20090412 08:31:49< wesbot> boucman * r34212 : avoid useless search of image in cache when invalidating a unit that fits in the hex
20090412 08:31:52< wesbot> URL: http://svn.gna.org/viewcvs/wesnoth?view=rev&rev=34212
20090412 08:32:14-!- gnutoo_ [n=gnutoo@host216-163-dynamic.51-79-r.retail.telecomitalia.it] has joined #wesnoth-dev
20090412 08:33:09< mordante> esr, regarding commit 34754 I see a line "#po: "be of heart" -> "have courage" in modern English." does this show in in the comment in po files?
20090412 08:34:23< esr> mordante: If you're using my new Python wmlgettext, yes.  But Ivanovic has not transitioned to that yet.  #po: is treated as a comment to be passed through into the .po.
20090412 08:36:02< mordante> esr, cool feature :-) I can remember 1 translation fix which I did, which would have been avoided by it :-)
20090412 08:37:36< esr> mordante: I wrote it out of compassion for the poor bastarfds who have to translare all the archaisms and dialect words I like to use. At leat they should get a hint... :-)
20090412 08:43:20< CIA-30> mordante * r34755 /trunk/data/campaigns/Northern_Rebirth/scenarios/01_Breaking_the_Chains.cfg: 
20090412 08:43:20< CIA-30> Add a translator hint.
20090412 08:43:20< CIA-30> esr's new wmlgettext can give hints to translators, so added one which
20090412 08:43:20< CIA-30> determines the gender of the tyrant.
20090412 08:43:44< mordante> esr, not only useful for archaisms :-)
20090412 08:43:51< esr> Indeed.
20090412 08:44:42< mordante> the German translation assumed the tyrant was a male
20090412 09:00:49-!- Sirp [n=me@wesnoth/developer/dave] has quit ["leaving"]
20090412 09:00:56-!- silene [n=plouf@ASte-Genev-Bois-152-1-31-188.w83-114.abo.wanadoo.fr] has joined #wesnoth-dev
20090412 09:02:34< CIA-30> esr * r34756 /trunk/src/unit.cpp: Backward-compattibility hack to accept "stoned" in savefiles.
20090412 09:02:34< nagbot> Hi GSoC Applicants!
20090412 09:02:34< nagbot> Time of SoC decision is nearing
20090412 09:02:34< nagbot> We are doing some last checks before working on our final "ranking", which we will do *soon*
20090412 09:02:38< nagbot> So, if you want to add anything, post a comment in your proposal
20090412 09:03:26< silene> hi
20090412 09:03:36< mordante> hi silene 
20090412 09:04:28< mordante> I just see mist seems to be on the forums again... 
20090412 09:04:34< esr> hi silene.  Any chance you'll untangle  #13294 	(No localedef) in the near future?
20090412 09:05:38< mordante> thanks esr 
20090412 09:07:20< silene> esr: i had forgot about it; i can just uncomment the code suokko commented, at least it will avoid the issue in the default setup
20090412 09:08:54< esr> That looked like it would work.  I'm just nudging because nudging people to clear bugs seems to be part of my job.
20090412 09:12:13< silene> that being said, his patch was also touching cmake and scons and i have no clue what he changed there
20090412 09:12:29< esr> Aaargh.
20090412 09:12:45< esr> Well, having reminded you, I need to go to sleep.
20090412 09:12:51< esr> Goodnight, all.
20090412 09:13:18< mordante> night esr 
20090412 09:20:23< CIA-30> silene * r34757 /trunk/Makefile.am: 
20090412 09:20:23< CIA-30> Reverted part of Suokko's patch 28786. (Fix for bug #13294, no localedef.)
20090412 09:20:23< CIA-30> That patch was also touching scons and cmake, I have no clue what he did there, so please review if it needs reverting there too.
20090412 09:23:06< mordante> I'll have a look at cmake
20090412 09:23:27< CIA-30> silene * r34758 /branches/1.6/Makefile.am: 
20090412 09:23:27< CIA-30> Backport of commit 34757 from trunk.
20090412 09:23:27< CIA-30> Reverted part of Suokko's patch 28786. (Fix for bug #13294, no localedef.)
20090412 09:23:57-!- Chusslove [n=caslav@brsg-d9bef6c4.pool.mediaWays.net] has joined #wesnoth-dev
20090412 09:24:39-!- Chusslove [n=caslav@brsg-d9bef6c4.pool.mediaWays.net] has left #wesnoth-dev ["Kopete 0.12.7 : http://kopete.kde.org"]
20090412 09:24:42-!- Chusslove [n=caslav@brsg-d9bef6c4.pool.mediaWays.net] has joined #wesnoth-dev
20090412 09:25:32-!- stikonas [n=stikonas@ctv-79-132-187-171.vinita.lt] has joined #wesnoth-dev
20090412 09:41:09-!- Synergizer [n=synergiz@c-75-73-225-141.hsd1.mn.comcast.net] has quit ["Leaving"]
20090412 09:43:51-!- busfahrer [n=busfahre@unixboard/user/busfahrer] has joined #wesnoth-dev
20090412 09:47:22-!- Synergizer [n=synergiz@c-75-73-225-141.hsd1.mn.comcast.net] has joined #wesnoth-dev
20090412 09:47:53< mordante> silene, just curious about r34757, but having dummy locales always generated shouldn't break a thing or am I missing something?
20090412 09:52:41-!- oak__ [n=oakus@252.108.broadband12.iol.cz] has quit ["( www.nnscript.de :: NoNameScript 4.02 :: www.XLhost.de )"]
20090412 10:05:04-!- noy [n=Noy@d75-157-52-251.bchsia.telus.net] has joined #wesnoth-dev
20090412 10:08:39< alink> boucman: sorry to notice that only now, but i see a serious problem with the change done in r34116 compared to my patch 1134
20090412 10:08:44< silene> mordante: it breaks because i have no clue how to do it except on my own computer; in particular, the bug report was about bsd not having localedef; i guess windows doesn't have it either; if someone finds a portable way to generate locales, that would be good; but in the meantime, no need to break builds
20090412 10:09:22< mordante> ok thanks
20090412 10:09:34< alink> boucman: seems that we misunderstood each other when i mentioned the debug output is_in_hex in image.cpp
20090412 10:10:14< alink> boucman: you tested it and told "alink: works like a charm, and loads images at first unit creation"
20090412 10:10:22< mordante> silene, shouldn't the --dummy-locales option for the game then also depend again on -DUSE_DUMMYLOCALES?
20090412 10:10:42< silene> mordante: yes, that's precisely what i'm implementing right now ;-)
20090412 10:11:08< mordante> ah ok, then my confusion is gone :-)
20090412 10:12:39< alink> boucman: which I assumed was like my patch, scan one (or maybe few) image by unit_type (when creating an unit of this type). But now i see thousand of image scans, and often the same ones. Maybe all the default anims or something like that
20090412 10:14:07< alink> boucman: this cause a load of a lot of images and the multiple scan of the same image seems really inefficient
20090412 10:15:27< alink> boucman: again sorry to didn't retest the debug code earlier, I really understood that the result was the same
20090412 10:16:00< Ivanovic> moin
20090412 10:16:10< alink> hello Ivanovic
20090412 10:18:32< mordante> hi Ivanovic 
20090412 10:21:27< Dragonking> Good morning everyone
20090412 10:21:45< mordante> hi Dragonking 
20090412 10:32:49-!- Newell [i=d31ab04c@gateway/web/ajax/mibbit.com/x-a6a978482ca8b7fb] has joined #wesnoth-dev
20090412 10:36:35-!- noy [n=Noy@wesnoth/developer/noy] has quit []
20090412 10:37:04-!- loonybot [n=loonybot@79.139.246.210] has joined #wesnoth-dev
20090412 10:37:49-!- loonycyborg [n=sergey@79.139.246.210] has joined #wesnoth-dev
20090412 10:39:21< CIA-30> silene * r34759 /trunk/src/ (game.cpp language.cpp): Made --dummy-locales option actually depend on support for dummy locales. Removed deprecated option --dummylocales.
20090412 10:44:26< CIA-30> silene * r34760 /branches/1.6/src/ (game.cpp language.cpp): Made --dummy-locales option actually depend on support for dummy locales.
20090412 10:44:58-!- Reisiger [n=Reisiger@adsl-84-226-76-63.adslplus.ch] has joined #wesnoth-dev
20090412 10:44:59-!- Newell [i=d31ab04c@gateway/web/ajax/mibbit.com/x-a6a978482ca8b7fb] has quit ["http://www.mibbit.com ajax IRC Client"]
20090412 11:02:33< mordante> loonycyborg, can you test this patch on Windows to see whether it fixes the gui2 error? http://paste.debian.net/33257/
20090412 11:04:10< loonycyborg> silene: I hope that you understand that suokko wanted to remove dummy-locales option from buildsystems and make it runtime option instead?
20090412 11:05:37< loonycyborg> Why not just make autotools check for localedef and make/install dummy locales if it's available?
20090412 11:07:34-!- zookeeper [n=l@wesnoth/developer/zookeeper] has joined #wesnoth-dev
20090412 11:09:32-!- Blueblaze [n=nick@c-98-199-143-139.hsd1.tx.comcast.net] has quit [Remote closed the connection]
20090412 11:10:49-!- Elvish_Pillager [n=eli@66-189-40-122.dhcp.oxfr.ma.charter.com] has joined #wesnoth-dev
20090412 11:12:05< loonycyborg> mordante: You mean the error that I previously reproduced on wine and windows? Sure.
20090412 11:12:44< mordante> loonycyborg, yes I still have no idea what goes wrong and I found some hints that setting the height might invalidate the ellipse settings
20090412 11:12:56< mordante> and a bugfix but that was in 1.20.5
20090412 11:13:11< mordante> already in*
20090412 11:15:12< _ping_> dfranke: well that was what i wrote (that is suspected the stats not the rng) anyway gald to see it is fixed, i can ask krimzon next time i see him what version he is on, also knowing what cause the error would be good (since that way we whould be sure it fxed not a random occuring errror)
20090412 11:16:23< CIA-30> fendrin * r34761 /trunk/po/wesnoth-low/de.po: Updated the german translation of LoW with a patch from Crommy.
20090412 11:17:50< dfranke> _ping_: don't need to know about krimzon now.  that would only be an issue if it were related to droid sides.
20090412 11:17:57< loonycyborg> silene: I really hope that you'll find a way to avoid requiring -DUSE_DUMMYLOCALES. Such conditional code should be IMO avoided if possible.
20090412 11:18:56< loonycyborg> Since compile errors in disabled code segments will go unnoticed unless you explicitly test all such options.
20090412 11:20:50< _ping_> dfranke: i came up with an eksample of why i think the current calculation i a bit odd, If an horseman hit a 31 hp assaissan (on 60%) once the stats will read (0.16*31 + 0.48 *30) = 19.36 if the same horse hit a 60 pionts assassin once the stats will read 30/24 so as it is now i'm more lucky to hit the first assassian?, in other words the same out come result in two different stat readings
20090412 11:22:25< dfranke> _ping_: no... stats are counted for the battle, so if the horse dies from retaliation after the first hit, he'll do poorly relative to EV.
20090412 11:23:01< _ping_> dfranke: in my ekample the horseman have no chance of dying
20090412 11:23:50< Chusslove> Ivanovic: Would it be possible to send yet another translation update now?
20090412 11:24:05< dfranke> _ping_: not certain I follow you then, but I'll get back to you in a bit.
20090412 11:25:14-!- stikonas [n=stikonas@wesnoth/translator/stikonas] has quit [Remote closed the connection]
20090412 11:25:30< silene> loonycyborg: the conditional in language.cpp is just there to reduce compile-time / binary-size; if you feel it is not worth it, just remove it
20090412 11:26:07< _ping_> okay put it this way if i hit and 100000 hp assassin (once)  i will be less lucky than if i hit and 31 hp assissain according to how stats are calculated right now (horse in both cases have for sake of the eksample lots of hp
20090412 11:28:32< Dragonking> Wow. Formula AI attack candidate move ignores teams and can make AI attack allies and kill them...
20090412 11:29:52< silene> _ping_: it doesn't have to do anything with luck, it all has to do with efficiency; the ev for the 31hp assassin is smaller than the ev for the 1mhp assassin because you are wasting your horseman on the first one
20090412 11:30:37< _ping_> silene: please re-read there was no 1hp assassion 
20090412 11:30:55-!- stikonas [n=stikonas@wesnoth/translator/stikonas] has joined #wesnoth-dev
20090412 11:30:56< silene> _ping_: come on, i wrote 1mhp, that's one million hp...
20090412 11:32:12< _ping_> if i hit an assiassin once and the result does not cuase the assassion to die would you say i'm more or less lucky if the hitted assassion had 31 hp or 100 hp?
20090412 11:33:05< silene> _ping_: it doesn't have anything to do with luck!!! the average damage is all about efficiency; you really have no clue about statistics, do you?
20090412 11:36:15< Ivanovic> Chusslove: don't ask, just send
20090412 11:36:26< Ivanovic> when you are too late, it is just in the next release
20090412 11:36:47< _ping_> silene: i dont think you get it
20090412 11:36:47< Ivanovic> (i am not this much around right now, going to visit relatives and all this, you know, it is easter sunday...)
20090412 11:36:50-!- Crab_ [n=Crab_@wesnoth/developer/crab] has joined #wesnoth-dev
20090412 11:36:54< Crab_> hi
20090412 11:36:57< Chusslove> Oh, right :)
20090412 11:38:19< Dragonking> hi Crab_ 
20090412 11:38:25< Crab_> hi Dragonking
20090412 11:38:26< CIA-30> dragonking * r34762 /branches/1.6/src/ (formula_ai.cpp formula_ai.hpp): Fixed bug that made 'attack' candidate move attack and kill allied units.
20090412 11:38:32< mordante> hi Crab_ 
20090412 11:38:38< Crab_> hi mordante
20090412 11:39:10< dfranke> Dragonking: nice bug :-)
20090412 11:39:18< Dragonking> Yes, amazing one. :)
20090412 11:40:59< Crab_> dragonking: a good bug )
20090412 11:41:44< Dragonking> Crab_: I think you forgot to add svn properties to ai_manager.[ch]pp
20090412 11:42:17< Crab_> Dragonking: what properties should I add ?
20090412 11:42:29< Crab_> Dragonking: that's why I want to make the ai read-only in regards the game state (except a small and well-defined interface) - because the game doesn't  allow human player to attack allied units, why should ai be allowed to do that ?)
20090412 11:42:36< Dragonking> Crab_: svn proplist -v formula_ai.cpp
20090412 11:42:37< mordante> Crab_, do you use svn or git?
20090412 11:42:42< Crab_> mordante: git
20090412 11:43:08-!- Chusslove [n=caslav@brsg-d9bef6c4.pool.mediaWays.net] has left #wesnoth-dev ["Kopete 0.12.7 : http://kopete.kde.org"]
20090412 11:43:09< Crab_> mordante: but I've got an svn copy lurking around, too
20090412 11:43:10< mordante> I think it's not possible to set properties with git :-(
20090412 11:43:23< Dragonking> Oh.
20090412 11:43:33< mordante> I normally use svn when I add new files so it automatically sets the properties
20090412 11:43:43< loonycyborg> That's why I still have a trunk svn checkout :P
20090412 11:44:07< Dragonking> Crab_: Yes... good idea.
20090412 11:44:54< Crab_> Dragonking: ok, I'll set those properties for ai_manager and ai_actions
20090412 11:45:03< Dragonking> Great :)
20090412 11:45:38< mordante> Crab_, when using svn you can set http://paste.debian.net/33263/ to ~/.subversion/config
20090412 11:46:22< Crab_> thanks. so, if I add new files with svn, they'll be auto-set ?
20090412 11:46:35< mordante> yes for .cpp .hpp and .cfg files
20090412 11:47:18< _ping_> silene: we are using applied statistics which is about predicting the outcome of events and our model is right now predicting two differnt statistics for the "same" event, moreover this is not "exact statictics" (we are not proving theorems). 
20090412 11:48:15< Crab_> _ping_: that is ok. the knowlegge was not the same.
20090412 11:49:19< Reisiger> Good morning everyone :)
20090412 11:49:20< Crab_> _ping_: if we are attacking 1000000 hp assassin, we *know* that he will not die and we will not waste any of our dmg. if we are attacking 36-hp assassin, we do not know whether he will die (in this case we would waste dmg) or not
20090412 11:50:09< _ping_> well again i'm not saying the calc based on current model is wrong, i'm saying (just as maboul made an arguement of why the old model there was some "strange" outcomes) this model have "strange" outcomes
20090412 11:50:13< mordante> hi Reisiger 
20090412 11:50:43< Reisiger> Great... I broke actions.cpp ... back in a bit with a cup of coffee :)
20090412 11:50:47< loonycyborg> mordante: It still crashes under wine with your patch.
20090412 11:50:55< Crab_>  _ping_: can you send that replay to me as well ? ( Crab on forums)
20090412 11:51:32< _ping_> Crab_: in my eksample i have two different stats readngs (30/19 and 30/24) for the even of a horse hitting an assassing without killing it
20090412 11:51:47< mordante> loonycyborg, :-( would it be possible to test with the lastest pango, or did I already ask that?
20090412 11:52:01< _ping_> to me that makes the model we use odd
20090412 11:52:12< _ping_> Crab_: your mail?
20090412 11:52:20< _ping_> oh as PM
20090412 11:52:22< Crab_> by forum pm
20090412 11:53:34< loonycyborg> I alreadyuse pango 1.24 in cross build.
20090412 11:54:53< mordante> :-(
20090412 11:55:56< mordante> I'm a bit out of options then, guess I need to try to create a stand-alone test case so we can see whether it's me or pango
20090412 11:56:46< loonycyborg> Though I may also test it on native in case it's a wine bug that just coincided to end up in the same place :P
20090412 11:57:40< mordante> it would be nice to test it (I however have little hope it will fix it)
20090412 11:58:05< mordante> but if it does it would save me quite some bug hunting
20090412 11:58:09< silene> _ping_: this is not the same event, the amount of hitpoints changes between your two examples; and the game engine also computes exact statistics: they are obtained by dynamic programming and the model we use is mathematically proved (up to rounding errors and implementation bugs obviously)
20090412 12:00:36< CIA-30> fendrin * r34763 /trunk/po/wesnoth-low/de.po: Reverted my commit 34761. It broke the translation process.
20090412 12:02:35< _ping_> silene: yes again it is choosing a different approch i want
20090412 12:04:00< silene> _ping_: so what? the game computes the exact same damage that what would be obtained by running thousands of simulations; yours would not, so it doesn't interest us
20090412 12:05:00< _ping_> silene: easy now
20090412 12:09:12-!- 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, 232 feature requests, 11 patches | logs: http://irclogs.wesnoth.org/ | Don't paste on IRC! Use a pastebin: http://wesnoth.pastebin.com | http://imagebin.org
20090412 12:12:54-!- cyber_11 [n=muratcan@78.165.99.242] has joined #wesnoth-dev
20090412 12:15:46< _ping_> i just want the stats to be more according to how they are used by players as a messaure of their "luck". I hit an 31 hp assassion with one hit for 30 hp was i really more lucky than hitting an 60 hp assassion once? (notice it the same kind of arguement that change the model in first place)
20090412 12:16:17< CIA-30> loonycyborg * r34764 /trunk/ (configure.ac src/game.cpp src/language.cpp): 
20090412 12:16:17< CIA-30> Removed conditional compilation for dummy_locales option
20090412 12:16:17< CIA-30> Since it doesn't save much binary size/compile time but adds a
20090412 12:16:17< CIA-30> possibility of uncaught compile errors unless both values of the options
20090412 12:16:17< CIA-30> are regularly tested.
20090412 12:17:06< zookeeper> boucman, i'm planning on applying the WC death anim patch before tagging (assuming it's in >6h). just saying in case anyone starts worrying about it..
20090412 12:20:18< Crab_> _ping_: yes, in fact, you were more lucky. because you've *avoided* a wasteful hit-the-assassin twice outcome.
20090412 12:21:48< _ping_> Crab_: I you present the expected result as in chance for an event to occur you where not more lucky
20090412 12:22:15< _ping_> the chance of me hitting once is ofcause the same
20090412 12:24:16-!- zookeeper [n=l@wesnoth/developer/zookeeper] has quit []
20090412 12:26:41< Crab_> _ping_:  it all comes down to a moment of "expectation". If we *expect* before the attacking (when we make the decision to attack but not yet know the outcome), then "hit 36-hp assassin once" is more lucky than "hit 60-hp assassin once". 
20090412 12:26:50< Crab_> _ping_: see the following simple example:
20090412 12:27:50< Crab_> situation A: get 100$ with 50% probability or get 0$ with 50% probability.
20090412 12:27:50< Crab_> situation B: get 100$ with 50% probability or get 200$ with 50% probability.
20090412 12:27:59-!- Elvish_Pillage2 [n=eli@66-189-40-122.dhcp.oxfr.ma.charter.com] has joined #wesnoth-dev
20090412 12:28:06< Crab_> both A and B have "get 100$ with 50% probability" part
20090412 12:28:34< Crab_> but in fact, person who draws 100$ in A is lucky, and person who gets 100$ in B is not lucky at all.
20090412 12:28:54< Crab_> the chance of getting 100$ is, of course, the same )
20090412 12:29:11-!- Elvish_Pillager [n=eli@66-189-40-122.dhcp.oxfr.ma.charter.com] has quit [Read error: 110 (Connection timed out)]
20090412 12:29:18< _ping_> Crab_: yes i agree that, that is what we model i just think it is not the right way to do it
20090412 12:29:25< silene> loonycyborg: your patch is bad, there is no way to detect whether the --dummy-locales is meaningful now, you shouldn't have removed the macro from game.cpp
20090412 12:31:17< Crab_> _ping_: what is the right way ? current way *is* the measure of luck, and it is the accurate measure of expected value.
20090412 12:31:55< silene> loonycyborg: in particular, if the user uses the option now, all the languages are activated but once they are used it will break bad
20090412 12:31:59< Crab_> _ping_: propose another interesting way )
20090412 12:32:08< _ping_> Crab_: to me the current way is not the messure of luck  :)
20090412 12:32:18< mordante> silene, loonycyborg what are we going to do with the dummy locales? I've a cmake patch pending.
20090412 12:32:18< Crab_> _ping_: yes, I understand )
20090412 12:33:30< Crab_> situation A: get 100$ with 50% probability or get 0$ with 50% probability.
20090412 12:33:30< Crab_> situation B: get 100$ with 50% probability or get 200$ with 50% probability.
20090412 12:33:30< Crab_> current way will give A100 = 100/50,     A0 = 0/50,       B100 = 100/150,             B200 = 200/150
20090412 12:33:32-!- stikonas [n=stikonas@wesnoth/translator/stikonas] has quit [Read error: 104 (Connection reset by peer)]
20090412 12:33:34< silene> mordante: as far as i'm concerned, i would revert loonycyborg's patch (except for the language.cpp part)
20090412 12:34:17< _ping_> Crab_: yes either it could be present as % chance of an event to occur but hat migth be too hard for some to read, or it could be present as # of hit vs # of hit expected
20090412 12:35:55< Crab_> _ping_: # of hit vs # of hit expected is interesting..
20090412 12:36:26< loonycyborg> silene: Revert language.cpp part too since it causes problems on windows :P
20090412 12:37:05< CIA-30> fendrin * r34765 /branches/1.6/po/wesnoth-low/de.po: Updated LoW's german translation with a patch from Crommy.
20090412 12:37:46< Crab_> _ping_: if the archmage does 4 actual 70% attacks to target, and 3 of them hit, then we'll have 3/ 4*0.7 =  3/2.8   ? 
20090412 12:38:28< Crab_> _ping_: like this ?
20090412 12:38:37< alink> loonybot << bool()
20090412 12:38:42< loonybot> false
20090412 12:38:56< _ping_> Crab_: thinking btw i sended the replays
20090412 12:39:05< alink> ^always false for all compiler / system ?
20090412 12:39:25< Polarina> alink: Who knows?
20090412 12:39:45< Crab_> _ping_ ok. I'l see it in 10-12 hours (I'll have more time then)
20090412 12:39:49< _ping_> Crab_: yes like that
20090412 12:40:14< Crab_> _ping_: so, this is *additional* statistic, not a replacement for current expected dmg
20090412 12:40:28< CIA-30> silene * r34766 /trunk/ (configure.ac src/game.cpp src/language.cpp): 
20090412 12:40:28< CIA-30> Reverted "Removed conditional compilation for dummy_locales option" (revision 34764)
20090412 12:40:28< CIA-30> It doesn't detect disabled option and it causes problems on Windows.
20090412 12:40:30< alink> Polarina: someone I hope
20090412 12:40:41< _ping_> well the current way is also an result of a discussion
20090412 12:40:50< Polarina> loonybot << bool(false)
20090412 12:40:51< loonybot> false
20090412 12:40:54< alink> ^always 'false' for standard compiler / system ?
20090412 12:40:55< Polarina> alink: Use that.
20090412 12:41:12< Crab_> _ping_: and, only attacks actually made are to be counted. if the mage killed the target with his 2nd attack, then we'll have 2/1.4 , not 2/2.8
20090412 12:41:36< _ping_> Crab_: yes
20090412 12:41:57< Polarina> Happy Easter!
20090412 12:42:03< Crab_> _ping_: so, we'll not have this "problem" with 6-hp leftovers in this stat - since it'll count hits, not dmg
20090412 12:42:15< silene> Crab_: that won't work (we already had this discussion years ago), as it will show a bias toward succesful hits; indeed, if you hit and kill, you are decreasing the pool of actual hits/misses
20090412 12:42:44< Crab_> silene: can you illustrate with a small example ?
20090412 12:42:53< alink> Polarina: hum, it's a template that i can't change. It uses T() where T is = bool
20090412 12:43:04< Crab_> ah, seems I got it
20090412 12:45:23-!- Chusslove [n=caslav@brsg-d9bef4e8.pool.mediaWays.net] has joined #wesnoth-dev
20090412 12:45:23< _ping_> well as it is right now if a mage having 200 attacks attacks a 1 hp unit and misses all 200 times the stats will "show" that he was not that unlucky as he did 0/1 :)
20090412 12:45:56< Crab_> _ping_: good point
20090412 12:46:32< silene> Crab_: i have an example, do you still need it?
20090412 12:46:39< mordante> alink, afaik bool() should always return false
20090412 12:47:27< mordante> s/should/must/
20090412 12:47:27< silene> s/must/shall/ ;-)
20090412 12:47:27< mordante> thanks 
20090412 12:48:20< alink> yeah seems to make sense if int()=0  , but false and true had a less obvious default
20090412 12:48:26< Crab_> silene: I thing I've got it, but I may be wrong, so go on )
20090412 12:48:36< Crab_> s/thing/think
20090412 12:50:36< mordante> alink, zero initialization
20090412 12:52:34< alink> mordante: so that's not for my initial problem, but that means that in the initialization list of ctor, you could use some_bool() instead of some_bool(false). But it's obviously less clear
20090412 12:52:58< silene> Crab_: suppose you have 50% chance to hit and you can kill in one hit and you have two chances; the possible outcomes are 50% hit | 25% miss hit | 25% miss miss, that is 50% 1/0.5 | 25% 1/1 | 25% 0/1; so the total probability is suddenly 1.25 instead of the sane 1 (which you would obtain if the unit couldn't be killed in one hit)
20090412 12:54:07< mordante> alink, yes you can, I also prefer the some_bool(false) but both versions do the same
20090412 12:54:15< silene> when doing probabilities, the pool of events has to be fixed beforehand; if it changes in the middle of a run, you suddenly get insane results
20090412 12:54:47< Crab_> silene: yes, this method will not calculate *probability*.
20090412 12:54:48< alink> mordante: ok, thanks
20090412 12:54:52-!- stikonas [n=stikonas@wesnoth/translator/stikonas] has joined #wesnoth-dev
20090412 12:55:41< silene> Crab_: the question is: what does it compute? in the example, it would tell the user that he is +25% lucky every time, which doesn't make sense
20090412 12:56:28< silene> s/every time/on average/
20090412 12:56:56< CIA-30> loonycyborg * r34767 /trunk/SConstruct: Added dummy_locales option to scons.
20090412 12:57:40< silene> Crab_: the current engine already has this statistic, and it fixes the issue by not considering indidual hits, but complete sequence of hits
20090412 12:58:19< silene> and i shouldn't even say "fix", i should say "alleviate"
20090412 12:58:49< Crab_> silene: yes, but _ping_ is not happy with that ) so, I'm thinking 'are any additional ways to compute those stats available?"
20090412 12:59:34< silene> Crab_: perhaps; but my point was just that evaluating actual hits is biased toward successful hits
20090412 13:00:13< Crab_> silene: and what about calculating the actual %-chance to hit for each of possible chances-to-hit. for example, show the user that he, actually, hit 52%-of 50-percent-chance-to-hit attacks,  59% of 60-percent-chance-to-hit-attacks, 73% of 70-percent-chance-to-hit-attacks  ? 
20090412 13:00:19< silene> a way to fix would be to generate dummy attacks to fill the sequence, but then the user would complain that the values don't match their game
20090412 13:01:39< silene> Crab_: no, it still doesn't work; consider the example again: it would tell the user that he hits 75% of its 50%-cth attacks
20090412 13:02:17< silene> 62.5%, sorry
20090412 13:03:19< Crab_> yes, you're right
20090412 13:05:43< mordante> loonycyborg, any luck with testing on Windows or do you want to reboot later?
20090412 13:06:51< CIA-30> ivanovic * r34768 /branches/1.6/po/wesnoth-low/de.po: update/fix german translation of LoW
20090412 13:07:03< Polarina> Still not too late to submit new translations?
20090412 13:07:09< loonycyborg> mordante: I'll reboot later.
20090412 13:07:28< CIA-30> ivanovic * r34769 /trunk/po/wesnoth-low/de.po: updated German translation (ported fendrins update for LoW to trunk)
20090412 13:07:56< Polarina> Ivanovic: Is there still time to submit translations?
20090412 13:07:59< mordante> loonycyborg, ok, if it doesn't work I'd like to limit the number of characters before the 1.6.1 release
20090412 13:09:00< silene> Crab_: you know, it's equivalent to this classical exercise: there are three doors and a treasure behind one of them, you can only open one door so you choose one at random, say A; so the treasure has probability 1/3 to be behind A
20090412 13:09:53< Crab_> yes, i know that one...
20090412 13:09:57< silene> Crab_: now, before you open the door, someone comes and tells you that the treasure is not behind B; what is the probability of it being behind A?
20090412 13:10:04< Ivanovic> btw: a clear case of *WOW*: cycholka seems back, at least Mist is active again in the forums
20090412 13:10:17< _ping_> 1/3 and 2/3 behind c
20090412 13:10:32< loonycyborg> Ivanovic: I saw him on #wesnoth-umc-dev recently :P
20090412 13:12:27< silene> _ping_: precisely, and it shows how bad the probabilities get biased once you modify the event pool; users would kill us...
20090412 13:14:04< _ping_> silene: the reason for 2/3 is becuse someone transfered knowladge to the system .. you had 2/3 chance choosing the worng door and if in first place and be changing you will have 2/3 of chance of winning
20090412 13:14:54< silene> _ping_: yes, and when you kill a unit by a hit, you also transfer knowledge to the system; so you can't just count hits, this is not the correct metric
20090412 13:15:53< _ping_> if you modify the example choose door A and a random door of b and c opens, let say it was b and the tresure was not there then there would be 50% 50% and a and c
20090412 13:17:08-!- jmunro[a] is now known as jmunro
20090412 13:17:18< _ping_> well the preception of beng more lucky to hit a weak unit then a strong unit bugs me
20090412 13:18:24< _ping_> hitting an assian that haves 31 hp once and hitting once that have 61 hp once is the same (given the hitter does not have a chance to die)
20090412 13:18:59< Crab_> _ping_: if the target unit has any chance to die, then it is not the same
20090412 13:20:45< Crab_> _ping_: because if you hit 31-hp assassin once, you're avoided the result 'hit-31-hp-assassin twice and waste dmg'
20090412 13:21:45< Crab_> _ping_: so, the event *pool* is not the same in those cases.
20090412 13:21:47< _ping_> and that was bugs me, i would say i was actully a bit unlucky as i missed a ctk, also if the "real luck" should be calculated (which we cant) then a unit that had a chance to die but lived should not count full damage there after
20090412 13:22:46< Crab_> _ping_: current system doesn't take "kills" into account. it just takes dmg into account
20090412 13:22:53< _ping_> we have moved the calcualtion from pr attack to pr fight but it does not solve the pr "game" and also it introduced some (imo) bad side effects
20090412 13:24:05< CIA-30> ivanovic * r34770 /trunk/po/ (11 files in 6 dirs): updated Italian and Serbian translation
20090412 13:24:08< CIA-30> ivanovic * r34771 /branches/1.6/po/ (11 files in 6 dirs): updated Italian and Serbian translation
20090412 13:24:54< stikonas> mordante: I can send you a patch that reenables dummy locales in cmake. And should they be enabled by default?
20090412 13:25:01< Crab_> _ping_: also note that we can "evaluate" the value of the outcome differently, while keeping the current system of "expected value"
20090412 13:25:24< Crab_> _ping_: for example, "2 30-hp hits vs 36 hp assassin" is valued as 36. 
20090412 13:25:37< mordante> stikonas, I've a patch pending as well, but was waiting whether or not the code remains
20090412 13:25:52< Crab_> _ping_: we can attach some other value to this, to take into account the fact that the assassin was "killled"
20090412 13:25:55< mordante> and it's off by default, which used to be the default
20090412 13:26:53< Crab_> _ping_:  so,"as i missed a ctk"  can be taken into account in current system. we just need to evaluate all possible outcomes differently
20090412 13:26:59< loonycyborg> mordante: Conditional compilation will stay for now if it is what you mean.
20090412 13:27:07< Crab_> _ping_: for now it counts just damage.
20090412 13:27:07< Polarina> Ivanovic: I send you a update translation.
20090412 13:27:22< Reisiger> Crab_: If you got time, can you take a look at the unit::get_ability_attribute_bool ?  http://wesnoth.pastebin.com/d51086a58
20090412 13:27:50< silene> mordante: the code is already back there, and loonycyborg even added support for scons, so you can look at cmake whenever you want
20090412 13:28:40< mordante> ah ok, then I'll commit it after reviewing another patch
20090412 13:28:45< CIA-30> ivanovic * r34772 /trunk/po/ (wesnoth/is.po wesnoth-multiplayer/is.po): updated Icelandic translation
20090412 13:28:52< CIA-30> ivanovic * r34773 /branches/1.6/po/ (wesnoth/is.po wesnoth-multiplayer/is.po): updated Icelandic translation
20090412 13:30:09< _ping_> Crab_: yes and that is not a fair messure ) at least hits would be more fair as the above example the #hit would be 0.8 in both case (also still note that it only evaluate damage pr fight not damage pr game)
20090412 13:30:20< silene> Reisiger: what is it for? and note that for external abilities, you are actually testing the external unit, not the current unit
20090412 13:31:16< Reisiger> silene: It's ment to i.e. be used to test a unit if it can cure poison (heals poison=cured)
20090412 13:33:05< silene> Reisiger: then it probably doesn't do what you want; if a unit heals, but does not cure poison, but a unit next to it cures poison on adjacent units, you may wrongly believe that the unit also heals poison
20090412 13:33:38< Crab_> _ping_: it is a "weighting" measure, to attach more value to those hits who do more damage..
20090412 13:33:49< silene> (this part of the code has always been a bit convoluted)
20090412 13:34:16< Crab_> _ping_: of course, it fails to do its purpose when "no more damage is necessary, since we kill that target anyway"
20090412 13:34:31< _ping_> Crab_: is it how about posion? 
20090412 13:34:34< Reisiger> silene: I see... so I need to make different get_ability_bool and _abilities not checking surrounding units
20090412 13:34:44-!- edb [n=edb@101.117.88-79.rev.gaoland.net] has joined #wesnoth-dev
20090412 13:34:58< Crab_> _ping_: afaik, no
20090412 13:37:31< silene> Reisiger: not sure; if you could just live with get_ability_bool, it would probably the simplest (the AI is allowed to do some assumptions about the core units)
20090412 13:38:04< _ping_> Crab_: i agree on the no more damage we killed it, but by moving from pr attack to pr fight calcualtion did not solve the problem of calcualting game damage, also to me it introduced the "you are ecpected to do less damage even that you did not kill a unit but you had a chance to" problem
20090412 13:38:44< Reisiger> silene: Possible... I'll need to read which unit has the ability (and/or attribute/value) tho
20090412 13:40:05< _ping_> and btw the wigthing of attack by damage is not nessarry better then no wigthing as missing with posin, slow (attack with side effects) does not show
20090412 13:41:10< silene> _ping_: slow is taken into account
20090412 13:41:29< silene> as is berserk
20090412 13:41:31< _ping_> silene: how?
20090412 13:41:41< CIA-30> ivanovic * r34774 /trunk/po/ (wesnoth/de.po wesnoth-manpages/de.po): updated German translation
20090412 13:41:45< CIA-30> ivanovic * r34775 /branches/1.6/po/ (wesnoth/de.po wesnoth-manpages/de.po): updated German translation
20090412 13:41:54< _ping_> (about slow )
20090412 13:41:54< Crab_> silene: probably he means that the fact that slow *persists* after the combat, is not taken into account
20090412 13:42:05< _ping_> Crab_: yes
20090412 13:42:14< Ivanovic> silene: i saw you fixed some mistakes in the french translation
20090412 13:42:21< Ivanovic> have you already ported the files over to trunk, too?
20090412 13:42:28< Ivanovic> that is: ported and updated the po files
20090412 13:43:09< silene> Ivanovic: yes, i ported them (and updated the po files), and they were not mistakes... they were hundreds of fuzzy/untranslated sentences...
20090412 13:43:30< _ping_> Crab_: the fact that the argument as goes "you where more lucky not to hit the weak assassin twice" shows my point :)
20090412 13:43:49< Crab_> _ping_: no )
20090412 13:44:06< Crab_> if you hit the weak assasin twice, you'd have 36/20.  if you hit it once, you'd have 30/20
20090412 13:44:39< Crab_> you are more lucky to hit 36-hp assassin once that to hit 60-hp assassin once.
20090412 13:44:47< Crab_> because the *alternatives* are different
20090412 13:44:56< Crab_> from the dmg-viewpoint
20090412 13:45:12< _ping_> Crab_:  it was how you explained why the stats for hitting the weak and the strong onces (without killing it) are different
20090412 13:45:14< Ivanovic> silene: great
20090412 13:45:23< Crab_> yes
20090412 13:45:48< Crab_> so, dmg-viewpoint is limited, but it is objective
20090412 13:46:08< _ping_> it an okay explainating but is show my point of the limit in pr fight damage calcuation
20090412 13:46:23< Crab_> if we try to take into account kills, poison, experience gained, levels gainted, etc, it would be *more complete*, but subjective.
20090412 13:47:08< _ping_> in a way we removed some limitaion in pr attack damage calcuation but introduced some new in pr fight calcuation all becase we can not solve the pr game
20090412 13:47:45< Crab_> _ping_: you see, when evaluating, we use dmg because it is objective (there's only one way to count dmg)
20090412 13:48:32< _ping_> if used to messuare luck (and that what players do use them too) then it not fair as the chance to hit the two assasain in my examplees are the same
20090412 13:49:06< Crab_> _ping_: see: expected_dmg = sum_over_all_possible_outcomes ( value(outcome) * chance( outcome) )
20090412 13:49:14< Crab_> propose a value(outcome) function.
20090412 13:49:35< Crab_> now value(outcome) = damage_to_defender_in_outcome
20090412 13:49:53< Crab_> propose another value-function )
20090412 13:49:59< _ping_> Crab_: but it is calcuated pr unit fight that is if a unit had a chance to die but lived the next time and fight ouccur that is not taken into account
20090412 13:50:17< Crab_> yes. propose another value function that takes this into account.
20090412 13:50:29< Crab_> the formula will be still  expected_dmg = sum_over_all_possible_outcomes ( value(outcome) * chance( outcome) )
20090412 13:50:48< Crab_> but value(outcome) can be arbitrary.
20090412 13:51:36< _ping_> well the problem is that "sum_over_all_possible_outcomes" is darn big if view as a game :)
20090412 13:52:51< Crab_> "sum_over_all_possible_outcomes" is simply "law of total probability'
20090412 13:53:38< Crab_> the only thing that we can change (if we want this to be "expectation") is the *value* of each outcome from the outcome pool.
20090412 13:53:55< Crab_> (I've got to go, sorry)
20090412 13:56:18< CIA-30> loonycyborg * r34776 /branches/1.6/SConstruct: Added dummy_locales option to scons.
20090412 13:56:25< _ping_> Crab_:well in short what i'm saying is that going back to the old calcuation might be a better messure than the one we have now if used as a messure of luck, also i think that i have not made my self clear what i mean about "pr game" (sorry for my english) but if you see it as i do you know why i said that all possible outcomes are impossible to calcuated game wise
20090412 13:56:27< jmunro> hey guys, does anyone know what sort of time Sirp is likely to be online?
20090412 13:56:56-!- zitterbart [n=beard@dslb-092-072-047-067.pools.arcor-ip.net] has joined #wesnoth-dev
20090412 13:57:09< zitterbart> hi
20090412 13:57:30< Crab_> _ping_: why they are impossible to calculate ?
20090412 13:58:08< Crab_> in our example, we have "dead assassin, 16% chance", "6hp assassin, 48% chance", "36hp assassin, 36% chance"
20090412 13:59:00< _ping_> that is pr fight
20090412 13:59:08< zitterbart> i just read about the leadership-ai improvement that would be nice to have
20090412 13:59:12< Crab_> yes.
20090412 13:59:17< Crab_> and this we can calculate.
20090412 13:59:31< Crab_> and per-game outcome is subjective.
20090412 13:59:43< _ping_> pr game it would taken into account what happen before had the units a chance to die before etc...
20090412 13:59:48< Crab_> yes.
20090412 13:59:58< Crab_> we know the possible states of the game after combat, but we don't know how to value them.
20090412 14:00:03< Crab_> so we just take dmg.
20090412 14:00:06< Ivanovic> jmunro: he will for sure be around in 6h (when we do the *final* ranking of our SoC applicants)
20090412 14:00:23< Ivanovic> jmunro: in general i recommend just asking questions and hoping/waiting for an answer
20090412 14:00:46< jmunro> Ivanovic, thanks - I only received my comments yesterday as I have been on holiday, it looks like I'm now too late!
20090412 14:00:57< loonycyborg> Sirp doesn't appear to read backlogs.
20090412 14:01:01< Ivanovic> you are really late now, yes
20090412 14:01:04< zitterbart> if i would like to work on it, an i just look at the code and get started?
20090412 14:01:11< Ivanovic> loonycyborg: but others might be able to handle his questions
20090412 14:01:14< zitterbart> *can
20090412 14:01:43< _ping_> Crab_: but that why the "fairness" of the stats does not give a good messurement hence my mage missing 200 times on a 1 hp unit
20090412 14:02:18< jmunro> Ivanovic, thats tough luck, thanks for being honest
20090412 14:02:28< silene> zitterbart: sure, it's how all the developers, except for the first one, started; look at the code, write a patch, submit it
20090412 14:02:35< nagbot> Hi GSoC Applicants!
20090412 14:02:35< nagbot> SoC decision is near
20090412 14:02:35< nagbot> We are doing last checks before working on our final "ranking", which we will do *soon*
20090412 14:02:38< nagbot> So, if you got anything to add, post a comment in your proposal
20090412 14:02:41< Crab_> _ping_: that's fair. 1hp unit is not important :)))
20090412 14:02:42< jmunro> is there any point improving my proposal today?
20090412 14:03:07< _ping_> also in my example with the assasin my oppenent could claim after i hit the assassin once doing 30 hp : look how lucky you are you where only expected to do 19 damage
20090412 14:03:16< zitterbart> i thought there might be other people working on it, and it sure would be useful communicating with them, then :)
20090412 14:03:44< _ping_> well Crab_ we can take it another time you had to go :)
20090412 14:03:54< Crab_> _ping_: yes, D&D session )
20090412 14:04:07< silene> zitterbart: it doesn't ring a bell, but i can't be sure
20090412 14:04:36< Crab_> zitterbart: talk with boucman about that - he wants to introduce a new type of AI candidate move that would be more useful for leadership
20090412 14:04:38< _ping_> Crab_: good luck )
20090412 14:04:49-!- Crab_ [n=Crab_@wesnoth/developer/crab] has left #wesnoth-dev []
20090412 14:05:01< zitterbart> ah, very well
20090412 14:06:41< zitterbart> what are you guys working on at the moment?
20090412 14:07:46< silene> Ivanovic: in case you haven't noticed this patch: https://mail.gna.org/public/wesnoth-commits/2009-04/msg00523.html ; so it may be worth regenerating the pot files when you have time, so that translators don't work for nothing
20090412 14:08:10< Ivanovic> silene: since it is trunk only it can wait
20090412 14:08:16< zitterbart> and, how complicated is wesnoth's ai so far?
20090412 14:08:23< Ivanovic> and the other stuff they work on is in anyway
20090412 14:08:34< CIA-30> alink * r34777 /trunk/src/ (image.cpp image.hpp unit_frame.cpp unit_frame.hpp): 
20090412 14:08:34< CIA-30> Simplify and optimize the in_hex caching by remove it from amimation code
20090412 14:08:34< CIA-30> and instead implement it like other image caches (but just storing bool)
20090412 14:08:34< CIA-30> This allow to load and scan images only when needed.
20090412 14:08:34< CIA-30> Also fix a tiny bug when using :refresh after modifiying sprite's size.
20090412 14:08:42< Ivanovic> yes, translators are *not* encouraged to work on trunk atm
20090412 14:08:56< silene> Ivanovic: that's fine then
20090412 14:10:01< zitterbart> i'm off for lunch, but i'll get back later
20090412 14:10:08< zitterbart> see you
20090412 14:10:32< alink> boucman: ^ I think is a lot simpler like that . And also much more efficient (for memory, disk and cpu)
20090412 14:11:52< alink> boucman: also the fix of :refresh make me think that it's possibly more robust to WML actions
20090412 14:12:09-!- zitterbart [n=beard@dslb-092-072-047-067.pools.arcor-ip.net] has quit ["Verlassend"]
20090412 14:14:12< alink> mmh very short for porting it in 1.6, but I am really not fan of all these loading recently added there.
20090412 14:16:27< benetnash> boucman: do you copy?
20090412 14:17:00-!- stikonas [n=stikonas@wesnoth/translator/stikonas] has quit [Read error: 110 (Connection timed out)]
20090412 14:17:25< alink> benetnash: no he's away, I just report and comment my work because I think it might interest him
20090412 14:18:12< benetnash> alink: ok, thx
20090412 14:21:23< alink> btw, I think i will continue to hide the new faction_name WML key in 1.6.1. Even if compatible and possibly usefull, it's probably better to keep WML data identical regardless of the host's version
20090412 14:22:59< silene> alink: yes; as a matter of fact, i don't even understand why you added it to 1.6
20090412 14:23:47< alink> because the key was missing/broken in the info sent by 1.6.0, preventing me to fix the UI bug
20090412 14:24:28-!- Reisiger [n=Reisiger@adsl-84-226-76-63.adslplus.ch] has quit ["Verlassend"]
20090412 14:24:44< Polarina> Trunk is evil, unless I'm told otherwise.
20090412 14:27:16-!- molgrum [n=molgrum@c83-249-53-163.bredband.comhem.se] has joined #wesnoth-dev
20090412 14:28:11< alink> and as someone said, the bug was 'unfair' (because host was the only one who could see faction of other players) so it was not just a 'cosmetic' bug
20090412 14:34:29-!- Crab_ [i=crab@wesnoth/developer/crab] has joined #wesnoth-dev
20090412 14:35:59-!- Turuk [n=Turuk@wesnoth/forumsith/turuk] has joined #wesnoth-dev
20090412 14:36:50< Crab_> silene: I was thinking a bit.. isn't the current statistics system also biased towards successful hits ? for example, if we take enemy peasant and two dragonguards.each with 50% change to "hit and kill" that peasant, and we will attack with the second dragonguard only if the first misses, then our avg. damage will be more than expected damage.
20090412 14:38:34< CIA-30> alink * r34778 /trunk/src/unit_frame.cpp: 
20090412 14:38:34< CIA-30> Optimization of the in_hex scanning.
20090412 14:38:34< CIA-30> Skip it when not aligned to the hex grid (because we will not use it anyways).
20090412 14:38:34< CIA-30> This mainly avoid to scan most of the frames of melee animations.
20090412 14:38:45-!- zitterbart [n=beard@dslb-092-072-047-067.pools.arcor-ip.net] has joined #wesnoth-dev
20090412 14:38:59< zitterbart> hi
20090412 14:39:10< Polarina> alink: Please do not use slangs in SVN commits.
20090412 14:39:29< silene> Crab_: sure; any system is actually biased since you can't kill a unit with a miss; but at least the current is unbiased on the length of a complete fight; it only gets biased when you consider several fights in a row
20090412 14:39:53< alink> Polarina: slangs ? where ? and why can't use it there ? ;-p
20090412 14:40:16< Polarina> alink: "anyways" is an American slang.
20090412 14:40:22< zitterbart> why does it get biased?
20090412 14:40:57< silene> zitterbart: because the ultimate swing is necessarily a hit :-)
20090412 14:41:17< zitterbart> what is that? an attack?
20090412 14:41:19< Crab_> yes. I agree. but this explains the fact that in x/y  , x tends to "float" to be >y. and that's the reason _ping_ was wondering...
20090412 14:42:20< alink> Polarina: mmmh ok, sorry, I was not aware of that. I live in belgium. But I suppose that it's not a very 'rude' word
20090412 14:42:23< zitterbart> sorry, i'm playing german version of wesnoth
20090412 14:42:34< alink> just some idiom not very nice
20090412 14:42:50< Polarina> alink: No, but the correct word is "anyway".
20090412 14:44:17< alink> Polarina: yes thanks for the info (even if I probably knew that)
20090412 14:44:32< Polarina> alink: No problem.
20090412 14:45:02< silene> Crab_: i'm not sure it tends that much; at least i can't think of an example where it would
20090412 14:45:30< Crab_> silene: he's got some replays..
20090412 14:45:38< Crab_> 700/450 or something like this..
20090412 14:45:58< _ping_> Crab_: i just got the perfect example
20090412 14:46:19< silene> Crab_: dfranke said that the stored value was not correct; that if you replayed the game, the values were matching
20090412 14:46:32< _ping_> of what i mean in or game as pr fight lets say 2 horsemen attack the same assaigan having 60 hp
20090412 14:47:08< _ping_> if first misses both and second hit once the stats will read 30 of 48
20090412 14:47:53< _ping_> if first hits once and second misses both the stat read 30 of 43
20090412 14:48:01 * AnMaster wonders why the last scenario of Northen rebirth is so slow... Even moving mouse in it. I guess all those units cause it (around 100-150 units on the map I guess, total from all sides)
20090412 14:48:09< AnMaster> is this a known issue?
20090412 14:48:28< AnMaster> or should I go mess with oprofile or something
20090412 14:48:37< benetnash> Crab_: how can I include file with formula ai defs in other script?
20090412 14:48:43< Crab_> yes
20090412 14:49:19< Crab_> 2 ways. 'run_file', and  include (ask Dragonking for the syntax)
20090412 14:49:42< CIA-30> alink * r34779 /trunk/src/unit_frame.cpp: Tiny optimization of frequently changing/moving big unit (like shadow)
20090412 14:49:44< AnMaster> because this is unplayable on my computer (Sempron 3300+ (2 GHz) x86_64, 1.5 GB RAM, Linux 2.6.27)
20090412 14:49:58< Dragonking> benetnash: {path_to_other_file}     same like you include files in scenario cfgs
20090412 14:50:09< AnMaster> seeing all those optimisations alink is working atm, are they related to this in any way?
20090412 14:50:22< AnMaster> working on atm*
20090412 14:50:43< silene> _ping_: what do you suggest? that the game should ask the user to make all his decisions beforehand? you can't really compare these two cases because we allow the user to change his mind between the two fights
20090412 14:51:02< alink> AnMaster: it will indeed help the framerate and cpu when there is a lot of units on screen
20090412 14:51:27< AnMaster> alink, Well I think wesnoth was causing swap trashing too
20090412 14:51:53< AnMaster> alink, at least once I exited wesnoth it went from 0 MB free to 782 MB free RAM.
20090412 14:51:59< alink> AnMaster: ah sorry no memory usage change (at least by me)
20090412 14:52:55< AnMaster> alink, any issues with memory leaks? Maybe I should try using valgrind on wesnoth...
20090412 14:52:56< alink> AnMaster: unless SDL was causing it, because by drawing stuff less often, we use less temporary SDL surface
20090412 14:53:29< AnMaster> almost 800 MB sounds insane to me...
20090412 14:53:44< Ivanovic> AnMaster: the problem is northern rebirth
20090412 14:53:55< Ivanovic> it is just too huge...
20090412 14:54:03< AnMaster> Ivanovic, oh? Yeah I never had so many units in any other campaign
20090412 14:54:08< Ivanovic> (that is: too big maps with too many units and all this stuff)
20090412 14:54:14< AnMaster> but I like it. Fun, epic.
20090412 14:54:47< alink> AnMaster: indeed. It would be nice to find if it comes from a memory's leaks amplified by NR, or just comes from the NR crazyness 
20090412 14:54:53< AnMaster> still, when moving a drake takes like 5 seconds per hex it moves over something is seriously wrong.
20090412 14:55:20< _ping_> silene: i suggest you aknowledge that the system is not perfect right now and that there could be room for improvement :) Well i think going back to old stats with the modification that once can not i once attack do more damage than unit have would be better
20090412 14:55:28< AnMaster> alink, since valgrind more than doubles the needed RAM I don't think my system could handle NR when running wesnoth under valgrind
20090412 14:55:34< AnMaster> since I "only" have 1.5 GB RAM
20090412 14:56:11-!- stikonas [n=stikonas@ctv-79-132-187-171.vinita.lt] has joined #wesnoth-dev
20090412 14:56:37< alink> well 1GB here 
20090412 14:57:10-!- molgrum [n=molgrum@c83-249-53-163.bredband.comhem.se] has quit [Read error: 104 (Connection reset by peer)]
20090412 14:57:18< AnMaster> I guess I could try rebuilding it as 32-bit (first I need to build 32-bit SDL and so on though), don't know how much it would help for wesnoth, is it pointer heavy?
20090412 14:57:25< alink> btw what can cause the lost of half a DIMM (256MB of one of my two 500MB) ?
20090412 14:57:38-!- molgrum [n=molgrum@c83-249-53-163.bredband.comhem.se] has joined #wesnoth-dev
20090412 14:57:43< AnMaster> alink, hm... mislabeled?
20090412 14:57:48< silene> _ping_: i acknowledge that the system is not perfect (and if you look at the logs, you will notice that this is not the first time i acknowledge this); but what i don't acknowledge at all is that the old system was better
20090412 14:58:02< Polarina> Any schedule on when 1.6.1 will be released (source code)?
20090412 14:58:25< alink> AnMaster: no worked fine during months and one day only 0.75 GB detected
20090412 14:58:37< Ivanovic> Polarina: *soon*
20090412 14:58:40< alink> changing that DIMM fixed it
20090412 14:58:42< Polarina> Ivanovic: I see.
20090412 14:58:47< Polarina> Ivanovic: Today?
20090412 14:59:07< Ivanovic> regarding the topic of this chan: yes
20090412 14:59:08< Ivanovic> ^^
20090412 14:59:09< AnMaster> alink, that sounds _very_ odd. No idea really. If it broke I would have expected it to not work at all instead...
20090412 14:59:22< Polarina> Ivanovic: Awesome!
20090412 14:59:23 * alink hopes that boucman will return before the tagging
20090412 14:59:26< AnMaster> alink, ninja in at night replacing the ram stick?</silly>
20090412 14:59:38< alink> AnMaster: lol
20090412 14:59:49 * Polarina hopes no one forgets http://www.wesnoth.org/wiki/Credits
20090412 15:00:36< loonycyborg> alink: There are limitation on maximum size of each module and on total memory size depending on MB.
20090412 15:00:58< loonycyborg> Perhaps you've hit one of them :P
20090412 15:01:12< AnMaster> alink, I think I had about 90-120 units or so in the recall list in the last scenario in NR. When I pressed Alt-R  it took several seconds before the recall list showed up.
20090412 15:01:24< silene> AnMaster: it isn't that odd; memory builders don't discard chips that are not perfect, they just sell them as smaller ones; so chips have systems embedded in them to reduce their size; usually it is static, but it can also be dynamic
20090412 15:01:49< AnMaster> silene, aha, that would explain it.
20090412 15:01:59< AnMaster> alink, better run memtest then
20090412 15:02:18< alink> loonycyborg: yeah but i was using correct size etc... And this worked fine during months. I am mainly surprised that only half of it stopped working
20090412 15:02:40< silene> never run memtest, unless you want to be sure to your memory stick is definitely fried afterwards :-)
20090412 15:02:48< alink> mmh I didn't tried to replug it ...
20090412 15:02:50-!- gnutoo_ is now known as GNUtoo
20090412 15:03:07< alink> (mainly because i had another 500MB dimm and only 2 slots)
20090412 15:03:46< loonycyborg> alink: Those modules often contain multiple chips. Perhaps they allow some of them fail.
20090412 15:04:09< alink> loonycyborg: which is good I guess 
20090412 15:04:26< loonycyborg> Or the either MB or module just conform to standards poorly.
20090412 15:04:51< AnMaster> <silene> never run memtest, unless you want to be sure to your memory stick is definitely fried afterwards :-) <-- really?
20090412 15:05:11< AnMaster> sounds like low quality ram modules to me then.
20090412 15:05:18< alink> but maybe add a warning somewhere, no idea since how many days i had less memory (but not long I think)
20090412 15:05:57< _ping_> silene: okay i'm gald we at least agree that the system / model is is in this case here a choosing and not a exact math given.  well i think my suggestion with the combo of old model and current would be better but i will give it some more thoughts
20090412 15:06:05< silene> AnMaster: i'm exagerating; but memtest tends to stress memory sticks a lot more than what you would do with normal usage (disabled caches and so on), so the sticks will heat a lot more
20090412 15:07:04< silene> _ping_: no, it is an exact model for a single fight; where it becomes incorrect is when you consider several fights in a row; the old model wasn't even exact for a single fight
20090412 15:07:30< AnMaster> silene, hm, my system doesn't get very hot anyway, after compiling wesnoth CPU was 31 C for example.
20090412 15:07:46< AnMaster> there are advantages to using low end systems :)
20090412 15:08:03< AnMaster> (low end these days at least)
20090412 15:08:48< _ping_> silene: but choosing to use the model have some drawbacks
20090412 15:09:20< AnMaster> as for ram temp, I don't know, but iirc it doesn't get very hot either. Plus they are quite close to one of the fans.
20090412 15:09:21< _ping_> silene: yes for one fight it is good but the stats shows for a game so that will lead to wrongness
20090412 15:10:12-!- jmunro [i=james@unaffiliated/jmunro] has quit [Read error: 60 (Operation timed out)]
20090412 15:10:35< _ping_> and once again i dont dispute the model i dispute the use of the model
20090412 15:10:59-!- jmunro [i=james@217.112.83.12] has joined #wesnoth-dev
20090412 15:11:39-!- jmunro is now known as jmunro[a]
20090412 15:12:01< _ping_> it like saying 2*2 = 4 but if you want to apromisted PI it might not be the best way to do it, still 2*2 is without doubt 4 (10 based system)
20090412 15:13:05< Polarina> 2 * 2 is 11 (base 3)
20090412 15:15:12< benetnash> Crab_: it seems, that run_file() doesn't add defs in file to global defs table so you cannot include collection of defs usnig run_file. Also run_file doesn't support includes with {}. Will it be fixed in near future?
20090412 15:16:58< Crab_> yes
20090412 15:26:17-!- loonycyborg [n=sergey@wesnoth/developer/loonycyborg] has quit ["KVIrc 3.4.2 Shiny http://www.kvirc.net/"]
20090412 15:26:53-!- loonybot [n=loonybot@wesnoth/bot/loonybot] has quit [Remote closed the connection]
20090412 15:28:07< alink> hehe I almost hope to find a problem with r34777, so I would stop being tempted to port it to 1.6 ;-)
20090412 15:28:23< Ivanovic> tststs
20090412 15:30:10< alink> (note that it will reduce the change between 1.6.0 and 1.6.1, but it has almost no testing compared to the 2 weeks in trunk for the other slower way)
20090412 15:30:58< CIA-30> ivanovic * r34781 /branches/1.6/po/ (5 files in 5 dirs): updated Turkish translation
20090412 15:30:59< Ivanovic> short guess: the stuff in trunk had almost no testing either
20090412 15:31:00< Ivanovic> ;)
20090412 15:31:08< CIA-30> ivanovic * r34780 /trunk/po/ (5 files in 5 dirs): updated Turkish translation
20090412 15:31:26< alink> probably true :)
20090412 15:31:46-!- Turuk_ [n=Turuk@wesnoth/forumsith/turuk] has joined #wesnoth-dev
20090412 15:36:06< alink> especially because the main risk of glitch is some missing ear of a big sprite when moving mouse around it. Easily missed.
20090412 15:37:05< Polarina> So many update translations...
20090412 15:37:43< alink> I fear more a coding error causing a bad memory action on another system
20090412 15:38:37< alink> (even if I am confident, which means nothing for this kind of things ;-p )
20090412 15:39:11-!- Turuk [n=Turuk@wesnoth/forumsith/turuk] has quit [Connection timed out]
20090412 15:41:06-!- YogiHH [n=chatzill@c214211.adsl.hansenet.de] has joined #wesnoth-dev
20090412 15:41:33< YogiHH> hello everyone
20090412 15:41:41< alink> hi YogiHH
20090412 15:42:35< Ivanovic> hi YogiHH
20090412 15:43:03< zitterbart> hi yogihh
20090412 15:43:48< Dragonking> hi YogiHH 
20090412 15:44:28< alink> here is the image loaded and scanned by current 1.6  http://wesnoth.pastebin.com/d4fa455b when starting test-scenario
20090412 15:44:31 * zitterbart strikes his beard like Delfador
20090412 15:44:42< alink> and after r34777 on trunk : http://wesnoth.pastebin.com/m19ff15fd
20090412 15:45:47< Polarina> alink: And the point is?
20090412 15:46:58< alink> Polarina: the point is that really didn't like how current 1.6 do a lot of new useless work and I am annoyed to have noticed it so late
20090412 15:47:11< Polarina> alink: Port it to 1.6
20090412 15:48:26< zitterbart> why is there ai.cpp and ai2.cpp?
20090412 15:48:29< alink> Polarina: yeah i want to, but that a bit unwise, so i am hoping some second opinion of someone knowing this code (mainly boucman)
20090412 15:48:42< zitterbart> *hpp
20090412 15:48:55< Polarina> alink: I can help you test. What do you want me to do?
20090412 15:49:16< alink> zitterbart: never noticed ai2.hpp before :)
20090412 15:49:28< alink> Polarina: can you compile trunk ?
20090412 15:49:38< Polarina> alink: Yes.
20090412 15:49:39< zitterbart> yeah, it's quite short
20090412 15:49:59< alink> Polarina: so you know the :sunset command ?
20090412 15:50:05< alink> s/so/do
20090412 15:50:06< Polarina> alink: No...
20090412 15:50:17< silene> zitterbart: it was just in case someone would one day write an ai; it never happened
20090412 15:50:26< Polarina> alink: I don't even know how to use these commands.
20090412 15:50:38< silene> (in fact, it happened, but not there)
20090412 15:50:48< alink> Polarina: type ':' in debug mode (for all command)
20090412 15:50:51< Polarina> Trunk compiled.
20090412 15:50:57< Polarina> alink: How do I turn on debug mode?
20090412 15:51:01< alink> Polarina: (in game)
20090412 15:51:02< zitterbart> no i'm confused ... i think there already is an ai?
20090412 15:51:13< alink> Polarina: same thing ':' and type 'debug'
20090412 15:51:18< Polarina> alink: Ok.
20090412 15:51:20-!- cyber_11 [n=muratcan@78.165.99.242] has quit []
20090412 15:51:26< alink> Polarina: there is also a 'help' command
20090412 15:51:30< silene> zitterbart: there are several ais
20090412 15:51:45< Polarina> alink: Ok. Tell me what to do.
20090412 15:52:09< alink> Polarina: use the ':sunset' command
20090412 15:52:23< alink> the screen will become darker
20090412 15:52:53< alink> only refreshing what is updated, like when you move your mouse
20090412 15:53:00< Polarina> alink: Nothing happens.
20090412 15:53:30< alink> Polarina: did you saw the command line where you type your command ?
20090412 15:53:34< Polarina> alink: No.
20090412 15:53:49< zitterbart> silene: are they all used in the game? i noticed the first function i looked at had the comment of being an simple ai.
20090412 15:54:17< alink> Polarina: launch a game, press ':' and it must show up, unless you changed the hotkey
20090412 15:54:29< Polarina> alink: I haven't changed any hotkeys.
20090412 15:54:35< Polarina> alink: : does nothing.
20090412 15:55:01< alink> Well it's the "Enter user command" hotkey
20090412 15:55:14< zitterbart> or is there one for dwarves, one for elves, etc?
20090412 15:55:26< Polarina> alink: It's set to :
20090412 15:55:50< alink> Polarina: the line show up in the bottom of the screen with a yellow "command:"
20090412 15:55:53< silene> zitterbart: no, only the default ai is used; if it had not bitrotten, the dfool ai could be used to; and in the future, a full formulaai may be used
20090412 15:55:58< Polarina> alink: No such thing.
20090412 15:56:35< Polarina> alink: All I see on the bottom is the envelope and a label written on "Útgáfa 1.7.0-svn (34781)
20090412 15:56:36< alink> Polarina: strange, this is really usefull stuff btw, you can skip level, cheat xp hp etc...
20090412 15:57:00< alink> Polarina: oh ok sorry , I said start *a* game
20090412 15:57:01< zitterbart> ah, i see
20090412 15:57:05< Polarina> alink: Oh...
20090412 15:57:09< alink> not just *the* game
20090412 15:57:56< Polarina> alink: Unknown command :sunset
20090412 15:58:09< alink> Polarina: you need debug mode
20090412 15:58:14< alink> use :debug
20090412 15:58:24< Polarina> alink: Unknown command (:debug)
20090412 15:58:40< alink> first ':' then 'debug'
20090412 15:58:53< silene> Polarina: it's an option at configure time; if you haven't enabled it, you will have to recompile
20090412 15:58:54< Polarina> alink: "Aflúsunarham virkt!"
20090412 15:58:58< zitterbart> formulaai would be a formula to give values to position, like in chess engines? or something else?
20090412 15:59:22< Polarina> alink: sunset does nothing.
20090412 15:59:25< silene> zitterbart: it's more than that, but that's the idea
20090412 15:59:44< alink> silene: no I think game debug mode works for everybody, except in network MP
20090412 15:59:47< zitterbart> sounds cool
20090412 16:00:06< Polarina> alink: That mode is buggy as hell.
20090412 16:00:19< alink> Polarina: screen is darkening ...
20090412 16:00:30< Polarina> alink: Yes, why?
20090412 16:00:45< alink> Polarina: it's a debug tool, you can see what the engine redraw
20090412 16:00:56< alink> we want to minimize that
20090412 16:01:07< Polarina> alink: I see.
20090412 16:01:23< alink> in 1.6.0 any move near a pack of units triggers a refresh of all these units  
20090412 16:01:24< Polarina> alink: I see it drawing stuff not being animated.
20090412 16:01:28< zitterbart> unlike chess, wesnoth also depends on chance. seems not so easy to put that in.
20090412 16:01:31< Polarina> alink: For example, castle tiles.
20090412 16:02:11< alink> Polarina: yes we need to refresh adjacent hexes of changing things
20090412 16:02:20< Polarina> Ok.
20090412 16:02:26< Polarina> alink: What do you want me to do now?
20090412 16:02:48< alink> because SDL is rectangle based and wesnoth use hexagon so all hex 'overlap'
20090412 16:02:54< silene> alink: does it? at a time, the DEBUG macro had to be defined at compile time; but perhaps it was changed
20090412 16:03:25< alink> Polarina: this optimization detect which unit fit in a hex
20090412 16:03:26< Polarina> alink: Ok, and what do you want me to test?
20090412 16:03:55< alink> create a small dwarf and move you mouse on it, it must not refresh adjacent hexes
20090412 16:04:40< Polarina> alink: What do you mean?
20090412 16:04:45< zitterbart> a problem might be that all campaigns have to be overworked after an new ai is used
20090412 16:04:46< alink> then create a necromancer, because of his stick, it doesn't fin in hex and thus refresh 4 adjacent hex 
20090412 16:05:12< alink> Polarina: do you know how to create unit in debug mode ? (use context menu)
20090412 16:05:17< Polarina> alink: No...
20090412 16:05:33-!- fendrin [n=fabi@wesnoth/developer/fendrin] has quit [Remote closed the connection]
20090412 16:05:53< alink> Polarina: try context menu+ create unit, it's nice
20090412 16:06:14< Polarina> alink: Huh?
20090412 16:06:27< alink> right+click on tile
20090412 16:06:38< alink> * a tile/hex
20090412 16:07:00< Polarina> alink: When I hover over it, it updates tons of other tiles.
20090412 16:07:27< alink> yeah each tile you pass by
20090412 16:07:34-!- zitterbart [n=beard@dslb-092-072-047-067.pools.arcor-ip.net] has quit ["Verlassend"]
20090412 16:07:40< Polarina> alink: I haven't moved a single unit.
20090412 16:07:49< alink> + keep shift unit out of hexes, stay away of keep
20090412 16:07:54< alink> Polarina: scrolling ?
20090412 16:07:59< Polarina> alink: Tons, I mean, several tiles around it.
20090412 16:08:02< Polarina> alink: No, not scrolling.
20090412 16:08:11< Polarina> alink: Why stay away from the keep?
20090412 16:08:17< alink> Polarina: around units ?
20090412 16:08:30< Polarina> alink: Yes, that shouldn't make a difference.
20090412 16:09:01< alink> Polarina: because the optimization is ineffective there, units are shifted up and so don't fit in their hex
20090412 16:09:19< Polarina> I see.
20090412 16:09:47< Polarina> This mode is boring.
20090412 16:10:07< alink> lol, that's a debug tool, not a game ;-p
20090412 16:11:15< alink> a simpler test is to verify that the stick of the necromancer is not cutted when moving the mouse around him, like it does in 1.4
20090412 16:11:53< alink> or that units don't lose their head when on a keep
20090412 16:13:08-!- loonycyborgw32 [i=kvirc@79.139.246.210] has joined #wesnoth-dev
20090412 16:13:38< alink> well at least it didn't crashed :)
20090412 16:13:41< CIA-30> silene * r34782 /trunk/src/ (unit_animation.cpp unit_types.cpp): Removed unused iterators.
20090412 16:13:50< CIA-30> silene * r34783 /trunk/src/ (config.hpp unit_types.cpp unit_types.hpp): Removed config::const_child_iterator default constructor and unit_type_data::unit_type_map_wrapper copy constructor.
20090412 16:13:54< CIA-30> silene * r34784 /trunk/src/config.hpp: Removed config::child_iterator default constructor.
20090412 16:13:59< CIA-30> silene * r34785 /trunk/src/ (config.hpp variable.cpp variable.hpp): Removed config::all_children_iterator and vconfig::all_children_iterator default constructors.
20090412 16:14:01< CIA-30> silene * r34786 /trunk/src/config.hpp: Removed config::const_attribute_iterator default constructor. Defaulted its copy constructor.
20090412 16:14:11 * Polarina tests if trunk still compiles.
20090412 16:14:48< alink> use :create Necromancer  to test ^
20090412 16:15:10< Polarina> But why no debug mode in multiplayer?
20090412 16:15:11< alink> btw it's not nice to crash after a typo there
20090412 16:15:34< Polarina> It crashes on typoes?
20090412 16:15:36< alink> Polarina: why not allowing to cheat and create level 5 unit for free ?
20090412 16:15:36< Polarina> That's a bug.
20090412 16:15:53< Polarina> alink: But, technically, it would be possible (through the current protocol)?
20090412 16:15:57< alink> :create necrotypo crashes
20090412 16:16:06< Polarina> alink: Why not fix that?
20090412 16:16:30< alink> Polarina: I think if you modify code and use debug there, it will OOS 
20090412 16:16:37< Polarina> OOS?
20090412 16:16:59< silene> out of synchronization; and yes it does
20090412 16:16:59< alink> Out of Sync
20090412 16:17:06< Polarina> o.O
20090412 16:17:32< Polarina> Trunk compiles. :D
20090412 16:19:06< YogiHH> silene, you got a few minutes for me?
20090412 16:20:14< silene> YogiHH: yes
20090412 16:21:01< YogiHH> silene: Can you explain to me in simple words what a preproc_define is and what its main task is supposed to be?
20090412 16:21:48< silene> not sure, it didn't exist at the time i wrote the preprocessor, it was only added after i left; as far as i can tell, it's used to store campaign-wide macros
20090412 16:21:54< loonycyborgw32> mordante: Your patch doesn't fix that error on native windows either.
20090412 16:22:07< esr> YogiHH: have you looked into bug #13330 (saved moveto event does not fire) at all?
20090412 16:22:32< silene> since campaigns are preprocessed again for each new scenarios, as far as i know
20090412 16:24:08< YogiHH> silene: i see
20090412 16:24:08< alink> IIRC it's also used to define the difficulty define and campaign-name define
20090412 16:24:39< silene> alink: that's what i call campaign-wide macros :-)
20090412 16:24:46< alink> I mention it because you works on savegame, so there is maybe specific things about these two
20090412 16:25:16< silene> YogiHH: why are you interested in them?
20090412 16:25:17< alink> silene: yes i know, it's just because it not really a macro and is a bit 'special ' part in code
20090412 16:26:21< alink> for example these defines are encoded in the SHA hash key of the cache files
20090412 16:26:31< YogiHH> silene: they are making my life harder to make up a nice design for loading games. I am not sure, but from what i figured they shouldn't be known to savegame.cpp.
20090412 16:27:08< alink> well, as I just mentioned it's related to cache loading
20090412 16:27:32< alink> or savegame.CPP not just savegame, sorry
20090412 16:27:42< alink> s/or/oh
20090412 16:28:32< YogiHH> esr: no, i just got back from a two day off-time. Haven't looked into this one yet
20090412 16:29:24< silene> YogiHH: it depends; i don't what the goal of savegame.cpp, but if it has to restore the context of the savegame (e.g. unit types), then it has to know about the campaign-wide defines
20090412 16:29:48< esr> YogiHH: It's screwing up DNM pretty badly - I'vee seen at least two symptoms and there are probably others.
20090412 16:29:58< esr> s/DNM/DM/
20090412 16:31:47< YogiHH> silene: the main goal of savegame.cpp is managing savegames as a whole, but not too much dealing with the contents.
20090412 16:32:25< YogiHH> esr: i see, i will look into it as soon as preparations for GSoC are done
20090412 16:32:48< esr> YogiHH: Thanks.
20090412 16:33:33-!- crimson_penguin [n=ben@64.201.60.216] has joined #wesnoth-dev
20090412 16:34:12-!- loonycyborgw32 [i=kvirc@79.139.246.210] has quit ["KVIrc 3.4.0 Virgo http://www.kvirc.net/"]
20090412 16:34:59-!- edb [n=edb@101.117.88-79.rev.gaoland.net] has quit [Remote closed the connection]
20090412 16:35:27-!- Turuk_ [n=Turuk@wesnoth/forumsith/turuk] has quit [Read error: 60 (Operation timed out)]
20090412 16:36:13-!- Turuk [n=Turuk@wesnoth/forumsith/turuk] has joined #wesnoth-dev
20090412 16:39:13-!- Sirp [n=me@wesnoth/developer/dave] has joined #wesnoth-dev
20090412 16:40:01-!- zookeeper [n=l@wesnoth/developer/zookeeper] has joined #wesnoth-dev
20090412 16:41:10< silene> YogiHH: i'm not sure i follow; how are these defines making it harder?
20090412 16:41:52< YogiHH> silene: in game.cpp::load_game, they are mixed up with other code i am trying to move to savegame.cpp
20090412 16:42:27< alink> grmbmrml found a flaw in this hex's invalidation code (in 1.6.0, 1.6.0+svn and trunk) but currently only reproduced it with special hand-made sprites, and slightly harder to do with 1.6.0 (because it invalidate so much things)
20090412 16:44:59< Polarina> Ivanovic: What are the procedures for a translation maintainer to add a new translator?
20090412 16:45:35< silene> YogiHH: you won't avoid mixing them; for instance, in order to load units from a savegame, you need their unit types, which are present in campaigns, eras, whatever; so you have to load them too
20090412 16:46:10< silene> savegames are not self-contained (fortunately, otherwise they would be ridiculously huge)
20090412 16:46:18< Polarina> :(
20090412 16:49:06< YogiHH> silene: i know that. It's not that i want to change functionality at this stage. I am just restructuring code and try to figure out what belongs to where. And i am not sure if i should marry the cache processing code with the savegame managing code. So far, i feel it is not so good an idea.
20090412 16:51:10-!- loonycyborgw32 [i=kvirc@79.139.246.210] has joined #wesnoth-dev
20090412 16:58:06< Polarina> When is 1.6.1 comming? Today?
20090412 16:58:19< stikonas> Polarina: yes, see topic
20090412 16:58:57< Polarina> stikonas: I know, but will it be released, not only packaged, today?
20090412 16:59:51< zookeeper> since when has there been a [stone] action tag?
20090412 16:59:56< silene> YogiHH: why does it seem like a bad idea? it seems to me that getting rid of the game_controller monolith and having a separate file/class/... for loading context is a good idea; or is there another issue somewhere else?
20090412 16:59:57< stikonas> Polarina: it will be annouunced a couple of days later.
20090412 17:00:14< Polarina> Oh...
20090412 17:00:18< stikonas> s/annouunced/announced/
20090412 17:02:24< YogiHH> silene: to be honest, it's mainly lack of knowledge at the moment. It seems to me that caching covers a bit more than just savegame issues. That's why i wanted to know more about this preproc defines.
20090412 17:03:31< silene> YogiHH: yes, i'm not talking about putting context-loading into savegame.cpp, but i wouldn't mind having it off game.cpp
20090412 17:03:58< YogiHH> silene: yes, makes sense to me, too
20090412 17:04:01< crimson_penguin> Polarina: you can play it today though (assuming I get the dmg up today)
20090412 17:04:24< crimson_penguin> hopefully I have time before the party at 4
20090412 17:04:28< crimson_penguin> Ivanovic: ETA? :)
20090412 17:05:30< Polarina> crimson_penguin: :)
20090412 17:13:17< CIA-30> esr * r34787 /trunk/data/campaigns/Delfadors_Memoirs/ (TODO maps/manor.map maps/parthyn_ford.map maps/prince.map): DM map improvements along lines suggested by ShadowMaster.
20090412 17:14:02< zookeeper> esr, since when had there been a [stone] action tag?
20090412 17:14:15< zookeeper> are you sure there was one in the first place?
20090412 17:14:46< esr> zookeeper: Dunno.  I'm guessing it had been there for a long time but was mostly hidden by macros.
20090412 17:15:13< esr> Look at the diff for the big commit.  It was there, all right.
20090412 17:15:39< zookeeper> ok
20090412 17:15:42-!- busfahrer [n=busfahre@unixboard/user/busfahrer] has quit ["leaving"]
20090412 17:23:36-!- maxy [n=maxy@80-219-0-199.dclient.hispeed.ch] has joined #wesnoth-dev
20090412 17:24:50< Polarina> What are the procedures for a translation maintainer to add a new translator?
20090412 17:28:15< stikonas> Polarina: just send Ivanovic his name and nick so he will be added to credits
20090412 17:28:26< Polarina> stikonas: That's all?
20090412 17:29:09-!- alink_ [n=alink@78.129.18.215] has joined #wesnoth-dev
20090412 17:29:23< stikonas> Polarina: yes, though you will have to share the work somehow so there will be no conflicts
20090412 17:29:44< Polarina> stikonas: Of course.
20090412 17:31:15< Polarina> I found a person interested in translating too, but he hasn't responded for the last several hours after I told him to try the game first.
20090412 17:31:34< mordante> loonycyborgw32, thanks, can you test how many W it takes before crashing?
20090412 17:32:55< AnMaster> Polarina, is that a good or bad sign?
20090412 17:33:07< stikonas> Polarina: one person interested in Lithuanian translation has contacted me too, but has not send any translations yet.
20090412 17:33:10< Polarina> AnMaster: I don't know. It could be addiction to the game..
20090412 17:33:24< AnMaster> yeah
20090412 17:33:37< AnMaster> (and is that a good or bad thing? :)
20090412 17:33:49-!- Elvish_Pillager [n=eli@66-189-40-122.dhcp.oxfr.ma.charter.com] has joined #wesnoth-dev
20090412 17:33:53< Polarina> AnMaster: To be honest, I'm not really sure.
20090412 17:33:57< AnMaster> heh
20090412 17:34:33-!- ping_a [n=ping_c@56341227.rev.stofanet.dk] has joined #wesnoth-dev
20090412 17:36:55-!- Elvish_Pillage2 [n=eli@66-189-40-122.dhcp.oxfr.ma.charter.com] has quit [Read error: 60 (Operation timed out)]
20090412 17:37:06-!- ikarius [n=ross@216.27.182.3] has joined #wesnoth-dev
20090412 17:40:15< Ivanovic> Polarina: in general the procedure is: you tell the translator what he/she has to do
20090412 17:40:31< Ivanovic> when you get the first update from him, have a look at it and then send it in with your normal updates
20090412 17:40:46< Ivanovic> when sending it in, add how to add the new translator into the credits
20090412 17:41:11< Polarina> Ivanovic: Thank you.
20090412 17:42:29-!- alink [n=alink@wesnoth/developer/alink] has quit [Success]
20090412 17:42:35-!- alink_ is now known as alink
20090412 17:45:15< esr> Ivanovic: There's something screwed up in the translation machinery. po/wesnoth/en_GB.po has a bunch of untranslated strings in it that look like they're from Scepter of Fire::The Dragon.
20090412 17:45:29< Ivanovic> AI0867, alink, boucman, cjhopman_, Crab_, crimson_penguin, dfranke, Dragonking, esr, freim, loonycyborgw32, mordante, silene, Sirp, Soliton, Smar, stikonas, YogiHH, zookeeper: ready for 1.6.1?
20090412 17:45:29< CIA-30> mordante * r34788 /trunk/CMakeLists.txt: 
20090412 17:45:29< CIA-30> Only set the g++ compiler flags for g++.
20090412 17:45:29< CIA-30> There were some problems on MSVC with the g++ compiler flags. This patch
20090412 17:45:29< CIA-30> should fix it, but that's untested. (It's tested that it doesn't break
20090412 17:45:29< CIA-30> g++.)
20090412 17:45:37< CIA-30> mordante * r34789 /trunk/CMakeLists.txt: Make the dummy locales a target.
20090412 17:45:39< Ivanovic> if not please get ready or ping me about blockers
20090412 17:45:50< esr> Ready.
20090412 17:45:54< Ivanovic> esr: do not look at trunk!
20090412 17:46:01< crimson_penguin> Ivanovic: yup
20090412 17:46:01< esr> ?
20090412 17:46:09< Smar> nothing here...
20090412 17:46:19< Ivanovic> yes, trunk is a little screwed because of a textdomain declaration that was borked till yesterday
20090412 17:46:24< mordante> Ivanovic, no I want to part 34789 and add a workaround for Windows problem, should be trivial
20090412 17:46:24< Ivanovic> and thus borked at the last pot-update, too
20090412 17:46:28< esr> Ah, I see.
20090412 17:46:36< stikonas> Ivanovic: can you commit my translation update
20090412 17:46:43< YogiHH> Ivanovic: ready for 1.6.1
20090412 17:46:48< Ivanovic> silene: will do so "soon" (in probably 15mins)
20090412 17:46:55< Ivanovic> s/silene/stikonas
20090412 17:47:01< freim> Ivanovic: can we backport art to 1.6.x from trunk?
20090412 17:47:10< Crab_> Ivanovic: yes
20090412 17:47:26< Ivanovic> freim: as long as no wml is required for the art to work (as in eg new terrain letters and such): sure
20090412 17:48:14< CIA-30> mordante * r34790 /branches/1.6/CMakeLists.txt: Make the dummy locales an option.
20090412 17:48:27< freim> Ivanovic: need only to uncomment some wml which is already there
20090412 17:48:38< loonycyborgw32> mordante: That probably depends on screen resolution :P
20090412 17:48:43< Ivanovic> freim: will it add any new terrain letter or stuff like this?
20090412 17:48:53< Ivanovic> if not it is fine, if it does: no, you are not allowed to do so
20090412 17:49:12< mordante> loonycyborgw32, IIRC that dialog has a fixed size ;-)
20090412 17:49:18< freim> Ivanovic: no, it's just some more snow mountain variants
20090412 17:49:28< mordante> but I'll count under Linux
20090412 17:50:03< silene> Ivanovic: i'm ready
20090412 17:50:45< Ivanovic> freim: then commit it...
20090412 17:51:01< freim> Ivanovic: I have no idea how to work on branches
20090412 17:51:13< Ivanovic> freim: exactly the same way as on trunk
20090412 17:51:25< Ivanovic> that is: instead of a trunk checkout get yourself a checkout of branches/1.6
20090412 17:51:41-!- Psyche^ [n=Psyche@g224106008.adsl.alicedsl.de] has joined #wesnoth-dev
20090412 17:51:50< Ivanovic> s/instead/additional to
20090412 17:51:52< Ivanovic> ;)
20090412 17:53:12-!- _ping_ [n=ping_c@56341227.rev.stofanet.dk] has quit [Read error: 110 (Connection timed out)]
20090412 17:53:15< freim> ok, will take a while to do the checkout
20090412 17:53:18< alink> Ivanovic: I suppose so. I still wish to backport r34777 to 1.6 for preventing any new slowdown of loading there, but I am too tired to decide that today. So it will be tagged without it (unless boucman show up before and double-check it)
20090412 17:53:49< loonycyborgw32> mordante: That takes 30 'a' letters on windows.
20090412 17:54:11< mordante> and how many letters 'W'?
20090412 17:55:00< alink> 30W/a obviously <:o)
20090412 17:55:58< loonycyborgw32> Bzt! Wrong! 18 :P
20090412 17:56:05< alink> and how many \/\/ ?
20090412 17:56:51-!- Turuk_ [n=Turuk@wesnoth/forumsith/turuk] has joined #wesnoth-dev
20090412 17:56:55< mordante> oh that will be short messages ;-) :-P
20090412 17:57:22< mordante> lklyt t l33t spkrs lk tht :P
20090412 17:57:40< loonycyborgw32> It doesn't crash with '\/\/' :P
20090412 17:57:41< mordante> lkly*
20090412 17:58:46< alink> so (float)'W'/'a' = 0.896907
20090412 17:59:23< alink> and  30* (float)'W'/'a' = 26.9
20090412 18:00:14-!- molgrum [n=molgrum@c83-249-53-163.bredband.comhem.se] has quit ["Lämnar"]
20090412 18:00:17 * YogiHH starts to think a complete rebuild would have been faster than letting msvc++ fiddle with the dependencies...
20090412 18:01:25< CIA-30> silene * r34791 /trunk/src/game_events.cpp: Factorized code. Used vconfig::has_attribute. Put conditions as positive.
20090412 18:01:31< CIA-30> silene * r34792 /trunk/src/ (game_events.cpp variable.hpp): Removed almost unused function vconfig::get_attribute.
20090412 18:02:05< Polarina> Almost?!? :O
20090412 18:04:07-!- Turuk [n=Turuk@wesnoth/forumsith/turuk] has quit [Read error: 110 (Connection timed out)]
20090412 18:05:55< zookeeper> Ivanovic, when are you tagging?
20090412 18:06:06< CIA-30> jhinrichs * r34793 /trunk/src/ (6 files): 
20090412 18:06:06< CIA-30> Savegame reorganization Step 1: Providing a simpler interface to saving and loading.
20090412 18:06:06< CIA-30> Move menu_handler::load_game and parts of game_controller::load_game to savegame.cpp.
20090412 18:06:31< CIA-30> ivanovic * r34794 /trunk/ (9 files in 8 dirs): updated German, Lithuanian and Russian translation
20090412 18:06:32< CIA-30> ivanovic * r34795 /branches/1.6/ (10 files in 8 dirs): updated German, Lithuanian and Russian translation
20090412 18:06:39< Ivanovic> zookeeper: the only thing i will do right before is running a pot-update and regenerating the doc files
20090412 18:06:50< Ivanovic> this i will start now and it should be done in maybe 15mins
20090412 18:06:55< Ivanovic> then i would like to release
20090412 18:07:06< Ivanovic> what do you have left on your list, zookeeper?
20090412 18:07:08< zookeeper> ah ok, i'll commit one patch and then update changelogs...
20090412 18:07:14-!- Patterner [n=Psyche@e177225124.adsl.alicedsl.de] has quit [Read error: 110 (Connection timed out)]
20090412 18:07:14-!- Psyche^ is now known as Patterner
20090412 18:07:31< zookeeper> might not be done in 15 mins, but shouldn't take much longer
20090412 18:07:51< Ivanovic> okay
20090412 18:08:23< CIA-30> hogne * r34796 /branches/1.6/data/core/ (10 files in 2 dirs): Backport new snow mountain variants from trunk (r34591 and r34606)
20090412 18:08:33< freim> Ivanovic: done
20090412 18:08:38< Ivanovic> freim: great
20090412 18:08:45< Ivanovic> included a changelog entry?
20090412 18:08:46< Ivanovic> ;)
20090412 18:09:15-!- 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 | 52 bugs, 232 feature requests, 11 patches | logs: http://irclogs.wesnoth.org/ | Don't paste on IRC! Use a pastebin: http://wesnoth.pastebin.com | http://imagebin.org
20090412 18:09:30< freim> Ivanovic: no, forgot
20090412 18:09:42< Ivanovic> then please add one
20090412 18:13:17< freim> hang on
20090412 18:13:20< freim> got a conflict
20090412 18:14:31-!- molgrum [n=molgrum@c83-249-53-163.bredband.comhem.se] has joined #wesnoth-dev
20090412 18:14:43< CIA-30> hogne * r34797 /branches/1.6/changelog: Added entry on new snow mountain variants
20090412 18:16:05< Polarina> http://alokinnollahwe.com/randommusic/my%20music/examples/Renaissance%20Village.mp3 :P
20090412 18:16:33< loonycyborgw32> mordante: Weird. If I input *different* letters in that input box, it doesn't crash :/
20090412 18:16:38< CIA-30> mordante * r34798 /branches/1.6/ (4 files in 2 dirs): 
20090412 18:16:38< CIA-30> Limit the maximum length of the mp command dialog.
20090412 18:16:38< CIA-30> This limit is only imposed on Windows and Mac, since Linux is not affected by
20090412 18:16:38< CIA-30> the problem. This is a workaround for bug #13333.
20090412 18:16:52< mordante> loonycyborgw32, what do you mean?
20090412 18:17:15< mordante> loonycyborgw32, btw I just added a workaround for now
20090412 18:17:18-!- Skizzaltix [n=chatzill@66.92.79.71] has joined #wesnoth-dev
20090412 18:17:45< mordante> Ivanovic, I think I'm done for 1.6... depends whether loonycyborgw32 has more interesting info ;-)
20090412 18:17:55< Polarina> I'm done too. :)
20090412 18:18:04< loonycyborgw32> mordante: I tried randomly hitting keys randomly and it didn't crash.
20090412 18:18:23< cjhopman_> AnMaster: still around?
20090412 18:18:24< loonycyborgw32> Also, '\/\/\/ ...' doesn't cause the crash.
20090412 18:18:39< AnMaster> cjhopman_, yes?
20090412 18:18:39< loonycyborgw32> 'abab...' does cause it though.
20090412 18:18:54< cjhopman_> I just wanted to get a save from that last scenario of NR
20090412 18:18:56< mordante> you mean filling the entire line with \/ ?
20090412 18:18:56< CIA-30> zookeeper * r34799 /trunk/data/core/units/undead/ (Corpse_Soulless.cfg Corpse_Walking.cfg): Applied patch #1157 with little modifications (made the units lacking the variation-specific frames use the variation-independent part of the animation).
20090412 18:19:15< AnMaster> cjhopman_, oh I don't have that around any more, ended up deleting ~/.wesnoth1.7 by mistake
20090412 18:19:19< AnMaster> sorry
20090412 18:19:24< cjhopman_> ok
20090412 18:19:34< mordante> what happens if you do 17 or 18  W's and then ////// ?
20090412 18:20:19< stikonas> Ivanovic: will download links in 1.6 announcement remain the same or will you sed s/1.6a/1.6.1/ so people won't download older versions?
20090412 18:20:23< loonycyborgw32> mordante: Yes. Entire line.
20090412 18:20:44< loonycyborgw32> //// ... causes crash while \\\\\... doesn't.
20090412 18:20:53< Ivanovic> stikonas: i will write a short announcement only like for dev releases
20090412 18:21:05< stikonas> ok
20090412 18:21:14< AnMaster> argh, stop commiting faster than I can recompile ;)
20090412 18:21:50< CIA-30> zookeeper * r34800 /branches/1.6/data/core/units/undead/ (Corpse_Soulless.cfg Corpse_Walking.cfg): Ported r34799 to 1.6.
20090412 18:22:23< Ivanovic> okay, getting read for 1.6 now, running the pot-update preparing for it...
20090412 18:22:36< AnMaster> Ivanovic, someone just commited to it hm?
20090412 18:23:13-!- neary [n=neary@dynamic.rabat2-103-236-12-196.wanamaroc.com] has joined #wesnoth-dev
20090412 18:23:25< loonycyborgw32> mordante: 18 W's followed by /// causes the error, while 18 W's followed by \\\\ doesn't.
20090412 18:23:38 * Ivanovic just ignores AnMaster...
20090412 18:24:05< silene> YogiHH: the loadgame::show_dialog function is a bit confusing; it should always display the dialog box; it should be up to the callers not to call it if they don't want to display the dialog
20090412 18:24:29< silene> loonycyborgw32: doesn't that means it is a memory corruption?
20090412 18:24:37< YogiHH> silene: that's correct, spotted that myself already :-)
20090412 18:24:47< AnMaster> sorry, I'll just shut up
20090412 18:24:53< loonycyborgw32> silene: No idea. Most likely it's a pango bug.
20090412 18:25:22< silene> loonycyborgw32: do you have a backtrace?
20090412 18:26:00< loonycyborgw32> That's not a crash but an exception.
20090412 18:27:00< mordante> loonycyborgw32, I would be really interested in the output of the good and bad case with this patch http://paste.debian.net/33293/
20090412 18:27:14< silene> loonycyborgw32: is that an exception that prevents a backtrace?
20090412 18:27:42-!- edb [n=edb@131.12.95-79.rev.gaoland.net] has joined #wesnoth-dev
20090412 18:27:59< mordante> silene, it's an exception to not terminate the game and instead return in the main menu
20090412 18:28:07< loonycyborgw32> I could use catch throw of course, but I don't even have an up-to-date debug version compiled on windows.
20090412 18:28:32< mordante> this exception is normally only triggered if the WML input is bad
20090412 18:30:02< silene> mordante: even more the reason to catch throw and get a backtrace of it :-)
20090412 18:30:39< loonycyborgw32> Actually, it  looks like an assert.
20090412 18:31:45< loonycyborgw32> But if it's an assert then why it's able to catch it and show a dialog?
20090412 18:32:01< mordante> silene, the backtrace is not that interesting, I know where it comes from only not why
20090412 18:32:19< mordante> it's an exception not an assert
20090412 18:32:40< CIA-30> ivanovic * r34801 /branches/1.6/ (63 files in 11 dirs): 
20090412 18:32:40< CIA-30> pot-update (preparing for 1.6.1, no new/changed strings)
20090412 18:32:40< CIA-30> regenerated doc files
20090412 18:32:59< CIA-30> zookeeper * r34802 /branches/1.6/ (changelog players_changelog): Updated 1.6 changelogs.
20090412 18:33:21< mordante> the most annoying part of the problem is that it can only be reproduced on Mac and Windows
20090412 18:33:36< mordante> and I only have Linux here and also did all testing on Linux
20090412 18:33:49< Ivanovic> everybody ready for 1.6.1?
20090412 18:33:57< mordante> Ivanovic, I am
20090412 18:34:07< Polarina> Ivanovic: Yes.
20090412 18:34:25< Polarina> Is there a page on the wiki that shows the SVN checkout address?
20090412 18:35:09< silene> Polarina: the WesnothSVN page
20090412 18:35:15< Polarina> Found it.
20090412 18:35:26< zookeeper> Ivanovic, i don't think i have anything left..
20090412 18:36:09< zookeeper> ah well, i guess i'll do one more test on that patch to make sure it works right now..
20090412 18:36:20< stikonas> mordante: why don't you try to reproduce the problem in wine?
20090412 18:36:35< Polarina> I'm going to compile the branch from scratch, to see if there are any compiler errors sneaking inside. :)
20090412 18:37:11< mordante> stikonas, I don't have much hdd space left so don't want to use it for wine
20090412 18:38:26< zookeeper> yeah, looks fine..
20090412 18:42:29< CIA-30> silene * r34803 /trunk/src/ (savegame.cpp savegame.hpp): Fixed compilation issues.
20090412 18:42:56< CIA-30> ivanovic * r34804 /branches/1.6/ (changelog configure.ac players_changelog src/wesconfig.h): bump version to 1.6.1
20090412 18:44:39< loonycyborgw32> mordante: Your change cause a genuine assert :P
20090412 18:44:59< loonycyborgw32> And "Width 245 maximum 240 maximum set in pango 245760 ellipse 1." in stderr.txt when WML error happens.
20090412 18:46:21-!- Turuk_ [n=Turuk@wesnoth/forumsith/turuk] has quit [Read error: 104 (Connection reset by peer)]
20090412 18:46:37-!- Turuk [n=Turuk@wesnoth/forumsith/turuk] has joined #wesnoth-dev
20090412 18:47:05< loonycyborgw32> No output if text field is filled without causing the error.
20090412 18:49:13< loonycyborgw32> And the assert is: src/text.cpp line 458 expression 'false'.
20090412 18:51:46< mordante> :-( that was the old patch, I've to check whether I still have the new one around
20090412 18:53:08< Ivanovic> okay, releasing now, wish me luck
20090412 18:53:22< Polarina> A fresh checkout of branches/1.6 compiles without errors.
20090412 18:53:25< Polarina> Ivanovic: Good luck.
20090412 18:53:37< Ivanovic> though reading what loonycyborgw32 i fear that i got to wait a little longer...
20090412 18:54:04< Polarina> Ivanovic: It'll be fixed in 1.6.2.
20090412 18:54:36-!- noy [n=Noy@d75-157-52-251.bchsia.telus.net] has joined #wesnoth-dev
20090412 18:54:42< Ivanovic> tuest me, this is a too common crash to have it in the release...
20090412 18:55:38< Polarina> Just wrap a try...catch around it and you're good to go. :)
20090412 18:55:57< Ivanovic> you have no idea what you are writing, right?
20090412 18:58:02< silene> can someone test if replays work correctly? in particular, is it possible to leave a replay; on trunk, if i select "quit game" in the menu, the replay starts from the beginning again instead of letting me go back to the titlescreen
20090412 18:58:34< crimson_penguin> Polarina: I got all confused and thought you were a Mac user, but now I realized I had no reason to think that
20090412 18:59:48-!- busfahrer [n=busfahre@unixboard/user/busfahrer] has joined #wesnoth-dev
20090412 19:02:35< nagbot> Hi GSoC Applicants!
20090412 19:02:35< nagbot> Time of SoC decision is nearing
20090412 19:02:35< nagbot> We are doing last checks before working on our final "ranking", which we will do *soon*
20090412 19:02:38< nagbot> So, if you want to add anything, post a comment in your proposal
20090412 19:04:23< mordante> loonycyborgw32, commenting out the assert and make the if unconditionally fixes the problem, you want another patch for it?
20090412 19:05:01< mordante> Ivanovic, I added a workaround for that problem and I fear it's an upstream problem
20090412 19:05:19< loonycyborgw32> Why? You've got your output anyway.
20090412 19:05:32< mordante> not of the good case
20090412 19:05:48< loonycyborgw32> assert doesn't happen in good case.
20090412 19:06:16< mordante> I knew that ;-) I want to see the numbers which the patch shows
20090412 19:06:32< mordante> and test whether I get the same numbers on my system for the good case
20090412 19:07:44-!- alink [n=alink@78.129.18.215] has quit [Remote closed the connection]
20090412 19:08:26< loonycyborgw32> I can try a patch that doesn't assert and output something in 'good' case..
20090412 19:10:24-!- Blueblaze [n=nick@c-98-199-143-139.hsd1.tx.comcast.net] has joined #wesnoth-dev
20090412 19:10:58< loonycyborgw32> http://www.wesnoth.org/forum/viewtopic.php?p=351665#p351665 <- perhaps this is an instance of no music infinite loop which was fixed recently?
20090412 19:11:14-!- Turuk_ [n=Turuk@wesnoth/forumsith/turuk] has joined #wesnoth-dev
20090412 19:14:25-!- fendrin [n=fabi@wesnoth/developer/fendrin] has joined #wesnoth-dev
20090412 19:19:23-!- Turuk [n=Turuk@wesnoth/forumsith/turuk] has quit [Connection timed out]
20090412 19:23:10< Ivanovic> mordante, loonycyborgw32: so what is the status, any idea when stuff will be ready for 1.6.1?
20090412 19:24:36< loonycyborgw32> [21:05:01] <mordante> Ivanovic, I added a workaround for that problem and I fear it's an upstream problem
20090412 19:25:06-!- Blueblaze [n=nick@c-98-199-143-139.hsd1.tx.comcast.net] has quit [Remote closed the connection]
20090412 19:26:16< Polarina> Ivanovic: How's the release going?
20090412 19:26:19< Ivanovic> loonycyborgw32: so does it "basically work" on windows, too?
20090412 19:26:30< mordante> Ivanovic, yes
20090412 19:26:49< mordante> just that you can only use 18 chars in that dialog
20090412 19:26:57< Ivanovic> ah, okay
20090412 19:26:57< mordante> but the dialog is not critical
20090412 19:28:33< mordante> loonycyborgw32, I don't think the forum post has anything to do with the infinite loop, since this is a crash not a 'hang'
20090412 19:29:03< loonycyborgw32> Yes. I've also failed to repro it in 1.6a
20090412 19:31:13< silene> mordante: according to the poster, there is a hang with --nosound --nomusic
20090412 19:31:57< mordante> which fixes it at debian iirc
20090412 19:34:44-!- noy [n=Noy@wesnoth/developer/noy] has quit []
20090412 19:35:06< Ivanovic> no idea if the start time params fix it on debian, at least unclicking music in the preferences fixes it on debian
20090412 19:36:25< crimson_penguin> anything I should be testing on Mac?
20090412 19:36:41< Ivanovic> WTF?!?
20090412 19:37:01< crimson_penguin> hmm?
20090412 19:37:19< Ivanovic> problems with the release commands
20090412 19:37:26< crimson_penguin> uh oh
20090412 19:37:34< Polarina> :O
20090412 19:37:37< Ivanovic> problems i never had before (that is before it was basically a warning, nothing more)
20090412 19:42:53< Polarina> I hope you get it resolved soon.
20090412 19:48:21< CIA-30> ivanovic * r34805 /branches/1.6/doc/man/Makefile.am: fix problem with missing files when creating tarball (never needed this before, now it seels to be required, ARGH!)
20090412 19:48:42-!- Ivanovic changed the topic of #wesnoth-dev to: releasing 1.6.1, using r34805 | Summer of Code info: http://www.wesnoth.org/wiki/SummerOfCodeIdeas | 52 bugs, 232 feature requests, 11 patches | logs: http://irclogs.wesnoth.org/ | Don't paste on IRC! Use a pastebin: http://wesnoth.pastebin.com | http://imagebin.org
20090412 19:49:14 * Polarina thinks of maximum compression.
20090412 19:50:10< loonycyborgw32> Ivanovic: Why would it suddenly become needed? automake upgrade caused that?
20090412 19:50:22< YogiHH> silene: how come that the order of calling initializers in the constructor matters?
20090412 19:50:37< boucman> back
20090412 19:50:49< boucman> Ivanovic: what I comitted is good enough for 1.6.1
20090412 19:51:14< CIA-30> ivanovic * r34806 /trunk/doc/man/Makefile.am: port r34805 to trunk (fix dist hook for manpages)
20090412 19:51:24< Ivanovic> loonycyborgw32: probably a new make version or something like this
20090412 19:51:39< loonycyborgw32> YogiHH: It doesn't matter. That's the point :P
20090412 19:51:53< loonycyborgw32> If you mean initializers after :
20090412 19:53:08< YogiHH> loonycyborgw32: that's what i meant and silene rearranged them "to fix compilation issues"
20090412 19:53:57< Dragonking> boucman: I found nice candidate move bug: with 'attack' candidate move AI could attack and kill units that belonged to side allied to AI :)
20090412 19:53:57< loonycyborgw32> No matter in what order you specify them they'll be called in order in which they appear in struct.
20090412 19:54:10< boucman> hehe
20090412 19:54:15< boucman> hope you fixed it :)
20090412 19:54:27< Dragonking> boucman: Yes, in stable
20090412 19:54:30< loonycyborgw32> So specify them in the same order to avoid confusion and bugs caused by it.
20090412 19:54:45< Dragonking> boucman: Not in trunk cause I'm right in a middle of rewriting RCM
20090412 19:54:57< boucman> ok, you know what you're doing...
20090412 19:55:11 * boucman is impatient to see the new RCM
20090412 19:55:39 * YogiHH always thought the computer is doing what he tells him to do :/
20090412 19:55:58< Dragonking> boucman: I already have derived classes, now I'll write candidate move manager which will coordinate everything
20090412 19:58:51< Ivanovic> looking good so far, uploading to files.wesnoth.org while doing some more tests
20090412 19:59:03< Polarina> :D
20090412 19:59:05< CIA-30> jhinrichs * r34807 /trunk/src/ (game.cpp multiplayer_connect.cpp savegame.cpp savegame.hpp): 
20090412 19:59:05< CIA-30> Savegame reorganization Step 1: Providing a simpler interface to saving and loading.
20090412 19:59:05< CIA-30> Start to implement loading multiplayer games in savegame.cpp.
20090412 19:59:42< Ivanovic> puh, 5MB xdelta, probably mainly because of translation updates
20090412 19:59:47-!- edb [n=edb@131.12.95-79.rev.gaoland.net] has quit [Remote closed the connection]
20090412 19:59:57< Polarina> xdelta?
20090412 20:01:02< esr> mordante: Should bug #13333 (Text Input Crashes Wesnoth) be considered fixed at this point?
20090412 20:02:10< stikonas> Polarina: you can download small xdelta file if you have wesnoth-1.6a.tar and create wesnoth-1.6.1.tar
20090412 20:02:10 * Ivanovic tries to summon sapient
20090412 20:02:17< Polarina> Ah...
20090412 20:04:32< silene> YogiHH: yes, initializers are run in the order they are declared; so even if you write ": b(++i), a(++i)", if a is before b, then you will get a = 1 and b = 2; this is so misleading that compilers usually complain loudly about it; that's why my commit was about "fixing compilation issues"; not sure why you compiler didn't complain; did you disable warnings?
20090412 20:05:08< loonycyborgw32> YogiHH uses msvc :P
20090412 20:05:48< loonycyborgw32> Not sure whether it supports that warning and issues it by default.
20090412 20:06:55< Sirp> in Wesnoth our 'coding standard' is meant to be to always write initializer lists with items in the same order they are declared.
20090412 20:07:54< Polarina> How do I run stuff done with "scons test"?
20090412 20:08:09< loonycyborgw32> ./test
20090412 20:08:19< Polarina> Yay!
20090412 20:08:46< Polarina> So short...
20090412 20:09:53< mordante> YogiHH, members are constructed in the order they're declared in the class definition
20090412 20:10:15< mordante> esr, no
20090412 20:10:44 * mordante forgot to scroll down ;-)
20090412 20:10:50< esr> mordante: Noted.  I'm trying to clean up the buglist again.
20090412 20:11:06< mordante> I assumed that
20090412 20:12:00-!- Sapient [n=sapien-x@c-71-204-10-166.hsd1.ga.comcast.net] has joined #wesnoth-dev
20090412 20:12:30< loonycyborgw32> Ivanovic: Your summoning worked :P
20090412 20:12:53< mordante> corn, around?
20090412 20:13:40 * Ivanovic helps summoning corn
20090412 20:13:53-!- JW1 [n=X@c-71-57-85-168.hsd1.il.comcast.net] has quit [Read error: 110 (Connection timed out)]
20090412 20:23:05 * dfranke joins in
20090412 20:23:09< dfranke> Iä! Iä! corn fhtagn!
20090412 20:23:33< Elvish_Pillager> That does not summon.
20090412 20:23:39< CIA-30> ivanovic * r34808 /tags/1.6.1/: tag 1.6.1 (using r34805)
20090412 20:24:03-!- Ivanovic changed the topic of #wesnoth-dev to: released 1.6.1, announcing "soon" | Summer of Code info: http://www.wesnoth.org/wiki/SummerOfCodeIdeas | 52 bugs, 232 feature requests, 11 patches | logs: http://irclogs.wesnoth.org/ | Don't paste on IRC! Use a pastebin: http://wesnoth.pastebin.com | http://imagebin.org
20090412 20:24:45< Ivanovic> ah, and in 15mins even the upload to files.wesnoth.org should be done...
20090412 20:25:20< Polarina> Ivanovic: Yay!
20090412 20:25:37 * mordante things we better use other invocations
20090412 20:26:01< boucman> I have to admit that if that particular summoning works, it would be a bit frightening
20090412 20:26:28< Elvish_Pillager> I've always wanted to know a good literary summoning incantation
20090412 20:26:40< Elvish_Pillager> but I've never heard of any
20090412 20:27:00< boucman> wesbot: seen corn
20090412 20:27:00< wesbot> boucman: Queried user corn is currently in this channel.
20090412 20:27:06 * mordante switched his music so get into a summoning mood
20090412 20:27:30 * boucman his sure he had a black goat somewhere for summoning emergencies
20090412 20:27:39< Dragonking> There is one summoning sentence, that works suprisingly well... along the lines  "I hope that corn never notices that..." ;)
20090412 20:27:49< Ivanovic> boucman: the summoning of corn by dfranke won't work, the one by mordante and me will though
20090412 20:27:50< Ivanovic> ;)
20090412 20:28:35< mordante> :-)
20090412 20:28:58< mordante> I somehow fear corn is fleeing from all the summonings
20090412 20:29:43< dfranke> he's probably still at school
20090412 20:29:47< fendrin> is trunk closed for commits?
20090412 20:29:57< Ivanovic> dfranke: uhm, on easter sunday?
20090412 20:30:02< mordante> dfranke, on easter sunday?
20090412 20:30:04< dfranke> oh. duh.
20090412 20:30:05< dfranke> nm.
20090412 20:30:06< Ivanovic> fendrin: why should it be?
20090412 20:30:12< fendrin> I can't commit
20090412 20:30:45< dfranke> this is what happens when one works a strange schedule at a mostly-Jewish office :-)
20090412 20:31:13< CIA-30> fendrin * r34809 /trunk/data/campaigns/Under_the_Burning_Suns/scenarios/07a_Dealing_with_Dwarves.cfg: UtbS 7a: Added a TODO item.
20090412 20:31:24< fendrin> Okay, it's slow but it's working again.
20090412 20:37:34-!- Blueblaze [n=nick@c-98-199-143-139.hsd1.tx.comcast.net] has joined #wesnoth-dev
20090412 20:40:53< Polarina> Still uploading?
20090412 20:42:59< boucman> Polarina: if you are in a hurry, you could simply "svn up" to the future 1.6.1 release and compile yourself
20090412 20:43:05 * boucman is doing that right now
20090412 20:43:16< Polarina> No, just want to see stuff official. :P
20090412 20:43:43< boucman> then don't hold your breath, we usually wait a couple of days before announcing
20090412 20:43:57< Polarina> Why?
20090412 20:44:33< loonycyborgw32> Because SF is freaking slow during first days after upload.
20090412 20:44:38< Polarina> Oh.
20090412 20:44:52< boucman> to let packagers for the different OS have time to compile wesnoth so most OS are available at release time
20090412 20:47:36< silene> YogiHH: it seems to me you can kill the 16 last lines of multiplayer_lg::load_game; also i don't understand why you marked the function as virtual; do you expect it to be used in a context where the loader type is not known by the compiler?
20090412 20:48:43< silene> YogiHH: you could probably just merge the two classes into one; the inheritance doesn't seem to bring anythin interesting
20090412 20:49:11-!- BenUrban [n=benurban@unaffiliated/benurban] has quit ["Power failu"]
20090412 20:52:21< YogiHH> silene: i'll have a look at that
20090412 20:54:15< fendrin> esr: wmllint output "nonstandard word-wrap style within message" means?
20090412 20:55:41< CIA-30> ivanovic * r34810 /branches/1.6/ (changelog configure.ac players_changelog src/wesconfig.h): post release version bump to 1.6.1+svn
20090412 20:56:48-!- oak_ [n=oakus@252.108.broadband12.iol.cz] has joined #wesnoth-dev
20090412 21:00:04-!- Torangan [n=Torangan@mue-88-130-54-149.dsl.tropolys.de] has joined #wesnoth-dev
20090412 21:01:58-!- neary [n=neary@unaffiliated/neary] has quit [Read error: 110 (Connection timed out)]
20090412 21:03:16< Torangan> hi
20090412 21:03:36< Torangan> mordante: have there really been no updates to WesCamp 1.6 yet?
20090412 21:04:03< mordante> Torangan, I haven't gotten around setting it up and testing for 1.6 yet :-(
20090412 21:04:27< Torangan> Ah ok, I wondered because the directory got created.
20090412 21:04:36< mordante> not much time lately and most of it went to gsoc
20090412 21:04:41< mordante> I didn't do that :-/
20090412 21:04:59< Torangan> Who did?
20090412 21:05:17< Torangan> Ah, Ivanovic did. :-)
20090412 21:05:46< Ivanovic> i did to fix something with g.w.o
20090412 21:06:45-!- Turuk_ [n=Turuk@wesnoth/forumsith/turuk] has quit [Connection reset by peer]
20090412 21:06:46< Torangan> Well, only a few languages have 1.6 complete and are ready to focus on UMC :-)
20090412 21:06:52< mordante> :-)
20090412 21:07:00-!- Turuk [n=Turuk@wesnoth/forumsith/turuk] has joined #wesnoth-dev
20090412 21:07:07< Ivanovic> jupp
20090412 21:07:10< mordante> I hope to find some time soon
20090412 21:07:26< mordante> (and time to actually write some code :-/ )
20090412 21:07:49< Torangan> That much work for gsoc?
20090412 21:08:49< boucman> student evaluation is taking a lot of our time, yes
20090412 21:09:29< mordante> yes
20090412 21:09:57< mordante> but also some other RL things which eat time
20090412 21:13:20< dfranke> Something we should look into at some point: http://scan.coverity.com/index.html
20090412 21:14:03< dfranke> Coverity provides a free web service for open source projects to run their code through Coverity Prevent.
20090412 21:14:30< dfranke> I've never used Prevent, but I've heard a lot of raving about it.
20090412 21:15:10< CIA-30> fendrin * r34811 /trunk/data/campaigns/Legend_of_Wesmere/scenarios/01_The_Uprooting.cfg: LoW 1: Early finishing bonus only when the southern orc leader is defeated. Story for this is still missing.
20090412 21:16:15< mordante> dfranke, agreed, I wanted to have a look at it once, but never found time to look at it
20090412 21:17:00< Ivanovic> corn: alive and around?
20090412 21:17:10-!- ping_e [n=ping_c@56341227.rev.stofanet.dk] has joined #wesnoth-dev
20090412 21:17:22< Ivanovic> corn: you still wanted to show us some code (the code from your prototype implementation)
20090412 21:17:37-!- gabm [n=gabriel@72.0.215.2] has joined #wesnoth-dev
20090412 21:17:53-!- ancestral [n=martinpr@97-116-111-81.mpls.qwest.net] has joined #wesnoth-dev
20090412 21:18:05-!- _ping_ [n=ping_c@56341227.rev.stofanet.dk] has joined #wesnoth-dev
20090412 21:18:49-!- zitterbart [n=beard@dslb-092-072-047-067.pools.arcor-ip.net] has joined #wesnoth-dev
20090412 21:19:02< zitterbart> hi
20090412 21:20:43< Sapient> yo zitterbart
20090412 21:20:50-!- Sirp [n=me@wesnoth/developer/dave] has quit [Read error: 113 (No route to host)]
20090412 21:24:25< esr> fendrin: checking...
20090412 21:24:36-!- oak_ [n=oakus@252.108.broadband12.iol.cz] has quit ["( www.nnscript.de :: NoNameScript 4.02 :: www.XLhost.de )"]
20090412 21:24:39-!- JW1 [n=X@c-71-57-85-168.hsd1.il.comcast.net] has joined #wesnoth-dev
20090412 21:28:30-!- fendrin [n=fabi@wesnoth/developer/fendrin] has quit [Remote closed the connection]
20090412 21:28:53< esr> fendrin: that means wmllint sees somrthing that looks like it might be an unbalanced string, but can't quite be sure.
20090412 21:30:39-!- oak_ [n=oakus@252.108.broadband12.iol.cz] has joined #wesnoth-dev
20090412 21:30:44< silene> did i hear things? do berserker units laugh madly when they are about to slaughter an enemy?
20090412 21:30:59< boucman> silene: yes, it has been around for some time now
20090412 21:30:59< mordante> a dark adept
20090412 21:31:10< boucman> only if the unit can't retaliate (so only DA in mainline)
20090412 21:31:29< boucman> all hail the mighty animation engine :P
20090412 21:32:01< zitterbart> i like this idea of ai arena
20090412 21:32:01< silene> okay, if it only happens against da, that explains why it's the first i hear it
20090412 21:32:43< mordante> yes it's a kind of easter egg, so you found the right day to discover it ;-)
20090412 21:32:54< boucman> zitterbart: yes, we need it to test AI better
20090412 21:34:09< zitterbart> can you also confront two ais to see which works better?
20090412 21:34:49< boucman> yes you can
20090412 21:35:24-!- ping_a [n=ping_c@56341227.rev.stofanet.dk] has quit [Read error: 110 (Connection timed out)]
20090412 21:35:25-!- ping_e [n=ping_c@56341227.rev.stofanet.dk] has quit [Read error: 110 (Connection timed out)]
20090412 21:35:47< zitterbart> that's pretty cool
20090412 21:35:54< boucman> :)
20090412 21:36:02< boucman> thank Crab_ for that :)
20090412 21:36:03< silene> note that confronting ais is nothing new, there are even command-line options to do it
20090412 21:36:46< zitterbart> as i see, the idea of evolutionary ai i just had has already been taken :)
20090412 21:38:22< Dragonking> boucman: Hmm... I'm not sure wheter I finish refractoring of CM today, there were some unnecessary things in the existing code, still looking at what can be improved
20090412 21:39:21< boucman> no problem
20090412 21:39:30< boucman> do you intend to work on it tomorow ?
20090412 21:39:35< Dragonking> Yes
20090412 21:39:40< boucman> ok, cool
20090412 21:39:47< boucman> I should be around most of the day too
20090412 21:40:03< Sapient> CM = ?
20090412 21:40:16< boucman> Candidate Moves
20090412 21:40:27< boucman> it's a way to give the AI "good ideas"
20090412 21:40:43< boucman> like having a special case for units with a cerain ability
20090412 21:40:50< boucman> (poisoning will be handled that way)
20090412 21:41:01< CIA-30> fendrin * r34817 /trunk/data/tools/Makefile: added a rule for checking LoW only.
20090412 21:41:02< CIA-30> fendrin * r34816 /trunk/data/campaigns/Legend_of_Wesmere/scenarios/ (3 files): 
20090412 21:41:02< CIA-30> LoW 21,22,23
20090412 21:41:02< CIA-30> Recoded the recall list management to reflect the changes to scenario 19.
20090412 21:41:02< CIA-30> New story: Kalenz former army is splitt in two, two thirds supports Landar, one third is still loyal to
20090412 21:41:04< CIA-30> Kalenz. The player has to slay the army he growed so far, what a happy thought ;-)
20090412 21:41:06< CIA-30> fendrin * r34814 /trunk/data/campaigns/Legend_of_Wesmere/scenarios/19_Costly_Revenge.cfg: 
20090412 21:41:08< CIA-30> LoW 19:
20090412 21:41:10< CIA-30> Stored Cleodil in her own variable to avoid ugly code for restoring her but not the other fairies in 21.
20090412 21:41:12< CIA-30> fendrin * r34815 /trunk/data/campaigns/Legend_of_Wesmere/scenarios/20_Council_ruling.cfg: LoW 20: Added a note about early finishing bonus to the objectives.
20090412 21:41:17< CIA-30> fendrin * r34813 /trunk/data/campaigns/Legend_of_Wesmere/scenarios/07_Elves_last_stand.cfg: 
20090412 21:41:19< CIA-30> LoW 7:
20090412 21:41:21< CIA-30> code cleanups.
20090412 21:41:23< CIA-30> extra scenario objectives at Olurf's return.
20090412 21:41:25< CIA-30> Reenabled the dialogue for stronger/weaker ai with better explanation. (still needs to be reviewed by a
20090412 21:41:27< CIA-30> natural english speaker)
20090412 21:41:29< CIA-30> Added hints about the gold carryover to the objectives.
20090412 21:41:35< dfranke> mordante: I just assigned you a bug.  https://gna.org/bugs/index.php?13359
20090412 21:41:41< boucman> fendrin is in a commit spree
20090412 21:42:03-!- Crab_ [i=crab@wesnoth/developer/crab] has quit ["Leaving."]
20090412 21:42:07< zitterbart> boucman, i heard you are working on that leadership ai-improvement from the wiki
20090412 21:42:10< CIA-30> ivanovic * r34818 /trunk/po/wesnoth/id.po: updated Indonesian translation
20090412 21:42:18< boucman> am I ?
20090412 21:42:22< CIA-30> ivanovic * r34819 /branches/1.6/ (changelog players_changelog po/wesnoth/id.po): updated Indonesian translation
20090412 21:42:30< boucman> no, I suggested it as a way to learn about formula AI
20090412 21:42:37< boucman> but I'm not working on it mysel
20090412 21:42:38< boucman> f
20090412 21:42:57< Dragonking> I would like to work on it once I have some time.
20090412 21:43:16< boucman> Dragonking: noted
20090412 21:43:27< mordante> dfranke, thanks, I think it's a duplicate did you search?
20090412 21:43:32< boucman> Dragonking: you might want to state that in the EasyCoding page, then
20090412 21:43:35-!- Skizzaltix [n=chatzill@66.92.79.71] has quit ["ChatZilla 0.9.84 [Firefox 3.0.8/2009032608]"]
20090412 21:44:08< CIA-30> jhinrichs * r34820 /trunk/src/ (multiplayer_connect.cpp savegame.cpp savegame.hpp): 
20090412 21:44:08< CIA-30> Savegame reorganization Step 1: Providing a simpler interface to saving and loading.
20090412 21:44:08< CIA-30> Cleaning up multiplayer code in savegame.cpp.
20090412 21:44:16< Dragonking> boucman: Well, that are just my plans, if someone wishes to do something about it, he can. Maybe something good will emerge from that :)
20090412 21:44:26< boucman> :)
20090412 21:44:30< stikonas> wesnoth is project #7 at cia.vc today
20090412 21:45:09< Polarina> :)
20090412 21:45:13< zitterbart> i would like to give it a try
20090412 21:45:32< zitterbart> nevertheless, i'm just working into the source and can't promise anything
20090412 21:45:43< dfranke> mordante: yes; I don't see any duplicate.
20090412 21:46:10< boucman> zitterbart: ok, in that case, have a look at data/ai/formula/ (the poison stuff should be a good place to start) and at the FormulaAI related wiki pages
20090412 21:47:35< Sapient> cya later, have fun.
20090412 21:47:35-!- Sapient [n=sapien-x@wesnoth/developer/sapient] has left #wesnoth-dev []
20090412 21:47:48< zitterbart> i'll do that, thanks
20090412 21:48:16-!- fendrin [n=fabi@wesnoth/developer/fendrin] has joined #wesnoth-dev
20090412 21:48:32< boucman> and feel free to ask questions
20090412 21:48:43< silene> YogiHH: in load_mp_game, you are still testing twice for version compatibility, the first time being when you call load_game; that's why i said you could simply remove it; similarly for load errors, just get rid of it
20090412 21:48:51< CIA-30> mordante * r34821 /trunk/data/campaigns/Legend_of_Wesmere/scenarios/04_Elvish_Treasury.cfg: Fix capitalization.
20090412 21:49:18< fendrin> svn update
20090412 21:49:28< boucman> fendrin: wrong window :)
20090412 21:49:42< Ivanovic> hmm, still no corn
20090412 21:49:51< fendrin> :-)
20090412 21:49:55< Ivanovic> though i just had some bbq and corn really tastes great
20090412 21:50:17< mordante> dfranke, ok, well if it is, it will be assigned to me as well... so will find out later
20090412 21:50:26< YogiHH> silene: that load_game call is calling gamestatus.cpp:load_game, it is not testing for version compatibility
20090412 21:50:52< mordante> drat, I knew I should have used the resurrection version of the summoning :-P
20090412 21:50:56< silene> YogiHH: oh, sorry, i thought it was calling the one from the class
20090412 21:51:03< YogiHH> silene: i am just about to move that one now
20090412 21:52:16< CIA-30> mordante * r34822 /branches/1.6/data/campaigns/Legend_of_Wesmere/scenarios/04_Elvish_Treasury.cfg: Fix capitalization.
20090412 21:55:49< gabm> YogiHH: hi
20090412 21:55:59< YogiHH> hi gabm
20090412 21:56:16< gabm> YogiHH: I've been looking through the source code
20090412 21:56:31-!- Turuk [n=Turuk@wesnoth/forumsith/turuk] has quit [Read error: 104 (Connection reset by peer)]
20090412 21:56:42< gabm> can I throw random questions at you? :)
20090412 21:56:47-!- Turuk [n=Turuk@wesnoth/forumsith/turuk] has joined #wesnoth-dev
20090412 21:57:36< YogiHH> gabm: sure
20090412 21:57:52< gabm> by the way, I'm on track to submit an updated proposal by Monday evening
20090412 21:57:53-!- loonycyborgw32 [i=kvirc@79.139.246.210] has quit ["KVIrc 3.4.0 Virgo http://www.kvirc.net/"]
20090412 21:59:46< gabm> ok, question 1: I see many things are duplicated in the save file - some of that is necessary so that WML change change the map on-the-fly. But might we consider storing only what changed, instead of a copy of everything?
20090412 22:00:18-!- loonybot [n=loonybot@79.139.246.210] has joined #wesnoth-dev
20090412 22:00:28< gabm> Some attributes like "recruit" are at least at three different places :P
20090412 22:00:51-!- loonycyborg [n=sergey@79.139.246.210] has joined #wesnoth-dev
20090412 22:01:53< YogiHH> gabm: most of the duplicates between [snapshot] and [replay_start] are necessary, as they represent different gamestates: The current turn (snapshot) and the start of the scenario (replay_start).
20090412 22:02:22< YogiHH> gabm: there are still some values that will never change, those can be extracted, for example to the root level
20090412 22:02:41< YogiHH> gabm: your idea about detecting differences and only storing them is a good one
20090412 22:03:06< YogiHH> gabm: but it is most likely not feasible for GSoC because of the time frame
20090412 22:03:17< gabm> YogiHH: even for things that might change, can't we wait for them to change before writing them somewhere else than the root?
20090412 22:03:33< YogiHH> gabm: For example, this could be done to unit WML, which uses up a lot of space
20090412 22:03:42< gabm> YogiHH: right, you're probably more aware than me of what's feasible in the GSoC
20090412 22:04:07< YogiHH> gabm: i'd love to do that, but we need to get some structure into the savegame first :-)
20090412 22:04:19< gabm> :)
20090412 22:05:50< gabm> YogiHH: Ok, next question (2) : Unless it changed recently, it's not possible to change langage mid-campaign, because unit names and stuff are stored "localized"... could we fix this along the way?
20090412 22:07:17< gabm> YogiHH: I tried to change from french to english mid-campaign in HTTT, to know which units people were talking about on the forum, but no go.
20090412 22:07:26< silene> gabm: are you sure? as far as i remember, unit names have always been stored with their untranslated form
20090412 22:07:36< silene> (the automatically-generated ones are not though)
20090412 22:07:59< YogiHH> gabm: Well, to be honest, i have no idea :-). I'd have to check with someone who is more into WML to see what could be done with that.
20090412 22:08:00< Chusslove> Ivanovic: You might consider upgrading po4a to 0.36; current extraction of XML path into translator comments (# type: ...) is wrong, and should be extracted comment (#. type: ...), as is fixed in 0.36. 
20090412 22:08:26< gabm> silene: I don't remember precisely, but the ones in the recall list seemed to keep their localized name, and other stuff was messed up
20090412 22:09:16< gabm> YogiHH: ok - anyways that would be a secondary goal for me, it seems "in-scope"
20090412 22:11:32< gabm> YogiHH: question 3! : I read the thread "Multiplayer campaign HOWTO" on the forum, and making on seems a very hackish job. How would my work come into the picture? Do you plan on improving the presentation of multiplayer campaigns, and especially the way to declare them in scenario files?
20090412 22:11:51< gabm> YogiHH: it's kind of circular dependency, isn't it?
20090412 22:12:39< gabm> YogiHH: *"making a multiplayer campaign", not "making on" lol
20090412 22:12:39< zitterbart> boucman: which file contains the poisoning stuff?
20090412 22:12:41< YogiHH> gabm: To first answer the question: Yes, i plan to do that
20090412 22:13:12< boucman> poisoner_attack.fai and poisoner_eval.fai
20090412 22:13:18< boucman> (these are trunk only)
20090412 22:13:39< YogiHH> gabm: I don't think it's a circular dependency, you can basically view a regular mp scenario as a 1-scenario-campaign and a mp campaign as a regular campaign with network support.
20090412 22:14:23< CIA-30> fendrin * r34812 /trunk/data/campaigns/Legend_of_Wesmere/scenarios/03_Kalian.cfg: LoW 3: Added note to the objectives for the early finishing bonus.
20090412 22:14:26< zitterbart> ah, i see
20090412 22:14:39< YogiHH> gabm: I want it to become such that there is no real difference between the savegame formats used for mp and single player, just some additional stuff for mp.
20090412 22:15:00< YogiHH> gabm: like for example the multiplayer settings
20090412 22:15:12< gabm> YogiHH: I was under the impression that the scenario files had to change a lot for full support of MP campaigns, and that savegame format would depend on that change
20090412 22:15:34-!- Alesis-Novik [n=alesis@78.60.188.59] has joined #wesnoth-dev
20090412 22:16:00< gabm> YogiHH: but you seem to be saying it won't be that bad
20090412 22:16:00-!- maxy [n=maxy@80-219-0-199.dclient.hispeed.ch] has quit []
20090412 22:16:31< YogiHH> gabm: I think scenario files are pretty fine, they just don't get supported well on the server.
20090412 22:16:32< silene> gabm: hmm... after checking, you may be right, seems like someone broke unit names
20090412 22:16:58< gabm> silene: heh, I remember that from ... a lot of versions ago
20090412 22:17:03< YogiHH> gabm: I may be wrong, but i think the biggest problem is correct synchronization for mp campaigns
20090412 22:17:20< Dragonking> boucman: BTW, do you think we need whole "register_candidate_move", or we could make it shorter, like "candidate_move" ? Also this paremeter called "eval_list" which need to be set in cfg to "yes" in order to use CM... first: do we really need it (I know it can be used to quicly disable CM usage, but what do oyu think?) 2nd: maybe some better name too... ?
20090412 22:17:53< Dragonking> I'm currently putting all CM config parsing to candidate_move_manager constructor
20090412 22:18:06< boucman> RCM => CM : ok, but you might want to tell esr about it for wmllint
20090412 22:18:20< gabm> YogiHH: do savegame files play a role in synchronization?
20090412 22:18:25-!- Crab_ [n=Crab_@wesnoth/developer/crab] has joined #wesnoth-dev
20090412 22:18:29-!- Crab_ [n=Crab_@wesnoth/developer/crab] has left #wesnoth-dev []
20090412 22:18:32< boucman> and probably drop that parameter, yes
20090412 22:18:52< Dragonking> ok
20090412 22:19:14< YogiHH> gabm: not really
20090412 22:19:29< gabm> YogiHH: one less thing to worry about, then
20090412 22:19:37< YogiHH> gabm: however, mp savegames are different from single player and that is one source of the problems
20090412 22:19:43< zitterbart> so, as i'm free to ask questions :), how does the ai know if poisoning an enemy is better than killing him with an high-exp-unit? by the assigned score?
20090412 22:20:00< YogiHH> gabm: that's why i want them to be using the same format
20090412 22:20:23< gabm> YogiHH: I see - as I understand, it's the main motivation for this work
20090412 22:20:49< YogiHH> gabm: yes, it is
20090412 22:21:23< YogiHH> gabm: The other one is, that even i don't like to touch the savegame code because it is extremely easy to break something
20090412 22:21:29-!- fabi [n=fabi@g227079101.adsl.alicedsl.de] has joined #wesnoth-dev
20090412 22:21:32< boucman> zitterbart: when we have the whole AI working, yes, it will be based on the score, but atm (as you can see) we only use 1 as the returned value
20090412 22:21:34-!- Turuk_ [n=Turuk@wesnoth/forumsith/turuk] has joined #wesnoth-dev
20090412 22:21:40< YogiHH> gabm: and very very time consuming to test
20090412 22:21:40< boucman> this code is still very experimental
20090412 22:21:55< gabm> YogiHH: question 4: I don't remember if there's a multiplayer savegame feature?
20090412 22:22:16< gabm> YogiHH: and if so, how do we support that?
20090412 22:22:21< zitterbart> boucman: is the position of the units after the move also taken into account?
20090412 22:22:29 * loonycyborg kicks sourceforge
20090412 22:22:36-!- Crab_ [n=Crab_@wesnoth/developer/crab] has joined #wesnoth-dev
20090412 22:22:40< boucman> zitterbart: nope
20090412 22:22:56< YogiHH> gabm: a multiplayer savegame feature? What do you mean by that?
20090412 22:22:58< CIA-30> fendrin * r34823 /trunk/data/campaigns/Legend_of_Wesmere/scenarios/ (13 files): 
20090412 22:22:58< CIA-30> LoW all: Code cleanups.
20090412 22:22:58< CIA-30> Added missing objective notes to every scenario in LoW.
20090412 22:23:04< boucman> it's basically an "all poisoning units will use poison as much as possible" algorithm at this point
20090412 22:23:21< gabm> I mean, in a MP scenario or campaign, can players save the game to resume at a later time
20090412 22:23:39< Dragonking> boucman: Heh.. that made me remember that I have improved poisoner eval here..
20090412 22:23:45< zitterbart> okay
20090412 22:23:50< boucman> Dragonking: ready for commit ?
20090412 22:24:09< gabm> YogiHH:^^
20090412 22:24:23< Dragonking> boucman: Nothing fancy, just making leader poisoning more important ald ignorng units in villages
20090412 22:24:27< Dragonking> boucman: I think so
20090412 22:24:44< boucman> well, go for it, then :)
20090412 22:24:44-!- ancestral [n=martinpr@97-116-111-81.mpls.qwest.net] has left #wesnoth-dev []
20090412 22:24:51< YogiHH> gabm: Yes of course, they can. And it's nothing we have to implement, this works perfectly fine.
20090412 22:25:20< zitterbart> but at least the chance to die after leveling up (lonely level2-unit facing lots of level 3-units) might be important
20090412 22:25:21< gabm> YogiHH: good - I wasn't sure, it might have been something to keep in mind, or improve support for
20090412 22:26:14< boucman> zitterbart: well, you could add something like that before attacking leadership, leadership is quite complicated and warming up on something simpler is  good idea
20090412 22:26:21< crimson_penguin> is there some huge file updated? my svn up is taking a looong time, after U    data/core/images/story/landscape-battlefield.jpg
20090412 22:26:31< crimson_penguin> (actually, not up, svn switch)
20090412 22:26:51< zitterbart> sure :)
20090412 22:27:03< Dragonking> boucman: Done
20090412 22:27:04< CIA-30> dragonking * r34824 /trunk/data/ai/formula/poisoner_eval.fai: Small poisoner_eval formula improvement
20090412 22:27:17< gabm> YogiHH: question 5: there's a section I find weird because of the duplication: [trait]
20090412 22:27:35< gabm> YogiHH: you get first female_name = "female^loyal"
20090412 22:27:35< boucman> Dragonking: thx
20090412 22:27:44< gabm> YogiHH: then id = "loyal"
20090412 22:27:45< boucman> zitterbart: see Dragonking's last commit
20090412 22:28:00< gabm> YogiHH: then male_name = "loyal"
20090412 22:28:08< CIA-30> silene * r34825 /trunk/src/unit.hpp: Fixed units being incorrectly known by their translated name.
20090412 22:28:13< gabm> YogiHH: Is that a candidate for merging?
20090412 22:28:13 * Dragonking hopes it really works like he thinks it does... ;p
20090412 22:28:56< boucman> looks good from here...
20090412 22:29:25< gabm> YogiHH: oh - unless there's something with localization... in french, loyal is "loyal" when masculine, and "loyale" when feminine
20090412 22:29:26< YogiHH> gabm: i think this is translation support for units which have a male and female variant. But better ask zookeeper or sapient about it (shadow_master, esr and a few others might know as well)
20090412 22:29:39-!- Turuk [n=Turuk@wesnoth/forumsith/turuk] has quit [Read error: 110 (Connection timed out)]
20090412 22:30:58< CIA-30> silene * r34826 /branches/1.6/src/unit.hpp: Fixed units being incorrectly known by their translated name.
20090412 22:31:28< Dragonking> boucman: We also could add bonus agaisnt units that are on good def hex and malus agaisnt users with magic/marksman retaliation
20090412 22:31:46< boucman> Dragonking: yes, we could do all sorts of things :)
20090412 22:32:15< boucman> but I'd rather have you work on the CM refactoring and let zitterbart make his teeth on simple stuff like the one you suggest
20090412 22:32:20< boucman> :P
20090412 22:32:21-!- silene1 [n=plouf@ASte-Genev-Bois-152-1-31-188.w83-114.abo.wanadoo.fr] has joined #wesnoth-dev
20090412 22:32:24< gabm> zookeeper: Hi, I was wondering what's the purpose of the "female^" prefix in the female_name tag.
20090412 22:32:32< Dragonking> boucman: Sure, just throwing an idea :)
20090412 22:32:48< boucman> well, continue throwing,then ;)
20090412 22:32:53< gabm> zookeeper: see my discussion with YogiHH a few lines above
20090412 22:33:04< silene1> gabm: thanks for reporting the issue; names should now be properly localized; but obviously it won't solve the issue for existing savegames
20090412 22:33:21-!- silene [n=plouf@wesnoth/developer/silene] has quit [Nick collision from services.]
20090412 22:33:26-!- silene1 is now known as silene
20090412 22:33:30< gabm> silene1: wow, that was a fast fix :)
20090412 22:34:22-!- nital [n=nital@unaffiliated/nital] has quit [Connection reset by peer]
20090412 22:34:26< mordante> gabm, the female^ prefix is needed to make a difference between a male and female unit name
20090412 22:34:27< silene> gabm: yes, it was a trivial fix; it just that someone rewrote the code and did use the wrong type to store the name; so it was a one-liner
20090412 22:34:38< mordante> matters for some translations
20090412 22:34:52< YogiHH> gabm: anything more you want to know?
20090412 22:35:07< gabm> YogiHH: thanks, that's all for now
20090412 22:35:12< Dragonking> boucman: Generally we could also look if target is next ot healer or not, for example giving such unit score 2, unit on grassland would get score 3 unit on good def hex score 4 unit with good retaliation score 1 etc.. :)
20090412 22:35:16< YogiHH> ok, i am off then
20090412 22:35:20< YogiHH> good night everyone
20090412 22:35:24< mordante> night YogiHH 
20090412 22:35:28< gabm> YogiHH: night
20090412 22:35:36< Dragonking> night YogiHH 
20090412 22:35:41< fabi> bye
20090412 22:35:50< boucman> Dragonking: yes, could you write that down as easy coding tasks, so we don't forget ?
20090412 22:35:53-!- YogiHH [n=chatzill@c214211.adsl.hansenet.de] has left #wesnoth-dev []
20090412 22:36:14< gabm> mordante: it's not enough to know that the name is in a female_name attribute?
20090412 22:36:29< Dragonking> boucman: Sure
20090412 22:36:45< mordante> for gettext there need to be two different strings otherwise it considers them the same
20090412 22:36:45< Dragonking> boucman: Once I get tired by C++ :)
20090412 22:36:54< boucman> :P
20090412 22:36:59< mordante> also for the translator it's good to know the gender of the unit
20090412 22:37:04-!- fendrin [n=fabi@wesnoth/developer/fendrin] has quit [Connection timed out]
20090412 22:37:13< gabm> mordante: thanks, I understand better now
20090412 22:37:58< mordante> you're welcome
20090412 22:38:38< gabm> mordante: by the way, I'll have updated my proposal and probably submitted a patch by Monday midnight UTC - don't know if you saw my comment on the proposal page
20090412 22:39:31-!- nital [n=nital@mion.elka.pw.edu.pl] has joined #wesnoth-dev
20090412 22:42:51-!- heinz [n=heinz@p54B318F5.dip0.t-ipconnect.de] has joined #wesnoth-dev
20090412 22:43:26< mordante> gabm, I do now ;-)
20090412 22:43:43< heinz> with branch/1.6 after a finished game the stats I see are the ones of the first player, no matter whether that was me or someone else. is that to be expected?
20090412 22:43:47< gabm> mordante: perfect :)
20090412 22:45:40< silene> heinz: expected? sure, since wesnoth is (almost) determistic; useful? probably not ;-)
20090412 22:46:49< heinz> silene: I guess I can  follow this line of argumention ;-) (even if I believe there was some randomness involved somewhere...)
20090412 22:47:09< Crab_> Dragonking: I will start (now) replacing 'direct changes of gamestate' with 'using specific abstraction layer for game changes'. I will not touch the old interface, and I'll replace current unsafe gamestate changes one-by-one. 
20090412 22:47:17< Crab_> What is the preferred action to do if AI asks the game to do something stupid (such as 'ordering one of its units to attack an ally) ? is "log and do nothing" ok for now ?
20090412 22:48:25< silene> heinz: that being said, regardless of this bug, it would probably make sense to display all the statistics at the end, not just yours
20090412 22:48:25< heinz> silene: also, in such situation, a 'team chat message' will result in a 'illegal message removed' reply from the 'server'
20090412 22:48:39< Dragonking> Crab_: I think it is ok for now, maybe boucman has something to add.
20090412 22:48:41< boucman> Crab_: if you do that, try to add a keyword or use a specific log so we can easily grep through
20090412 22:48:42< heinz> silene: that would be even cooler 
20090412 22:48:43< silene> interesting
20090412 22:48:50< boucman> ok for me too
20090412 22:49:03< Crab_> boucman: I've added a specific log category for ai_actions
20090412 22:49:27-!- Torangan [n=Torangan@mue-88-130-54-149.dsl.tropolys.de] has quit [Remote closed the connection]
20090412 22:49:29< boucman> ok, if it allows to find all the places to plug easily, I'm fine with it
20090412 22:49:31< Crab_> boucman: and it is *error-level* log message. Since the ai is not supposed to do stupid moves.
20090412 22:49:36 * Ivanovic really wants to get rid of ./configure based bulding really soon
20090412 22:49:40< boucman> makes sense
20090412 22:49:47< boucman> Ivanovic: seconded
20090412 22:50:01< Ivanovic> that is: i think the next dev release will require autogen if you want to build using configure
20090412 22:50:20< Ivanovic> you have to add every f***ing file explicitly to the tarball to have it in
20090412 22:51:49< dfranke> Ivanovic: thirded, fourthed, and ω₀ed.
20090412 22:52:20< Dragonking> Hmm.... very..... "unique" char... ω
20090412 22:52:45< dfranke> It's a lower-case omega.  Omega naught is the first transfinite ordinal.
20090412 22:52:47< heinz> Dragonking: omega?
20090412 22:53:40< Dragonking> Heh... I know, just maybe after too much hours of coding I see something more than that... nevermind ;)
20090412 22:54:55< dfranke> If it helps autotools get removed sooner, I'll throw in some uncountables too.
20090412 22:55:23< mordante> well I think it will be hard to remove it as long as we have developers using it ;-)
20090412 22:55:24< CIA-30> ivanovic * r34827 /branches/1.6/Makefile.am: add missing file to tarball generation (f*** autotools!!!)
20090412 22:55:24< CIA-30> ivanovic * r34828 /trunk/Makefile.am: add missing file to tarball generation (f*** autotools!!!)
20090412 22:55:33-!- dapounet [n=user@cust-48-149-111-94.dyn.versateladsl.be] has joined #wesnoth-dev
20090412 22:55:59-!- Elvish_Pillage2 [n=eli@66-189-40-122.dhcp.oxfr.ma.charter.com] has joined #wesnoth-dev
20090412 22:56:23< loonycyborg> mordante: I think removing configure and stuff from tarball will be enough.
20090412 22:56:24< mordante> of course we can require users to install autogen for using the tarball
20090412 22:56:43< loonycyborg> So the tarball could be svn snapshot.
20090412 22:56:48< mordante> loonycyborg, agreed ;-)
20090412 22:56:57< Ivanovic> yes, the tarball for the next dev release will be a plain svn snapshot
20090412 22:57:17-!- nagbot [n=nagbot@smtp.terraninfo.net] has quit ["leaving"]
20090412 22:57:20< mordante> for scons and cmake we also require extra tools to be installed
20090412 22:58:21< Ivanovic> due to autotools making packaing overly complecated building using cmake ain't working in the tarball i just uploaded
20090412 22:58:25< Ivanovic> ARGH!
20090412 22:58:58< mordante> why does cmake not work?
20090412 22:59:08< Ivanovic> config.h.cmake is not in the package
20090412 22:59:39< loonycyborg> Currently I can't login to sourceforge to upload files because it seems to have technical issues :/
20090412 22:59:39< silene> mordante: we don't even have to require users to install autogen, since we already ship it; we just have to tell them to run it (and imo we don't even have to tell them that; it would be as if we were telling scons users that they are supposed to run scons)
20090412 23:00:17< mordante> silene, I meant the tools needed to run autogen
20090412 23:00:22< loonycyborg> silene: Doesn't autogen.sh require automake and friends?
20090412 23:00:47< mordante> yes it does
20090412 23:00:56< Ivanovic> autogen.sh does require automake and autoconf
20090412 23:00:58< silene> loonycyborg: yes and scons requires scons and cmake requires cmake ;-)
20090412 23:01:41< mordante> thusfar we shipped the tarball in a state where automake and friends was not required
20090412 23:01:48< Ivanovic> jupp
20090412 23:01:50< mordante> but I see no reason not to change that
20090412 23:02:01< mordante> and 1.7.x would be a nice moment to do so
20090412 23:02:27< silene> agreed, and as far as i'm concerned, i already suggested we do it for 1.6
20090412 23:03:24< mordante> I've no problem with that as well, borderline as no changes to a stable series but we can write it in the announcement
20090412 23:03:25< CIA-30> jhinrichs * r34829 /trunk/src/ (gamestatus.cpp savegame.cpp savegame.hpp): 
20090412 23:03:25< CIA-30> Savegame reorganization Step 1: Providing a simpler interface to saving and loading.
20090412 23:03:25< CIA-30> Move gamestatus.cpp::load_game to savegame.cpp.
20090412 23:04:43< zitterbart> to make s.o. teeth is a new expression to me
20090412 23:04:59< mordante> ?
20090412 23:05:35< Ivanovic> i intentionally have not done so for the 1.6.x series since i would have prefered giving this change some testing before
20090412 23:05:44< zitterbart> english is not my mother tongue
20090412 23:05:55< boucman> zitterbart: making your teeth on something means training, basicaly
20090412 23:06:40< zitterbart> okeydokey
20090412 23:06:42< loonycyborg> Ivanovic: How did you manage to upload the tarball with all that HTTP 500 errors? :P
20090412 23:07:16< CIA-30> jhinrichs * r34830 /trunk/src/savegame.cpp: oops, fixing the messed up comments
20090412 23:07:34< Ivanovic> loonycyborg: upload is pure rsync
20090412 23:07:36< Ivanovic> worked nicely
20090412 23:07:42< Ivanovic> rsync via ssh
20090412 23:08:04< loonycyborg> Yes. But you need to add it to release.
20090412 23:08:36< loonycyborg> And what was exact command for rsync? I always look it up in sf's docs.
20090412 23:09:04< Ivanovic> rsync -avP -e ssh FILE USERNAME@frs.sourceforge.net:uploads/
20090412 23:09:13< Ivanovic> and when i added stuff to the release it was still working nicely
20090412 23:09:34< loonycyborg> thx
20090412 23:09:36-!- JW1 [n=X@c-71-57-85-168.hsd1.il.comcast.net] has quit []
20090412 23:09:38< Ivanovic> (i have a list of "stuff to do at release time" and part of this is the upload command ;) )
20090412 23:09:52-!- dapounet` [n=user@cust-48-149-111-94.dyn.versateladsl.be] has joined #wesnoth-dev
20090412 23:12:02-!- Elvish_Pillager [n=eli@66-189-40-122.dhcp.oxfr.ma.charter.com] has quit [Read error: 110 (Connection timed out)]
20090412 23:12:10-!- Turuk [n=Turuk@wesnoth/forumsith/turuk] has joined #wesnoth-dev
20090412 23:13:32< zitterbart> in poisoner_eval, does target.max_experience - target.experience <= 8,-1, mean the target is not attacked if it is near to leveling up?
20090412 23:15:12< crimson_penguin> loonycyborg: I totally memorized it :D
20090412 23:15:45< boucman> zitterbart: yes
20090412 23:16:27< zitterbart> seems a good thing :)
20090412 23:18:12< AnMaster> a question, which of the three current build systems do you (the wesnoth devs) think will become the future one. Or don't you know yet? 
20090412 23:18:51< dfranke> If we knew, it would be the present one :-)
20090412 23:18:55< AnMaster> And on a related note, how do you build out of tree with scons? As in mkdir build; cd build; <configure command>
20090412 23:19:26< AnMaster> since I used both cmake and autotools a lot previously I have no trouble making those work, but scons is confusing me.
20090412 23:19:43-!- Turuk_ [n=Turuk@wesnoth/forumsith/turuk] has quit [Read error: 110 (Connection timed out)]
20090412 23:20:23< AnMaster> dfranke, any idea about that bit?
20090412 23:20:24-!- loonybot [n=loonybot@wesnoth/bot/loonybot] has quit [Read error: 104 (Connection reset by peer)]
20090412 23:20:35-!- dapounet` [n=user@cust-48-149-111-94.dyn.versateladsl.be] has quit [Remote closed the connection]
20090412 23:20:55< mordante> AnMaster, I think we'll be using the three of them for quite a while
20090412 23:21:06< AnMaster> also it doesn't seem to pick up my CC, which is very irritating.
20090412 23:21:10< AnMaster> (scons that is)
20090412 23:21:24< mordante> so if you prefer one, use that one
20090412 23:21:25< AnMaster> nor CXX
20090412 23:22:08< AnMaster> mordante, well I wanted to try them all out and give some feedback to the developers from a user viewpoint
20090412 23:22:26< loonycyborg> AnMaster: It already builds objects out of source by default.
20090412 23:22:42< mordante> AnMaster, ok
20090412 23:22:57< AnMaster> s/user/fellow programmer/ I guess, I don't think I represent a "typical" linux user these days...
20090412 23:23:03< dfranke> AnMaster: scons cxxtool=/usr/bin/foocc
20090412 23:23:12< AnMaster> dfranke, aha
20090412 23:24:03< AnMaster> loonycyborg, so what if you want to do two builds, like I would do mkdir build build_debug and them use -O3 in one and -ggdb3 in the other?
20090412 23:24:10< CIA-30> jhinrichs * r34831 /trunk/src/savegame.cpp: minor corrections for savegame.cpp
20090412 23:24:20< dfranke> AnMaster:  scons –help from within the tree will give a list of build options we support
20090412 23:24:20< AnMaster> or such
20090412 23:24:23< AnMaster> hm
20090412 23:24:36< zitterbart> good night
20090412 23:24:39-!- zitterbart [n=beard@dslb-092-072-047-067.pools.arcor-ip.net] has quit ["Verlassend"]
20090412 23:24:52< AnMaster> dfranke, ok. just saying having several build trees against one source tree can be rather useful sometimes.
20090412 23:24:54< dfranke> AnMaster: scons build=debug and scons build=release
20090412 23:24:54< AnMaster> :)
20090412 23:25:07< dfranke> AnMaster: object files for the two builds automatically go into separate directories
20090412 23:25:36< AnMaster> dfranke, also what about CXXFLAGS, it doesn't pick that up either, resulting in build errors when it try to link 64-bit objects against 32-bit libraries.
20090412 23:25:43< dfranke> AnMaster: I routinely do both builds out of the same tree.
20090412 23:25:46< AnMaster> (yes I use a multilib system)
20090412 23:25:51< AnMaster> dfranke, aha
20090412 23:26:14< dfranke> AnMaster: extra_flags_base= or extra_flags_debug=
20090412 23:26:23< AnMaster> interesting
20090412 23:26:43< crimson_penguin> Uploading Mac dmg - hopefully there isn't anything wrong with it; didn't have time to test much, since I'm trying to go to a party
20090412 23:26:44< dfranke> and we'll automatically split them into CXXFLAGS versus LDFLAGS.
20090412 23:26:58< AnMaster> dfranke, any reason why scons doesn't pick up CC, CXX, CFLAGS, CXXFLAGS and so on like autotools and cmake does (and most other build systems I ever had to use)
20090412 23:26:58< crimson_penguin> Would've packaged it sooner, but... was distracted by a game of Wesnoth :P
20090412 23:27:26< dfranke> AnMaster: no particular reason I'm aware of.
20090412 23:27:28< crimson_penguin> ~3:30 till it's up
20090412 23:27:41-!- dapounet [n=user@cust-48-149-111-94.dyn.versateladsl.be] has quit [Connection timed out]
20090412 23:27:53< AnMaster> dfranke, maybe scons is great, but it is a lot more different from the common autotools than cmake when it comes to end user experience.
20090412 23:28:16< dfranke> AnMaster: agreed.  It's *wonderful* once you learn it though.
20090412 23:28:37< dfranke> The company I worked for hired me to maintain their autotools build system.
20090412 23:28:42< dfranke> Instead I rewrote it scons.
20090412 23:28:43< AnMaster> dfranke, maybe, but most users won't use it a lot, and thus end up irritated. :/
20090412 23:29:18< dfranke> s/worked/work/
20090412 23:29:30< loonycyborg> AnMaster: Because original developers were often getting broken builds due to environment.
20090412 23:29:36< dfranke> it in*
20090412 23:29:39< AnMaster> and I don't like autotools either. (I have maintained such a build system, was a pain)
20090412 23:29:40 * dfranke can't type today
20090412 23:29:54< AnMaster> loonycyborg, heh.
20090412 23:30:36< loonycyborg> So they decided to ignore the environment to get more repeateble builds.
20090412 23:30:36< loonycyborg> *repeatable.
20090412 23:31:06< dfranke> autotools is a collection of decades upon decades of hacks.
20090412 23:31:07< AnMaster> interesting. Why did they not clean up their environment vars instead? heh.
20090412 23:31:15< dfranke> there's nothing at the center.
20090412 23:31:20< dfranke> it's hacks all the way down.
20090412 23:31:47< loonycyborg> There were times when make wasn't used a lot either :P
20090412 23:32:07< AnMaster> dfranke, yes, personally I use cmake for all new projects I write (if they are in C/C++ that is)
20090412 23:32:39< dfranke> I heard a story about make once, don't know if it's really true...
20090412 23:33:06< dfranke> some time in the 70s, someone (I heard it as Ken Thompson) complained about how much make's syntax sucked...
20090412 23:33:21< dfranke> and the author said he agreed, but it was too late to change it because there were already DOZENS of people using it.
20090412 23:33:26< AnMaster> I rather like make, for small non-C projects it can usually be enough.
20090412 23:33:50< AnMaster> dfranke, heh
20090412 23:33:51< mordante> even for small projects I prefer cmake
20090412 23:34:11< CIA-30> esr * r34832 /trunk/data/campaigns/Delfadors_Memoirs/ (TODO maps/shadows.map scenarios/14_shadows.cfg): DM map and object tweaks.
20090412 23:34:15< AnMaster> mordante, well, cmake doesn't work well for languages like Scheme or Haskell or such.
20090412 23:34:45< mordante> I missed the non-C part ;-)
20090412 23:35:05< esr> dfranke: That story is true.  I researched it for "The Art of Unix Programming"
20090412 23:35:20< AnMaster> mordante, erlang is another one I often code in, where I use erlang's own emake, with a wrapper makefile to do the make clean
20090412 23:35:39< AnMaster> esr, hi. Nice to meet you. 
20090412 23:35:42 * loonycyborg wrote a Haskell tool for scons :P
20090412 23:35:56< AnMaster> heh
20090412 23:36:23< dfranke> I'm happy with Cabal for Haskell.
20090412 23:36:51< esr> AnMaster: Wat is the "An" you master?  (I wrote an offstage Wesnoth character named An-Usrukhat.)
20090412 23:36:53 * AnMaster is still a beginner at haskell,
20090412 23:37:02< AnMaster> esr, intitials. Why do everyone ask that :(
20090412 23:37:25< dfranke> I wrote a fairly luxurious suite of SCons builders for Erlang.  This reminds me that I need to bug my company again about letting me open source it.
20090412 23:37:36< AnMaster> esr, By the way... did you know there is a channel where you are famous for C-INTERCAL?
20090412 23:38:42< esr> I'm vaguly aware there is an esoteric-languages culture for which I'm a sort of patriarch-in-absentia.  It amused mr when I learned this.
20090412 23:39:30< AnMaster> esr, hah. I'm rather interested in these sort of languages.
20090412 23:41:12< AnMaster> esr, hm someone went on and developed C-INTERCAL, extended it, after your version. Did you know that?
20090412 23:41:50< esr> I knew there was a fork at one time, dodn't know it was still maintained.  Is it?
20090412 23:42:13< AnMaster> yes it is.
20090412 23:42:29< AnMaster> esr, http://select.intercal.org.uk/ has both the new C-INTERCAL version and CLC-INTERCAL
20090412 23:43:13< AnMaster> I think the last release was something like 8 days ago or so.
20090412 23:44:02< AnMaster> wait a bit more. 12 days :)
20090412 23:44:06< AnMaster> anyway brb
20090412 23:46:21-!- ais523 [n=ais523@92-236-187-64.cable.ubr08.king.blueyonder.co.uk] has joined #wesnoth-dev
20090412 23:51:10-!- noy [n=Noy@d75-157-52-251.bchsia.telus.net] has joined #wesnoth-dev
20090412 23:53:01-!- JW1 [n=X@c-71-57-85-168.hsd1.il.comcast.net] has joined #wesnoth-dev
20090412 23:53:15< JW1> there's a spammer on the forums.
20090412 23:53:22< JW1> he needs ban/deletion.
20090412 23:53:36< JW1> rameshkumar
20090412 23:57:09< AnMaster> JW1, that almost sound like some in-game name heh
20090412 23:58:06< JW1> I'm afraid its just someone trying to sell apartments in austin or something
20090412 23:58:07< AnMaster> (alas I'm only a user, can't help)
20090412 23:58:08< mordante> Turuk, around ^^
20090412 23:58:37< loonycyborg> Ivanovic: 1.6.1 win binary is uploaded.
20090412 23:58:55< mordante> or Ivanovic ^^
20090412 23:59:10< mordante> I can do the deletion but not banning
--- Log closed Mon Apr 13 00:00:28 2009