--- Log opened Wed Sep 02 00:00:05 2009
--- Day changed Wed Sep 02 2009
20090902 00:00:05-!- thespaceinvader [n=chatzill@wesnoth/artist/thespaceinvader] has quit ["night all"]
20090902 00:03:25-!- Crab_ [n=Crab_@wesnoth/developer/crab] has quit [Read error: 110 (Connection timed out)]
20090902 00:09:15< YogiHH> night everyone
20090902 00:09:39-!- YogiHH [n=chatzill@d083221.adsl.hansenet.de] has left #wesnoth-dev []
20090902 00:32:24< CIA-62> esr * r38293 /trunk/data/ (2 files in 2 dirs): Fix bug #13504.
20090902 00:38:24-!- Sirp [n=user@wesnoth/developer/dave] has joined #wesnoth-dev
20090902 00:40:34-!- giusef [n=giusef@unaffiliated/giusef] has quit ["exit (-1);"]
20090902 00:40:37-!- Tesafilmchen [n=micha@p5B277441.dip.t-dialin.net] has quit [Remote closed the connection]
20090902 00:47:04< CIA-62> esr * r38294 /trunk/data/core/units/drakes/Thrasher.cfg: Compatibility hack removed as anticipated. It was confusing wmllint.
20090902 00:47:21-!- grzywacz [n=grzywacz@wesnoth/developer/grzywacz] has quit [Remote closed the connection]
20090902 00:51:40-!- stikonas [n=and@wesnoth/translator/stikonas] has quit [Read error: 110 (Connection timed out)]
20090902 00:53:55-!- cib0 [n=cib@p5DD344A7.dip.t-dialin.net] has quit [Remote closed the connection]
20090902 00:54:55< CIA-62> esr * r38295 /trunk/data/core/units/drakes/Arbiter.cfg: And remove another compatibility hack.
20090902 01:27:39-!- n3penthe [i=black@adsl-85-157-188-175.regionline.fi] has quit [Read error: 110 (Connection timed out)]
20090902 01:31:55-!- loonycyborg [n=sergey@wesnoth/developer/loonycyborg] has quit ["Zzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzz"]
20090902 01:32:00-!- Chusslove [n=Chusslov@adsl-ppp-0026.yubc.net] has quit [Remote closed the connection]
20090902 01:32:29-!- loonybot [n=loonybot@wesnoth/bot/loonybot] has quit [Remote closed the connection]
20090902 01:34:37< CIA-62> esr * r38296 /trunk/data/ (8 files in 5 dirs): 
20090902 01:34:37< CIA-62> wmllint fixes (including some enhancements to wmllint itself to coope
20090902 01:34:37< CIA-62> with new AI-related tags). Minor prose polish in some Drake unit descriptions.
20090902 01:44:07-!- Espreon [n=espreon@wesnoth/developer/espreon] has quit [Read error: 110 (Connection timed out)]
20090902 01:50:15-!- fabi [n=fabi@wesnoth/developer/fendrin] has quit [Read error: 104 (Connection reset by peer)]
20090902 01:51:22-!- noy_ [n=noy@24.114.236.37] has joined #wesnoth-dev
20090902 02:02:04-!- Espreon [n=espreon@wesnoth/developer/espreon] has joined #wesnoth-dev
20090902 02:03:59-!- ilor_ [n=user@wesnoth/developer/ilor] has joined #wesnoth-dev
20090902 02:07:37-!- noy [n=Noy@wesnoth/developer/noy] has quit [Connection timed out]
20090902 02:07:37-!- noy_ is now known as noy
20090902 02:16:00-!- ilor [n=user@wesnoth/developer/ilor] has quit [Read error: 110 (Connection timed out)]
20090902 02:24:41-!- ancestral [n=ancestra@166.205.7.35] has joined #wesnoth-dev
20090902 02:24:43-!- fabi [n=fabi@wesnoth/developer/fendrin] has joined #wesnoth-dev
20090902 02:33:01-!- Zarel [i=Zarel@warzone2100/developer/Zarel] has joined #wesnoth-dev
20090902 02:33:30-!- noy [n=noy@24.114.236.37] has quit []
20090902 02:34:17-!- BenUrban [n=benurban@unaffiliated/benurban] has joined #wesnoth-dev
20090902 02:50:07-!- happygrue [n=George@wesnoth/developer/wintermute] has joined #wesnoth-dev
20090902 02:50:27< corn> Ivanovic: stats upload is active if you use --new-uploader flag
20090902 02:51:16< corn> otherwise, it uses the now inactive old uploader
20090902 02:54:20-!- ancestral [n=ancestra@166.205.7.35] has quit ["Get Colloquy for iPhone! http://mobile.colloquy.info/"]
20090902 03:10:27-!- nonsensical [n=ars@209-210-207-246.dslnorthwest.net] has joined #wesnoth-dev
20090902 03:11:57< nonsensical> Can I make a suggestion, perhaps someone here could add it to the todo list or implement it themselves.  On the load save game screen, the delete button brings up a confirm dialog, it should do this every time, and when you click the don't ask me again, it should apply for that session only, while the window is open and be reset once the window closes
20090902 03:16:33< Sirp> nonsensical: I don't really think it should apply for that session only; I find dialogs annoying.
20090902 03:17:02< nonsensical> when it comes to deleting things, if the dialog isn't there it's really easy to accidently delete a save you didn't want to delete
20090902 03:17:42< Sirp> then never click "don't show me again" ? :)
20090902 03:17:47< nonsensical> and you rarely go about deleting lots of saves at once
20090902 03:19:28< nonsensical> further, there's no way to turn that back on in the game interface :P
20090902 03:20:04< nonsensical> Another way you could do it is have the check box "confirm save deletion" in the load screen itself next to one of the other checkboxes
20090902 03:20:30< nonsensical> then you can turn it off and on at will
20090902 03:23:23< esr> fabi: ping!
20090902 03:42:35-!- DDR [n=chatzill@66.183.125.196] has joined #wesnoth-dev
20090902 03:46:43-!- fabi [n=fabi@wesnoth/developer/fendrin] has quit [Success]
20090902 03:50:35-!- Tesafilmchen [n=micha@p5B277441.dip.t-dialin.net] has joined #wesnoth-dev
20090902 03:59:02-!- Tesafilmchen [n=micha@p5B277441.dip.t-dialin.net] has quit [Remote closed the connection]
20090902 04:01:39-!- Ivanovic_ [n=ivanovic@77.178.196.127] has joined #wesnoth-dev
20090902 04:04:16-!- Blueblaze [n=nick@adsl-76-202-22-1.dsl.hstntx.sbcglobal.net] has joined #wesnoth-dev
20090902 04:10:52-!- dfranke [i=dfranke@wesnoth/developer/dfranke] has joined #wesnoth-dev
20090902 04:16:57-!- Ivanovic [n=ivanovic@wesnoth/developer/ivanovic] has quit [Read error: 113 (No route to host)]
20090902 04:17:37-!- Ivanovic_ is now known as Ivanovic
20090902 04:18:19-!- alink [n=alink@wesnoth/developer/alink] has quit [Read error: 104 (Connection reset by peer)]
20090902 04:18:20-!- alink_ [n=alink@host-85-27-114-223.brutele.be] has quit [Remote closed the connection]
20090902 04:25:29-!- Appleman1234 [n=Appleman@131.181.100.77] has joined #wesnoth-dev
20090902 04:29:51-!- dfranke [i=dfranke@wesnoth/developer/dfranke] has quit [Read error: 110 (Connection timed out)]
20090902 04:37:16-!- dfranke [i=dfranke@wesnoth/developer/dfranke] has joined #wesnoth-dev
20090902 04:40:47-!- dfranke [i=dfranke@wesnoth/developer/dfranke] has quit [Read error: 104 (Connection reset by peer)]
20090902 04:56:23-!- wesbot changed the topic of #wesnoth-dev to: 83 bugs, 241 feature requests, 11 patches | logs: http://irclogs.wesnoth.org/ | Don't paste on IRC! Use a pastebin: http://wesnoth.pastebin.com | http://imagebin.org
20090902 05:00:00-!- Ivanovic_ [n=ivanovic@77.178.32.119] has joined #wesnoth-dev
20090902 05:03:51-!- dfranke [i=dfranke@wesnoth/developer/dfranke] has joined #wesnoth-dev
20090902 05:15:49-!- Ivanovic [n=ivanovic@77.178.196.127] has quit [Read error: 110 (Connection timed out)]
20090902 05:15:58-!- Ivanovic_ is now known as Ivanovic
20090902 05:19:00-!- ancestral [n=ancestra@97-116-126-99.mpls.qwest.net] has joined #wesnoth-dev
20090902 05:24:43-!- wajimba [n=Andrew_A@24-158-30-63.dhcp.dlth.mn.charter.com] has joined #wesnoth-dev
20090902 05:25:06-!- wajimba [n=Andrew_A@24-158-30-63.dhcp.dlth.mn.charter.com] has quit [Client Quit]
20090902 05:43:46-!- fabi [n=fabi@wesnoth/developer/fendrin] has joined #wesnoth-dev
20090902 05:52:09< CIA-62> esr * r38297 /trunk/data/ (10 files in 2 dirs): 
20090902 05:52:09< CIA-62> Prose-polishing pass on LoW. Add "woodscraft" as a global spelling
20090902 05:52:09< CIA-62> exception.
20090902 05:54:06-!- dfranke [i=dfranke@wesnoth/developer/dfranke] has quit [Read error: 131 (Connection reset by peer)]
20090902 06:11:38-!- dfranke [i=dfranke@wesnoth/developer/dfranke] has joined #wesnoth-dev
20090902 06:30:45-!- ancestral [n=ancestra@97-116-126-99.mpls.qwest.net] has quit ["And that’s the end of THAT chapter."]
20090902 06:32:05-!- ancestral [n=ancestra@97-116-126-99.mpls.qwest.net] has joined #wesnoth-dev
20090902 06:35:09-!- DDR is now known as DDR_away
20090902 06:57:20< CIA-62> esr * r38298 /trunk/data/campaigns/ (21 files in 10 dirs): 
20090902 06:57:20< CIA-62> Clear all current SpellingMistakes items other than those associated with
20090902 06:57:20< CIA-62> SoF, that needs to wait until I get some sleep...
20090902 07:09:48-!- crimson_penguin [n=ben@wesnoth/developer/crimsonpenguin] has quit []
20090902 07:11:10-!- dfranke [i=dfranke@wesnoth/developer/dfranke] has quit [Read error: 104 (Connection reset by peer)]
20090902 07:29:33< CIA-62> cornmander * r38299 /website/stats.wesnoth.org/ (5 files in 3 dirs): 
20090902 07:29:33< CIA-62> Enabled logging in wsgi app. Changed log type to RotatingFileHandler and changed logging file to /var/tmp/wesstats.log. Added support for multiple
20090902 07:29:33< CIA-62> tables, although still broken and incomplete. Refactored some long string catenations into a cleaner style.
20090902 07:33:13-!- dfranke [i=dfranke@wesnoth/developer/dfranke] has joined #wesnoth-dev
20090902 07:42:51-!- fabi [n=fabi@wesnoth/developer/fendrin] has quit [Remote closed the connection]
20090902 07:43:12-!- fabi [n=fabi@wesnoth/developer/fendrin] has joined #wesnoth-dev
20090902 07:43:33-!- DDR_away [n=chatzill@66.183.125.196] has quit [Remote closed the connection]
20090902 07:43:48-!- [Relic] [n=[Relic]@99-58-54-211.lightspeed.milwwi.sbcglobal.net] has quit ["Leaving"]
20090902 07:45:29-!- Sirp [n=user@wesnoth/developer/dave] has quit [Read error: 113 (No route to host)]
20090902 07:54:25-!- dfranke [i=dfranke@wesnoth/developer/dfranke] has quit [Connection timed out]
20090902 07:56:42-!- happygrue_ [n=George@c-67-176-145-41.hsd1.in.comcast.net] has joined #wesnoth-dev
20090902 07:57:14-!- zookeeper [n=l@wesnoth/developer/zookeeper] has joined #wesnoth-dev
20090902 08:02:13-!- dfranke [i=dfranke@wesnoth/developer/dfranke] has joined #wesnoth-dev
20090902 08:03:45< Soliton> corn: any reason not to activate the new uploader by default now?
20090902 08:04:37< corn> Soliton: I'd like to test it more with sp campaigns
20090902 08:05:02< corn> give me about an hour and I can say for sure that we can switch it to default
20090902 08:06:47< corn> also, I want to fix 2 things before it's made default: 1. get rid of the mod_proxy business (I already switched it on my server, I can switch it on wesnoth.org with coordination from you or Rhonda) and 2. fix 404 errors being shown instead of internal server error
20090902 08:10:32< Soliton> ok, cool.
20090902 08:14:23-!- happygrue [n=George@wesnoth/developer/wintermute] has quit [Read error: 110 (Connection timed out)]
20090902 08:15:28-!- dfranke [i=dfranke@wesnoth/developer/dfranke] has quit [Read error: 60 (Operation timed out)]
20090902 08:18:39< corn> ok, I fixed the 404 problem
20090902 08:18:45< corn> (not commited yet)
20090902 08:19:24-!- Appleman1234 [n=Appleman@131.181.100.77] has quit [Read error: 54 (Connection reset by peer)]
20090902 08:20:28< CIA-62> cornmander * r38300 /website/stats.wesnoth.org/ (4 files in 2 dirs): 
20090902 08:20:28< CIA-62> Debugged remaining problems with multiple table support. Fixed duplicate lines in logfile. Fixed 500 Internal Server Error being shown as 404 Not
20090902 08:20:28< CIA-62> Found.
20090902 08:21:49-!- Chusslove [n=Chusslov@212.124.170.26] has joined #wesnoth-dev
20090902 08:30:25-!- stikonas [n=and@ctv-213-164-125-176.vinita.lt] has joined #wesnoth-dev
20090902 08:31:07< CIA-62> cornmander * r38301 /website/stats.wesnoth.org/wesstats/ (controllers/root.py templates/upload.html): Added browser-viewable info to the uploader page. Made the view add/delete pages inaccessible since they're not done.
20090902 08:31:26< corn> http://cornmander.com/wesstats/upload
20090902 08:31:30< corn> no longer a 404 :)
20090902 08:35:36-!- dfranke [i=dfranke@wesnoth/developer/dfranke] has joined #wesnoth-dev
20090902 08:37:59< corn> going to bed now - tomorrow is definitely the day that I'll make the uploader switch
20090902 08:40:35-!- Appleman1234 [n=Appleman@131.181.100.77] has joined #wesnoth-dev
20090902 08:45:11-!- euschn [n=chatzill@wesnoth/developer/euschn] has joined #wesnoth-dev
20090902 08:45:42-!- Crab_ [n=Crab_@wesnoth/developer/crab] has joined #wesnoth-dev
20090902 08:59:20< Espreon> Chusslove: http://imagebin.org/62091 — Do you think that is of decent quality?
20090902 09:00:20-!- noy [n=Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20090902 09:00:36< ancestral> corn: your 404 looks like mine: mproud.com
20090902 09:00:42< ancestral> http://mproud.com
20090902 09:00:52< ancestral> Well mine's not technically a 404
20090902 09:02:28< Chusslove> Espreon: I've got to go away in a few moments, so just shortly...
20090902 09:02:32< ancestral> I kinda like this one: http://www.flickr.com/photos/apelad/1266913887/
20090902 09:02:40< Espreon> Chusslove: Very well...
20090902 09:03:09< Chusslove> Espreon: The words batalla, por, and Wesnoth are not on same horizontal baseline -- is that intentional?
20090902 09:03:58< Espreon> Are you referring to the fact that they are not perfectly aligned?
20090902 09:04:10< Espreon> If so, no.
20090902 09:04:26< Chusslove> Well, that ought to be fixed.
20090902 09:04:32< ancestral> Espreon: why not type them in one text box?
20090902 09:05:05< Espreon> Because when I tried that, it screwed up the spacing between the "p" and the "o" in "por."
20090902 09:05:34< Espreon> It would kinda smoosh them together like "œ"
20090902 09:06:20< Chusslove> Espreon: The other thing is that "La" (with initial capital) is up, and "batalla" (initial lower case) is in second row. Is this usual for titles in Spanish? Because it kind of leaves strange impression for me.
20090902 09:06:32< Espreon> Yes.
20090902 09:07:09< Chusslove> What I want to say is: could it be that "La <something>" must be kept together in titles? Or not?
20090902 09:07:11< Espreon> I am not 100% sure, but, from what I read and from what I saw in the Spanish translation, this is the case.
20090902 09:07:29< Espreon> Huh?
20090902 09:07:33< Chusslove> No, no... I'm refereing to the fact that they are split accross lines here.
20090902 09:07:53< Espreon> Well... I was following the way the English logo does it.
20090902 09:08:19< Chusslove> Espreon: Follow the way Spanish language does it :)
20090902 09:08:58< Chusslove> I mean, if there is never a title, logo, etc. which breaks line between "La" and the noun, then this logo shouldn't either.
20090902 09:09:15< Espreon> I'd have to research that.
20090902 09:09:51< Espreon> But, other than style and the "por" out of alignment, is everything else all right?
20090902 09:09:57< Chusslove> Ok, that's it from me (the allignment and breaking after "La"), so ...
20090902 09:10:05< Chusslove> Right, I see no other problem.
20090902 09:10:58< Espreon> Hmmm... I'd have no idea how to research the title style. I'd have to wait for shadowmaster to return.
20090902 09:11:30< Chusslove> There's no hurry :)
20090902 09:11:30< Espreon> Spanish isn't my native language, I am just learning it.
20090902 09:11:38< Chusslove> Ok, I'm off.
20090902 09:11:59< Espreon> Very well.
20090902 09:12:00< Espreon> Thanks.
20090902 09:12:19-!- Chusslove [n=Chusslov@212.124.170.26] has quit [Remote closed the connection]
20090902 09:18:01-!- YogiHH [i=d4ca9d15@wesnoth/developer/yogihh] has joined #wesnoth-dev
20090902 09:18:05< YogiHH> hello
20090902 09:19:19< Espreon> Hello YogiHH.
20090902 09:22:17-!- Appleman1234 [n=Appleman@131.181.100.77] has quit [Read error: 110 (Connection timed out)]
20090902 09:22:44-!- ancestral [n=ancestra@97-116-126-99.mpls.qwest.net] has quit ["And that’s the end of THAT chapter."]
20090902 09:23:55-!- Espreon [n=espreon@wesnoth/developer/espreon] has quit ["WRYYYYYYYYYYYYYYYYYYYY!"]
20090902 09:35:06-!- stikonas [n=and@wesnoth/translator/stikonas] has quit [Remote closed the connection]
20090902 09:44:06-!- dfranke [i=dfranke@wesnoth/developer/dfranke] has quit [Connection timed out]
20090902 09:51:36-!- dfranke [i=dfranke@wesnoth/developer/dfranke] has joined #wesnoth-dev
20090902 10:08:08< CIA-62> thespaceinvader * r38302 /trunk/data/core/units/drakes/ (Hurricane.cfg Sky.cfg): Minor fix to Sky and Hurricane drake animations.
20090902 10:09:34< Ivanovic> corn: since the old status uploader is gone it *might* make sense to switch on the new one instead of the old one
20090902 10:09:36< Ivanovic> ;)
20090902 10:11:44< Ivanovic> okay, saw the followup
20090902 10:12:00< Ivanovic> good to see that the new statusuploader will be included in 1.7.5
20090902 10:19:18-!- dfranke [i=dfranke@wesnoth/developer/dfranke] has quit [Read error: 104 (Connection reset by peer)]
20090902 10:22:36< CIA-62> soliton * r38303 /branches/1.6/src/server/ban.cpp: take both masks into account so we match wider bans when looking for a too narrow one
20090902 10:22:42< CIA-62> soliton * r38304 /branches/1.6/src/server/ban.cpp: don't list ban groups when looking for specific bans
20090902 10:24:28< CIA-62> soliton * r38305 /trunk/src/server/ban.cpp: take both masks into account so we match wider bans when looking for a too narrow one
20090902 10:24:33< CIA-62> soliton * r38306 /trunk/src/server/ban.cpp: don't list ban groups when looking for specific bans
20090902 10:26:00< Ivanovic> moin
20090902 10:28:50< Ivanovic> uhm, was r38302 really intentional?
20090902 10:28:52< Ivanovic> -    [portrait]
20090902 10:28:53< Ivanovic> +    portrait]
20090902 10:35:06< YogiHH> zookeeper: Is it allowed to use a "-" (minus) sign for WML variable names?
20090902 10:35:24< zookeeper> YogiHH, yes
20090902 10:37:19< Ivanovic> oaky, it was *clearly* not intentional
20090902 10:38:33< YogiHH> zookeeper: http://www.wesnoth.org/forum/viewtopic.php?f=21&t=26559&start=15
20090902 10:38:53-!- dfranke [i=dfranke@wesnoth/developer/dfranke] has joined #wesnoth-dev
20090902 10:39:08< YogiHH> zookeeper: looks fine to me, maybe it is a problem of execution order, depending in which events that snippets are placed?
20090902 10:39:23< CIA-62> ivanovic * r38307 /trunk/data/core/units/drakes/Hurricane.cfg: 
20090902 10:39:23< CIA-62> fix an accidentially commited file/change from r38302
20090902 10:39:23< CIA-62> TSI: please do check svn diff before commiting
20090902 10:40:35< zookeeper> YogiHH, i kinda skipped that since i figured it's probably something to do with wrong sides or missing recall lists since he's doing complex stuff.
20090902 10:41:06< YogiHH> ah, ok
20090902 10:42:18< zookeeper> if the sides are correct and the later side actually has a recall list, then i guess it should work just fine, but since i have no idea whether he just messed something up..
20090902 10:45:00< fabi> esr: pong
20090902 10:47:55-!- fabi [n=fabi@wesnoth/developer/fendrin] has quit [Remote closed the connection]
20090902 10:51:06-!- Zarel [i=Zarel@warzone2100/developer/Zarel] has quit [Read error: 110 (Connection timed out)]
20090902 10:52:27-!- euschn [n=chatzill@wesnoth/developer/euschn] has quit [Read error: 60 (Operation timed out)]
20090902 10:56:49-!- fendrin [n=fabi@wesnoth/developer/fendrin] has joined #wesnoth-dev
20090902 10:59:16-!- dfranke [i=dfranke@wesnoth/developer/dfranke] has quit [Read error: 110 (Connection timed out)]
20090902 11:04:16-!- thespaceinvader [n=chatzill@wesnoth/artist/thespaceinvader] has joined #wesnoth-dev
20090902 11:05:54-!- dfranke [i=dfranke@wesnoth/developer/dfranke] has joined #wesnoth-dev
20090902 11:21:13-!- Blueblaze [n=nick@adsl-76-202-22-1.dsl.hstntx.sbcglobal.net] has quit [Remote closed the connection]
20090902 11:30:23-!- euschn [n=chatzill@wesnoth/developer/euschn] has joined #wesnoth-dev
20090902 11:46:21< euschn> Ivanovic: I have shortened my gsoc summary and sent it to you via pm. is this more like what you had in mind?
20090902 11:46:48-!- noy [n=Noy@wesnoth/developer/noy] has quit []
20090902 11:47:25< Ivanovic> euschn: looks perfect
20090902 11:47:34< euschn> great
20090902 11:47:36< Ivanovic> now i would just like to know what happened to DK before posting it
20090902 11:51:12-!- loonybot [n=loonybot@79.139.139.133] has joined #wesnoth-dev
20090902 11:52:00-!- loonycyborg [n=sergey@79.139.139.133] has joined #wesnoth-dev
20090902 12:51:53-!- AnMaster [n=AnMaster@unaffiliated/anmaster] has quit [Connection timed out]
20090902 12:58:24-!- Espreon [n=espreon@wesnoth/developer/espreon] has joined #wesnoth-dev
20090902 12:58:58-!- dfranke [i=dfranke@wesnoth/developer/dfranke] has quit [Read error: 60 (Operation timed out)]
20090902 13:22:13-!- dfranke [i=dfranke@wesnoth/developer/dfranke] has joined #wesnoth-dev
20090902 13:30:52< CIA-62> fendrin * r38308 /trunk/data/campaigns/Delfadors_Memoirs/scenarios/08_ur_thorodor.cfg: DM 08: Replaced a depricated result=continue.
20090902 13:31:39< CIA-62> fendrin * r38309 /trunk/data/campaigns/Delfadors_Memoirs/scenarios/09_houses_of_the_dead.cfg: LoW 09: Made use of translated team name for the player.
20090902 13:53:41-!- Appleman1234 [n=Appleman@131.181.100.77] has joined #wesnoth-dev
20090902 13:57:51-!- lizard_r [n=Miranda@wesnoth/umc-dev/developer/lizard] has joined #wesnoth-dev
20090902 14:01:23-!- YogiHH [i=d4ca9d15@wesnoth/developer/yogihh] has quit [Ping timeout: 180 seconds]
20090902 14:04:14-!- YogiHH [i=d4ca9d15@wesnoth/developer/yogihh] has joined #wesnoth-dev
20090902 14:15:00-!- Chusslove [n=Chusslov@212.124.170.166] has joined #wesnoth-dev
20090902 14:17:16-!- Appleman1234 [n=Appleman@131.181.100.77] has quit [Read error: 145 (Connection timed out)]
20090902 14:17:41< Ivanovic> thespaceinvader: you saw my commit fixing yours?
20090902 14:18:00< thespaceinvader> Ivanovic: which one?
20090902 14:18:02< Ivanovic> thespaceinvader: that is: in general r38302 looks really suspicous to me regarding the flying animation stuff
20090902 14:18:20< thespaceinvader> wesbot log 38302
20090902 14:18:21< wesbot> thespaceinvader * r38302 : Minor fix to Sky and Hurricane drake animations.
20090902 14:18:24< wesbot> URL: http://svn.gna.org/viewcvs/wesnoth?view=rev&rev=38302
20090902 14:18:35< Ivanovic> if you are not subscribed there yet, it might make sense to subscribe to wesnoth-commits (the mainling list that gets all the commits)
20090902 14:18:51< thespaceinvader> i'm not sure whether I am or not
20090902 14:18:51< Ivanovic> in r38307 i made the game at least start again
20090902 14:19:06< corn> is there any way I can find out in which revision a particular line of sourcecode was added
20090902 14:19:17< thespaceinvader> but to be honest, the signal-to-noise ratio on it would be huge for me, since 90% of the commits are irrelevant
20090902 14:19:23< Soliton> corn: svn blame
20090902 14:19:26< Ivanovic> if you get a mail for every revision that starts with [Wesnoth-commits] in the topic you are subscribed
20090902 14:19:36< thespaceinvader> wait, what i did broke the game that much?
20090902 14:19:48< Ivanovic> jupp
20090902 14:19:54< thespaceinvader> all i did was remove a filtered standing animation to match all the other current drakes
20090902 14:19:59< Ivanovic> in the hurricane config you removed an opening [
20090902 14:20:15< Ivanovic> that is the hurricane still has this line:
20090902 14:20:17< Ivanovic> {DRAKE_FLYING_ANIM "units/drakes/hurricane.png" "units/drakes/hurricane-fly-1.png"}
20090902 14:20:30< corn> Soliton: cool, I didn't know this command worked that way
20090902 14:20:38< Ivanovic> from the sky you removed the respective one in that commit
20090902 14:20:41< Ivanovic> {DRAKE_FLYING_ANIM "units/drakes/sky.png" "units/drakes/sky-fly-1.png"}
20090902 14:20:57< thespaceinvader> ok
20090902 14:21:03< Ivanovic> please check your changes before commiting
20090902 14:21:06< corn> wesbot log 37610
20090902 14:21:07< wesbot> cornmander * r37610 : Enabled log uploads for ai vs. ai multiplayer matches (logs are still incomplete, however). refactored read_replay(). Wrote new upload_log.start() without deprecated parameters.
20090902 14:21:11< wesbot> URL: http://svn.gna.org/viewcvs/wesnoth?view=rev&rev=37610
20090902 14:21:17< Ivanovic> no idea how to do so with the client you use, i would just run "svn diff"
20090902 14:21:25< corn> I wonder why I added this...
20090902 14:22:12< thespaceinvader> sure, will do
20090902 14:23:29< CIA-62> fendrin * r38310 /trunk/data/campaigns/Delfadors_Memoirs/scenarios/ (8 files): 
20090902 14:23:29< CIA-62> DM: Added translatable team names.
20090902 14:23:29< CIA-62> Replaced die events with last breath ones when the dying unit still has a last word.
20090902 14:27:25-!- dfranke [i=dfranke@wesnoth/developer/dfranke] has quit [Connection timed out]
20090902 14:34:25< fendrin> zookeeper: do you maintain AOI?
20090902 14:35:00-!- dfranke [i=dfranke@wesnoth/developer/dfranke] has joined #wesnoth-dev
20090902 14:35:04< zookeeper> fendrin, no, that's esr's job IIRC
20090902 14:35:25< fendrin> and the sceptre of fire?
20090902 14:35:25 * zookeeper thinks all campaigns should have an explicit "campaign maintenance" field in the credits..
20090902 14:35:31< zookeeper> SoF is mine, yes
20090902 14:35:48< fendrin> There is a foe named landar in sof
20090902 14:36:00< zookeeper> hmh?
20090902 14:36:12< fendrin> scenario 8 the dragon
20090902 14:36:24< zookeeper> ah, right
20090902 14:36:34< YogiHH> thespaceinvader: You use TortoiseSVN?
20090902 14:36:50< fendrin> and it looks from the speech in other scenarios that it is the same as LoW's Landar
20090902 14:37:10< zookeeper> fendrin, is it a continuity problem?
20090902 14:37:13< fendrin> Does this fit in the time?
20090902 14:37:18< thespaceinvader> YogiHH: yes
20090902 14:37:40< zookeeper> the elves making trouble in SoF are supposed to be landar's elves, at least according to SoF
20090902 14:37:50< YogiHH> thespaceinvader: Then in the commit window, simply doubleclick on every file that is checked
20090902 14:37:53< zookeeper> i don't know about the time
20090902 14:38:19< thespaceinvader> ok YogiHH
20090902 14:38:25< fendrin> I am wondering if landar is still alive. Maybe some of his followers are still fighting somewhere that shouldn't be a problem.
20090902 14:38:46< YogiHH> thespaceinvader: unless it is binary information like png files, of course
20090902 14:38:55< fendrin> But the Landar in SoF can be killed by the player. And he shouldn't be an elvish marshall. 
20090902 14:39:44< thespaceinvader> of course
20090902 14:39:45< Crab_> fendrin: from http://www.wesnoth.org/wiki/WesnothHistory it seems that the timing is ok.
20090902 14:39:48< AI0867> according to the wiki, SoF starts in 25 YW
20090902 14:40:03< AI0867> Landar starts the civil war in 26 YW
20090902 14:40:21< AI0867> wait, the events leading up to SoF are in 25 YW
20090902 14:40:35< AI0867> so the timeframe fits, but landar shouldn't be in that cave
20090902 14:40:41< zookeeper> fendrin, well, his type can be changed, and the player shouldn't be able to kill him. i can make him invincible.
20090902 14:40:42-!- yamokidzu-it [n=rainy-it@host186-67-dynamic.21-79-r.retail.telecomitalia.it] has joined #wesnoth-dev
20090902 14:41:13< fendrin> And what wonders me most is: Landar is an dwarven hater. He kills some dwarves in LoW. Now he is fighting with dwarves against other dwarves.
20090902 14:41:50< zookeeper> well, the dwarves he fights with are some sort of mercenaries
20090902 14:41:50< AI0867> fendrin: what better way to get rid of dwarves than to hire one half to kill the other half?
20090902 14:42:04< zookeeper> probably considered to be scum by other dwarves :p
20090902 14:42:16< fendrin> zookeeper: yes, but landar is a nazi.
20090902 14:42:27< fendrin> ho wouldn't fight with them side on side.
20090902 14:42:33< fendrin> at that time.
20090902 14:42:38< zookeeper> ah, hmm.
20090902 14:42:53-!- euschn [n=chatzill@wesnoth/developer/euschn] has quit [Read error: 110 (Connection timed out)]
20090902 14:43:30< fendrin> maybe just change his name. all that elves can be somehow in leage with landar. but he shouldn't take action in persion. everything else can be left as it is.
20090902 14:44:54< fendrin> AI0867: Right, the timeframe seems to fit.
20090902 14:45:34< fendrin> since there is no speech for him in that scenario anyway it shouldn't be a problem.
20090902 14:46:00< zookeeper> fendrin, would there be a suitable named elf in LoW that could be re-used?
20090902 14:46:23< fendrin> zookeeper: someone like landar's right hand?
20090902 14:46:26< zookeeper> yeah
20090902 14:46:35< fendrin> Cintril
20090902 14:47:21< zookeeper> got it. is elvish marshal ok for him (it's a he, right?)?
20090902 14:49:00< fendrin> "Crintil" is a elven captain in low. But I can make him a marshal, no problem.
20090902 14:50:42< zookeeper> well, is that before or after his appearance in SoF, timeline-wise?
20090902 14:50:47< fendrin> he shouldn't be killable as well. He has to live on to the mid of the elvish civil war.
20090902 14:52:08< fendrin> he is killed in low 22. That should be long after the sof scenario 1.
20090902 14:53:06< zookeeper> well, we're talking of scenario 8
20090902 14:53:14< zookeeper> which is a couple of years after scenario 1
20090902 14:54:05-!- lizard_r [n=Miranda@wesnoth/umc-dev/developer/lizard] has quit ["Saurian Augur - I'll heal you by 4 hp if you leave next to me"]
20090902 14:54:16< fendrin> there is something messed up. kalenz and his gang tell haldric2 in 23YW that he must put the ruby of fire in gold.
20090902 14:54:46< fendrin> in sof scenario 1 is 10 years after that.
20090902 14:54:58< AI0867> fendrin: he has to dig up the ruby first
20090902 14:55:06< AI0867> he sends an expedition to do that in 26 YW
20090902 14:55:27< fendrin> dig up?
20090902 14:55:36< Crab_> AI0867: note - wiki says "In 25 YW Haldric II sends an expedition to retrieve the Ruby of Fire from its place of concealment. "
20090902 14:55:40< AI0867> it is implied that it was hidden after his father's death
20090902 14:55:55< AI0867> -1 year
20090902 14:57:34< Soliton> it is hidden in trow while haldric lives.
20090902 14:58:16< fendrin> Okay, the timeline seems to fit. I will make Crintil to an elvish Marshall in LoW. 
20090902 14:58:28< zookeeper> haha, it stays in the troll hole? :P i never thought they'd _not_ go pick it up from there
20090902 15:00:27< fendrin> zookeeper: Would you mind some speech that connects the two campaigns somehow? 
20090902 15:00:51< corn> Crab_: there was a regression in the ai vs ai upload code that broke it... do you think it was any changes you made
20090902 15:01:42< corn> it is almost definitely a change that I made
20090902 15:01:59< Crab_> corn: what exactly happens ?
20090902 15:02:23< corn> Crab_: the internal 'config_' variable of the upload_log is never written to
20090902 15:02:31< corn> or it is written to, but then emptied
20090902 15:02:42< corn> so the upload log is blank at upload-time
20090902 15:02:49< zookeeper> fendrin, sure, that'd be good
20090902 15:03:16< Crab_> corn: 'getting the faulty commit by binary search' is, probably, the way to go
20090902 15:03:22< zookeeper> especially in scenarios 6 and 8. landar is only mentioned in scenario 1 atm.
20090902 15:03:50< corn> Crab_: does svn have a feature to automate that? I remember reading that it's a git feature
20090902 15:04:07< corn> (fetch revisions in a binary search way and present them to the user)
20090902 15:04:42< Crab_> no, it's a git feature. of course, there are scripts for svn, too.
20090902 15:04:52< fendrin> zookeeper: I am not exactly sure about the motives of the elves in that campaign. Every elvish foe in this campaign belongs to Landar's rebells?
20090902 15:04:57< corn> Crab_: ok, thanks
20090902 15:05:04< zookeeper> fendrin, yeah
20090902 15:05:50< fendrin> and why do they interfear?
20090902 15:06:08< zookeeper> i don't remember
20090902 15:06:34< Crab_> fendrin: in 1st scenario of SoF, elves just (pretend that) they want gold :)
20090902 15:06:45< fendrin> Crab_: right
20090902 15:07:15< zookeeper> seems like they just try to sabotage haldric's dealings
20090902 15:07:22< fendrin> But later on they should have more intense motives to keep in the game.
20090902 15:07:26< Crab_> zookeeper: kalenz's dealings
20090902 15:07:34-!- YogiHH [i=d4ca9d15@wesnoth/developer/yogihh] has quit [Ping timeout: 180 seconds]
20090902 15:07:42< zookeeper> Crab_, kalenz doesn't have any dealings there, it's just haldric and the dwarves
20090902 15:08:21< Crab_> zookeeper: kalenz was instrumental in the new treaty between Haldric 2 and elves. so, sabotaging this treaty is a very important.
20090902 15:08:34< Crab_> 'I was never among those who agreed to that treaty. Not all of us Elves are as low as Kalenz, or make deals with those who betrayed us'
20090902 15:08:55< zookeeper> Crab_, well, haldric already remarks that they're the "rebel elves", so how could they sabotage the treaty by just being pests?
20090902 15:08:56< fendrin> landar was present at that council. so he knows about the ruby of fire and it's problem.
20090902 15:09:55< fendrin> The elves in sof scenario1 are only after the gold. But they report the ongoings to landar who knows about the ruby.
20090902 15:10:18< fendrin> All elves that come after the player after that should have some special task from landar.
20090902 15:10:59< zookeeper> well, what's landar's motive? get the ruby or just prevent haldric from getting it back shielded?
20090902 15:11:14< corn> Crab_: found the bug, it was something I had introduced in my working copy
20090902 15:12:09< Crab_> zookeeper: rebels want support from the 'rest of the elves' in their struggle. So, after a fight with Haldric, they can appeal to blood and persuade other elves to join them in their fight vs kalenz (i.e. "kalenz conspires and signes treaties with those who have killed our brothers and sisters, therefore kalenz is evil. join us to defeat him")
20090902 15:12:11< fendrin> zookeeper: both outcomings would fit his purposes.
20090902 15:14:11< Crab_> zookeeper: i.e. if the elven rebels have wanted the gold, they could just had attacked the caravan in transit, without warnings or demands.
20090902 15:14:11< zookeeper> all right
20090902 15:15:45-!- happygrue [n=George@wesnoth/developer/wintermute] has joined #wesnoth-dev
20090902 15:16:19< zookeeper> maybe the road is actually elvish territory as the rebels IIRC claim, but haldric has the right of passage due to kalenz's treat
20090902 15:16:26< zookeeper> err, treaty :P
20090902 15:16:57< zookeeper> so, obviously the rebels feel justified in attacking them since they don't recognize the treaty
20090902 15:17:51< Crab_> zookeeper: the rebels fell justified in attacking them because they're dwarvish scum and human scum, and they must die in the name of Landar.
20090902 15:18:21< Crab_> zookeeper: 'territory issue' seems like a pretext to make Haldric the 'aggressor' side.
20090902 15:18:57< zookeeper> sure, pretext is fine, as long as it can be rationalized in some way
20090902 15:19:10< Crab_> yes..
20090902 15:19:32< fendrin> Crab_: But zookeeper's plan fits to your ideas. The elves want a reason to declare the human enemies that holds for all elves not just the nazi ones.
20090902 15:19:40< zookeeper> better to have the fanatical baddie have some kind of a proper excuse for attacking instead of just attacking for the heck of it
20090902 15:19:48< Crab_> fendrin: exactly.
20090902 15:20:32< Crab_> zookeeper: well, 'attacking for the heck of it' is better from a tactical point of view - since it can be done without warnings :)
20090902 15:20:57< Crab_> zookeeper: so, they had a reason for actually stating that proper excuse :)
20090902 15:21:15< zookeeper> yes, something like that ;)
20090902 15:22:21< fendrin> Some of that history can be flow into LoW as well.
20090902 15:22:32< esr> fendrin: I did the orose polish om LoW last night, ut there were three "ESR" tags I left imn place because I couldn't tell what you wanted me to do there.
20090902 15:22:55< fendrin> esr: okay, I will have a look now.
20090902 15:23:01-!- dfranke [i=dfranke@wesnoth/developer/dfranke] has quit [Connection timed out]
20090902 15:23:39< fendrin> esr: All in 02 hostile mountains?
20090902 15:23:43-!- happygrue__ [n=George@c-67-176-145-41.hsd1.in.comcast.net] has joined #wesnoth-dev
20090902 15:23:53< esr> Yes.
20090902 15:24:31< fendrin> The first question is: Is "Enemies" a proper name for a team?
20090902 15:25:06< esr> Don't see why not.
20090902 15:25:14-!- BenUrban [n=benurban@unaffiliated/benurban] has quit ["Power failu"]
20090902 15:26:01< fendrin> I have read all the diffs of the prosa change you did, they fit well. Thank you
20090902 15:27:39< fendrin> esr, zookeeper, shadowmaster, all campaign maintainers: I have noticed that in many campaigns the "die" event is used instead of the "last breath" one. That leads to empty hexes speaking. Maybe that should be fixed before 1.8.
20090902 15:29:27< fendrin> zookeeper: If you allow me I am doing some map updates like exchanging old with new terrain to sof. 
20090902 15:29:39< esr> I could write a script to find all files in which "event=die" is used but event=last breath" is not.  That would clue us into scenarios written before "last breath" was defined.
20090902 15:30:18< fendrin> esr: You could check for die events that use the died person as a speaker. But this task is a little bit harder.
20090902 15:30:44< esr> Hm.
20090902 15:31:13-!- dfranke [i=dfranke@wesnoth/developer/dfranke] has joined #wesnoth-dev
20090902 15:31:59< CIA-62> cornmander * r38311 /trunk/src/ (playsingle_controller.cpp upload_log.cpp): Fixed bug introduced in r37610: upload log always contains a [quit] child, even during victory and draw.
20090902 15:32:09< CIA-62> cornmander * r38312 /website/stats.wesnoth.org/wesstats/controllers/root.py: Fixed logic error that caused all singleplayer upload_logs to be marked as ai vs ai multiplayer logs.
20090902 15:32:31< Crab_> esr: thanks for LoW prosa fixes. actually, there's one string that doesn't look 'ok' in all cases. In LoW 3,  orc slayer returns to orc warlord. If slayer returns before fog is lifted, they exchange dialogue about the battlefield situation (because it's fog, and warlord hasn't seen it all himself). 
20090902 15:32:32< Crab_> Then, orc warlord orders the slayer to "Go, report this news to the warlord Grubr.".
20090902 15:33:00< Crab_> but, if orc slayer returns after the fog is lifted, the line "Go, report this news to the warlord Grubr." is used alone. so, in this case, it's not very clear what 'this news' is.
20090902 15:33:20-!- happygrue_ [n=George@c-67-176-145-41.hsd1.in.comcast.net] has quit [Read error: 110 (Connection timed out)]
20090902 15:33:21< Crab_> Esr: maybe change 'this news' to 'the news' ?
20090902 15:33:46< esr> fendrin: Your should fix that logic.  I'll look into detecting anomalous die events.
20090902 15:34:15< fendrin> esr: What logic? 
20090902 15:35:01< esr> Look at what Crab just said.  You probably want to make that bit of dialog conditional on whether there is actual news.
20090902 15:36:43< fendrin> Crab_: Do we need additional speech if the fog has lifted before the slayer returns?
20090902 15:36:52< Crab_> fendrin: if orc slayer returns after the fog is lifted, then we only need to explain his disappearance.
20090902 15:37:29< Crab_> fendrin: this is explained by  "Go, report this news to the warlord Grubr.", but what 'this news' refer to, is not clear.
20090902 15:38:07< Crab_> fendrin: ( I used 'the situation' instead of 'this news' before, which was 'good enough' for both cases)
20090902 15:38:38< fendrin> I see.
20090902 15:39:33-!- happygrue_ [n=George@c-67-176-145-41.hsd1.in.comcast.net] has joined #wesnoth-dev
20090902 15:40:10< fendrin> Crab_: Let's tell the situation from the mouth of the orc leader. Like he is telling his scout what to say.
20090902 15:41:06-!- happygrue [n=George@wesnoth/developer/wintermute] has quit [Read error: 110 (Connection timed out)]
20090902 15:41:08< Crab_> fendrin: it's somewhat hard, since to do this correctly after the fog lifts, requires analysis of Kalenz's force
20090902 15:42:24< fendrin> Crab_: Then the news that need to be reported is the arrival of Kalenz?
20090902 15:43:01< Crab_> fendrin: the turn when the slayer reaches his boss is somewhat undetermined.
20090902 15:43:10< Crab_> fendrin: since there are 5-9 turns for fog
20090902 15:43:30< Crab_> fendrin: and the elves might delay the slayer, too
20090902 15:44:00< Crab_> fendrin: and, the orc strategy is already decided and communicated to the player at this point.
20090902 15:44:14< fendrin> Crab_: okay
20090902 15:44:23< fendrin> Crab_: what is the problem with all of that?
20090902 15:45:11-!- AnMaster [n=AnMaster@unaffiliated/anmaster] has joined #wesnoth-dev
20090902 15:46:08< Crab_> fendrin: the problem is that, when slayer reaches his boss after the fog lifts, the prosa is somewhat strange: W: -"Go, report this news to the warlord Grubr." S: - "I obey."  
20090902 15:46:57< Crab_> fendrin: so, I suggest changing 'this news' or just omitting them. i.e.  W: -"Go, report to the warlord Grubr." S: - "I obey."  
20090902 15:47:52< fendrin> Crab_: If he reaches his boss after the fog lifts he tells: "An unknown elvish force has arrived. Go tell this news..."
20090902 15:48:12< Crab_> fendrin: ok
20090902 15:48:17< Crab_> fendrin: I'll change this.
20090902 15:48:56< fendrin> Crab_: We will never be ready for string frezze :-)
20090902 15:49:20< Crab_> fendrin: well, things are a lot better than yesterday :)
20090902 15:52:07-!- dfranke [i=dfranke@wesnoth/developer/dfranke] has quit [Read error: 104 (Connection reset by peer)]
20090902 15:53:42< fendrin> Crab_: That is true. The new scenario 3 is a fresh breeze.
20090902 15:55:49< CIA-62> fendrin * r38313 /trunk/data/campaigns/Delfadors_Memoirs/utils/deaths.cfg: DM: changed some die to last breath events.
20090902 15:56:23-!- happygrue__ [n=George@c-67-176-145-41.hsd1.in.comcast.net] has quit [Read error: 110 (Connection timed out)]
20090902 16:01:54< CIA-62> fendrin * r38314 /trunk/data/campaigns/Legend_of_Wesmere/scenarios/03_Kalian.cfg: LoW 3: Different music playlists for each Phase of the scenario.
20090902 16:02:56-!- Nobun [i=5704f17b@gateway/web/freenode/x-jrpsxrgroedgrsrh] has joined #wesnoth-dev
20090902 16:03:04< zookeeper> fendrin, sure, go ahead. is it just very minor changes or something bigger?
20090902 16:03:16< CIA-62> fendrin * r38315 /trunk/data/campaigns/Legend_of_Wesmere/scenarios/22_Northern_Battle.cfg: LoW 22: Promoted Crintil to Elvish Marshall.
20090902 16:03:28< fendrin> zookeeper: only terrain changes that behave the same.
20090902 16:04:24< zookeeper> fendrin, ok...for example what?
20090902 16:05:25< fendrin> zookeeper: in the first scenario I will transform the elves keep in an elvish keep. The road which is made of dirt will become elvish road.
20090902 16:05:58< zookeeper> fendrin, ah, ok, sounds good
20090902 16:09:44< fendrin> zookeeper: maintains will be transformed to snow mountains if sourrounded by snow hills. But I won't change the west wall because this could destroy the balance of the scenario.
20090902 16:14:53-!- dfranke [i=dfranke@wesnoth/developer/dfranke] has joined #wesnoth-dev
20090902 16:14:58-!- YogiHH [i=d4ca9d15@wesnoth/developer/yogihh] has joined #wesnoth-dev
20090902 16:16:34< Crab_> fendrin: I tried playing up with ai parameters to make it 'hold position' better, but there are reasons why it doesn't work well atm. 
20090902 16:16:54< Crab_> one of the solutions I'm thinking about is to add a parameter to default ai recruitment to allow us to choose whether the ai will prefer quality (few high-level units) vs quantity (many low-level units) - this will allow the ai to take more units, and last longer. Anything else that we can parametrize in recruitment ?
20090902 16:20:50< esr> Ivanovic: SpellingMistakes are clreared.  As Lord of Translationsm, you should be aware of the new page NotSpellingMistakes.
20090902 16:21:41-!- Nobun [i=5704f17b@gateway/web/freenode/x-jrpsxrgroedgrsrh] has quit ["Page closed"]
20090902 16:22:55-!- Baufo [n=thomas@wesnoth/developer/baufo] has joined #wesnoth-dev
20090902 16:23:06 * Soliton . o O (FrequentlyProposedSpellingMistakes?)
20090902 16:27:23< esr> Soliton: The most frequently proposed one is "marsh" for "march". I've gotten really tired of having to explain that one.
20090902 16:28:54< CIA-62> esr * r38316 /trunk/data/campaigns/ (11 files in 2 dirs): Clear up spelling mistakes.
20090902 16:29:58< Soliton> esr: yeah, i noticed. i was pointing out the similarity to our FPI (FrequentlyProposedIdeas) list.
20090902 16:31:17< CIA-62> esr * r38317 /trunk/data/campaigns/Two_Brothers/scenarios/2_The_Chase.cfg: Fix another Gray -> Grey glitch.
20090902 16:33:13< zookeeper> esr, any idea why people seem to constantly typo canon as cannon? -.-
20090902 16:33:18< zookeeper> (not wesnoth-related)
20090902 16:33:46< fendrin> Crab_: That is a good idea. The second thing maybe the FR about the lvl of the scouts. A parameter for the same thing but scouts.
20090902 16:34:15< Crab_> fendrin: btw, why L3 wolf has usage=scout ?
20090902 16:34:27< esr> zookeeper: Because 'canon' is a word only rather educated speakers know.  Others try to "correct" it to something familiar.
20090902 16:36:02< fendrin> Crab_: Maybe that is the problem.
20090902 16:37:10< fendrin> Crab_: But sometimes you may want to have high level scouts.
20090902 16:37:28< Soliton> the ai should just prefer low level scouts if it has the choice.
20090902 16:37:44< Soliton> for actual scout usage that is.
20090902 16:37:47< Crab_> fendrin: for what purpose ? the intent in some places of the code is for scouts to avoid enemies...
20090902 16:38:20< zookeeper> esr, educated? O.o uh, well, ok...i guess
20090902 16:38:27< Soliton> maybe the ai should just figure out what a scout is on its own. doesn't seem that difficult.
20090902 16:39:02< Soliton> just compare movement of available units considering the map.
20090902 16:39:22< esr> zookeeper: You speak English as a second language.  That actually gives you the vocabulary of a fairly educated speaker, by U.S. and British standards.
20090902 16:39:29< Soliton> and probably take cost into account as well.
20090902 16:40:27< Crab_> Soliton: our intent is to fix current ai recruitment with fewest changes possible, by parametrization, since we already have an experimental edition of recruitment done by DK.
20090902 16:41:01< Soliton> well, as i said just go by level then.
20090902 16:41:48< Ivanovic> esr: have you linked the "NotSpellingMistakes" page at teh "SpellingMistakes" page?
20090902 16:41:51< Ivanovic> ;)
20090902 16:42:17< Ivanovic> ah, okay, i see that you have linked it
20090902 16:42:54< Ivanovic> esr: uhm, were the changes too big to have them added to pofix or will you do this "soon"?
20090902 16:43:21< esr> Ivanovic: As I said, I got tired of explaining that "march" is not a typo for "marsh" ... :-(
20090902 16:43:26< Ivanovic> :)
20090902 16:44:00< esr> Does the "-mark" element occur in German as well as Scandinavian placenames?
20090902 16:44:59< esr> (I'm wondering if a German-speaker wasily reads "Estmark Hills" as "hills of the eastern border".
20090902 16:45:10< Ivanovic> yes, it does exist
20090902 16:45:15< esr> s/wasily/easily/
20090902 16:45:52< Ivanovic> that is: those knowing "old sounding german" will understand stuff like "In den östlichen Marken des Reiches" and all the likes
--- Log opened Wed Sep 02 17:03:56 2009
20090902 17:04:04-!- lobby [n=wesnoth@wesnoth/bot/lobby] has joined #wesnoth-dev
20090902 17:04:04-!- Topic for #wesnoth-dev: 83 bugs, 242 feature requests, 11 patches | logs: http://irclogs.wesnoth.org/ | Don't paste on IRC! Use a pastebin: http://wesnoth.pastebin.com | http://imagebin.org
20090902 17:04:04-!- Topic set by wesbot [] [Wed Sep  2 16:56:23 2009]
20090902 17:04:04[Users #wesnoth-dev]
20090902 17:04:04[ ABCD        ] [ Crab_     ] [ isaac_      ] [ shikadibot     ] 
20090902 17:04:04[ Aethaeryn   ] [ dfranke   ] [ Ivanovic    ] [ Smar           ] 
20090902 17:04:04[ AI0867      ] [ Doppp     ] [ kimrhh      ] [ thespaceinvader] 
20090902 17:04:04[ Alesis-Novik] [ erl_      ] [ knotwork    ] [ Tigge          ] 
20090902 17:04:04[ Amu_        ] [ Espreon   ] [ lobby       ] [ wesbot         ] 
20090902 17:04:04[ AnMaster    ] [ esr       ] [ loonybot    ] [ yamokidzu-it   ] 
20090902 17:04:04[ Baufo       ] [ ettin     ] [ loonycyborg ] [ yann           ] 
20090902 17:04:04[ Chusslove   ] [ fendrin   ] [ nonsensical ] [ zookeeper      ] 
20090902 17:04:04[ CIA-62      ] [ happygrue_] [ Rhonda      ] 
20090902 17:04:04[ corn        ] [ ilor_     ] [ shadowmaster] 
20090902 17:04:04-!- Irssi: #wesnoth-dev: Total of 38 nicks [0 ops, 0 halfops, 0 voices, 38 normal]
20090902 17:04:11-!- Soliton [n=Soliton@wesnoth/developer/soliton] has joined #wesnoth-dev
20090902 17:04:11-!- Soliton [n=Soliton@wesnoth/developer/soliton] has quit [Nick collision from services.]
20090902 17:04:18-!- Soliton [n=Soliton@wesnoth/developer/soliton] has joined #wesnoth-dev
20090902 17:04:20-!- Channel #wesnoth-dev created Tue Jan 27 06:28:41 2009
20090902 17:04:29-!- Soliton [n=Soliton@wesnoth/developer/soliton] has quit [Client Quit]
20090902 17:05:27-!- Soliton [n=Soliton@wesnoth/developer/soliton] has joined #wesnoth-dev
20090902 17:05:36-!- Irssi: Join to #wesnoth-dev was synced in 99 secs
20090902 17:06:21-!- Tesafilmchen [n=micha@p5B277F4A.dip.t-dialin.net] has joined #wesnoth-dev
20090902 17:08:00-!- alink [n=alink@wesnoth/developer/alink] has joined #wesnoth-dev
20090902 17:10:19< CIA-62> fendrin * r38318 /trunk/data/campaigns/Sceptre_of_Fire/maps/ (3 files): SoF 1,3,7: Some map updates.
20090902 17:11:19< CIA-62> fendrin * r38319 /trunk/data/campaigns/Legend_of_Wesmere/scenarios/02_Hostile_mountains.cfg: LoW 02: Removed TODOs.
20090902 17:12:25< CIA-62> fendrin * r38320 /trunk/data/campaigns/Legend_of_Wesmere/scenarios/22_Northern_Battle.cfg: LoW 22: Removed a line that forces the default.
20090902 17:13:08< Ivanovic> esr: anyway, is there any reason to not add the stuff from r38316 (your typo fix commit) to utils/pofix.py?
20090902 17:13:53< esr> I suppose not.
20090902 17:14:46-!- crimson_penguin [n=ben@wesnoth/developer/crimsonpenguin] has joined #wesnoth-dev
20090902 17:17:00< Ivanovic> fendrin: you mentioned some problem in LoW that basically prevents players being able to finish that campaign in 1.6.x
20090902 17:17:11< Ivanovic> any estimates when you might have a workaround ready?
20090902 17:17:49-!- Ivanovic changed the topic of #wesnoth-dev to: 1.7.5 planned for Thursday, September 3rd | 83 bugs, 242 feature requests, 11 patches | logs: http://irclogs.wesnoth.org/ | Don't paste on IRC! Use a pastebin: http://wesnoth.pastebin.com | http://imagebin.org
20090902 17:18:21< Ivanovic> currently i consider getting 1.6.5 (most likely the last 1.6.x release!) out either on sunday or some time next week
20090902 17:18:31-!- busfahrer [n=busfahre@unixboard/user/busfahrer] has joined #wesnoth-dev
20090902 17:19:29-!- dfranke [i=dfranke@wesnoth/developer/dfranke] has quit [Read error: 60 (Operation timed out)]
20090902 17:31:00< fendrin> Ivanovic: https://gna.org/bugs/?14172 ?
20090902 17:31:21< Ivanovic> jupp
20090902 17:32:01-!- Nobun [i=570ef7b3@gateway/web/freenode/x-mstmaoxwuoxvpyvd] has joined #wesnoth-dev
20090902 17:33:05< fendrin> Ivanovic: It's somewhere in the finishing dialogue of the scenario. Should be easy to avoid the failing wml code.
20090902 17:33:15< CIA-62> esr * r38321 /trunk/data/tools/wmllint: Fix a broken syntax check.
20090902 17:34:00< Ivanovic> fendrin: okay
20090902 17:35:03< fendrin> Crab_: I have a idea for scenario 21 elvish asassins.
20090902 17:35:12< Crab_> yes ?
20090902 17:35:21< fendrin> Crab_: are you familiar with the story?
20090902 17:35:27< Crab_> fendrin: somewhat
20090902 17:35:27-!- Tesafilmchen [n=micha@p5B277F4A.dip.t-dialin.net] has quit [Read error: 104 (Connection reset by peer)]
20090902 17:36:11< fendrin> Some of Landar's followers apear and want to murder kalenz.
20090902 17:36:34< fendrin> The scenario has nothing special in it.
20090902 17:37:48-!- stikonas [n=and@ctv-213-164-125-176.vinita.lt] has joined #wesnoth-dev
20090902 17:41:15< fendrin> Crab_: Let's give Kalenz an allied side, but the leader betrays klanenz. Before the first attack to one of Landar's soldiers we let him give the order to attack kalenz instead.
20090902 17:41:32-!- cib0 [n=cib@p5DD34351.dip.t-dialin.net] has joined #wesnoth-dev
20090902 17:41:35< fendrin> and change the team of that side
20090902 17:41:42< Crab_> fendrin: well, what the player will do when replaying that scenario a second time ?
20090902 17:41:53-!- dfranke [i=dfranke@wesnoth/developer/dfranke] has joined #wesnoth-dev
20090902 17:42:14< Soliton> yawn.
20090902 17:42:43< fendrin> He can try to obtain a good position for the comming side switch. But this can be difficult if we adjust the ai of the betrayer to stay close to kalenz troups.
20090902 17:46:42< CIA-62> esr * r38322 /trunk/data/tools/wmllint: Refactor to reduce parsing overhead.
20090902 17:47:51< Crab_> fendrin: we can give Kalenz two allied sides, and make one of them betray him :)
20090902 17:48:00< Crab_> fendrin: randomly chosen
20090902 17:48:42< fendrin> Crab_: great. but we choose that at scenario start. So the player must go back to start if he isn't satisfied with the randomness.
20090902 17:49:13< Crab_> ok
20090902 18:04:27-!- dfranke [i=dfranke@wesnoth/developer/dfranke] has quit [Read error: 60 (Operation timed out)]
20090902 18:05:52< CIA-62> esr * r38323 /trunk/data/tools/wmllint: Begin moving checks to reduce scan overhead.
20090902 18:17:14-!- Nobun [i=570ef7b3@gateway/web/freenode/x-mstmaoxwuoxvpyvd] has quit ["Page closed"]
20090902 18:21:47-!- Tesafilmchen [n=micha@p5B277F4A.dip.t-dialin.net] has joined #wesnoth-dev
20090902 18:27:16-!- dfranke [i=dfranke@wesnoth/developer/dfranke] has joined #wesnoth-dev
20090902 18:32:37-!- Aethaeryn [n=Michael@wesnoth/umc-dev/developer/aethaeryn] has quit ["bbl"]
20090902 18:38:07< alink> does trunk compiles for you ?
20090902 18:38:25< CIA-62> esr * r38324 /trunk/data/tools/wmllint: More refactoring to reduce code bulk.
20090902 18:38:30< alink> savegame.cpp:(.text+0x4578): undefined reference to `savegame::write_events(config&)'
20090902 18:38:59< Soliton> sounds like a linker error.
20090902 18:39:32< Soliton> it compiles for me but i'm not fully uptodate atm.
20090902 18:39:44< loonycyborg> alink: Probably someone neglected to update Makefile.am
20090902 18:39:58< alink> i use makefile indeed
20090902 18:39:58< loonycyborg> iirc there was such commit recently
20090902 18:40:13< CIA-62> soliton * r38325 /branches/1.6/src/server/ (game.cpp game.hpp server.cpp): wrapped most of the output_compressed() calls in a try-catch block with trivial error handling and refactored some code along the way
20090902 18:40:19< CIA-62> soliton * r38326 /branches/1.6/src/server/server.cpp: 
20090902 18:40:19< CIA-62> just save the era id from the previous scenario and use it in case no new era is sent
20090902 18:40:19< CIA-62> fixes the OOS part of bug #14223: MP Campaigns: Leader traits are not restored by save_id (OOS)
20090902 18:41:07< CIA-62> soliton * r38327 /trunk/src/server/ (5 files): wrapped most of the output_compressed() calls in a try-catch block with trivial error handling and refactored some code along the way
20090902 18:41:15< CIA-62> soliton * r38328 /trunk/src/server/server.cpp: 
20090902 18:41:15< CIA-62> just save the era id from the previous scenario and use it in case no new era is sent
20090902 18:41:15< CIA-62> fixes the OOS part of bug #14223: MP Campaigns: Leader traits are not restored by save_id (OOS)
20090902 18:41:31< alink> wesbot: log r38282
20090902 18:41:32< wesbot> euschn * r38282 : added savegame_config.cpp; added savegame::write_music_playlist and removed include of sound.hpp from savegame.cpp
20090902 18:41:35< wesbot> URL: http://svn.gna.org/viewcvs/wesnoth?view=rev&rev=38282
20090902 18:41:58< alink> seems that i need to port this to Makifile.am, I try
20090902 18:44:41< CIA-62> alink * r38329 /trunk/src/Makefile.am: Also add savegame_config.cpp (added by r38282) to Makefile.am
20090902 18:44:52< alink> done
20090902 18:47:04-!- dfranke [i=dfranke@wesnoth/developer/dfranke] has quit [Connection timed out]
20090902 18:54:35-!- dfranke [i=dfranke@wesnoth/developer/dfranke] has joined #wesnoth-dev
20090902 18:58:10< Crab_> lol... on LoW 22: objective is to kill all enemy leaders, including landar. I intend to skip the scenario, so I debug-create 3 armageddon drakes and used them to kill enemy leader, then kill landar. 
20090902 18:58:11< Crab_> Immediately after that, Kalenz cried "nooo! we cannot continue without landar" and the level ends with defeat msg - but, I still continue to next level.
20090902 18:58:14< Crab_> fendrin: ^
20090902 18:58:57< Soliton> kalenz is just really attached to him.
20090902 18:59:22< Soliton> probably another issue of the delayed end level stuff.
20090902 19:00:15-!- MikeJB [n=Michael@wesnoth/umc-dev/developer/aethaeryn] has joined #Wesnoth-dev
20090902 19:00:27< Soliton> or rather a non-sensical die event that wasn't noticed before.
20090902 19:00:42-!- giusef [n=giusef@unaffiliated/giusef] has joined #wesnoth-dev
20090902 19:01:00< Crab_> Soliton: well, normally, that scenario ends via a moveto event near landar, so he's not to be killed)
20090902 19:01:48< Soliton> right, so the die event is hidden.
20090902 19:01:53< Crab_> yes
20090902 19:02:45-!- mjs-de [n=mjs-de@vpw.wh.uni-dortmund.de] has joined #wesnoth-dev
20090902 19:03:16< CIA-62> alink * r38330 /trunk/src/astarsearch.cpp: 
20090902 19:03:16< CIA-62> Better pathfinding heuristic for A*.
20090902 19:03:16< CIA-62> Seems to helps nicely in some cases (East-West path on flat, O(d) instead of O(d^2))
20090902 19:03:16< CIA-62> but only few % better for more complex cases.
20090902 19:03:53< alink> Crab_: ^ thanks, your question about how A* scale with distance makes me spot (and solve) that flaw :)
20090902 19:04:01< Crab_> oh, nice :)
20090902 19:04:47< alink> it should be always O(d) now (when there is no obstacles or maze of terrain with higher MP cost)
20090902 19:06:13< alink> btw i also spotted a ugly perf killer for pathfinding: WML set unreachable to 99 and we often ask path with a >500 MP limit
20090902 19:07:13< Crab_> alink: well, with 100x100 maps, move from one corner to another might take many MP
20090902 19:07:15< Soliton> why such high limit?
20090902 19:07:18< alink> so if a target is in unreachable deep water, the pathfinding try to explore the all map to find a (impossible) way to access it
20090902 19:07:56< Soliton> is that for multi-turn paths?
20090902 19:08:09< alink> Soliton: yeah seems unneeded, IIRC these values were tweaked during the 1.6 AI bugs hunt
20090902 19:08:26< alink> Soliton: also for AI targetting, I check 
20090902 19:08:34< Crab_> alink: for ai targeting, yes
20090902 19:08:41< stikonas> esr: in Tutorial_2.cfg:28 you changed reference to Shaman from it to she. There are more instances of "it's" and "itself" on the same line.
20090902 19:08:51< Soliton> but can't you just limit it by the available MP then?
20090902 19:09:04< Crab_> Soliton: multi-turn targeting.
20090902 19:09:15< Soliton> right, ok.
20090902 19:09:20< esr> stikonas: I'l look at it.
20090902 19:09:26< alink> Soliton: well if you do that, the AI will consider far target unreachable.
20090902 19:09:46< Crab_> 'in how many turns unit at one corner of 100x100 map will reach enemy leader at opposite corner of 100x100 map?'
20090902 19:09:55< alink> it's tricky because the optimal value depends of map size, maze complexity, and terrain types
20090902 19:10:04< Soliton> well, we could easily change the impassable default to 999 if that helps.
20090902 19:10:22-!- MikeJB [n=Michael@wesnoth/umc-dev/developer/aethaeryn] has quit ["bbl"]
20090902 19:10:35< Crab_> alink: can we check the last hex and adjacent-to-last hexes for being passable ?
20090902 19:10:46< alink> Soliton: that's the easy solution that i wanted to propose :)
20090902 19:11:10< Crab_> (adjacent-to-last because it's possible to attack target which is standing on deep water/inside cave/etc)
20090902 19:11:12< alink> Crab_: and we can also do that
20090902 19:11:12< Soliton> or maybe something like -1 would work to use as a non-valid value?
20090902 19:12:54< alink> Soliton: I prefer avoiding special values in pathfinding code (for perf), but maybe we can translate them when reading/displaying WML
20090902 19:13:18< Soliton> sure, not all the time.
20090902 19:14:15-!- yamokidzu-it [n=rainy-it@host186-67-dynamic.21-79-r.retail.telecomitalia.it] has quit [Read error: 110 (Connection timed out)]
20090902 19:15:40< Soliton> currently we're resetting values below 0 to 1.
20090902 19:18:17< alink> seems a good place to do that 99 thing too, but maybe we need to take care to keep the 99 when writing savegame, even if don't see why some WML would check against this 99
20090902 19:18:34< alink> the movement_type code is a bit messy
20090902 19:18:39-!- Tesafilmchen [n=micha@p5B277F4A.dip.t-dialin.net] has quit [Remote closed the connection]
20090902 19:19:07< Soliton> well, in c++ impassable is actually 10000000.
20090902 19:19:48< Soliton> and yes 99 is assumed in a couple of places as impassable but that can be changed...
20090902 19:19:48< alink> there is also the  42424242.0 of GetNoPath()
20090902 19:20:04< alink> *GetNoPathValue()
20090902 19:21:46-!- Zarel [i=Zarel@c-24-118-17-28.hsd1.mn.comcast.net] has joined #wesnoth-dev
20090902 19:22:05< Soliton> probably not something that ever gets back to WML though.
20090902 19:24:03< alink> yeah 42424242 is purely pathfinding internal, 1000000 seems to be a default value for unspecified move_cost which using the 99 values avoid
20090902 19:25:50< alink> OTOH maybe we can be smarter and not only check move_cost but also always compare it with the max_MP of the unit (as we already do in some cases)
20090902 19:26:26< alink> I investigate that, if possible, it's probably more efficient and will solve this
20090902 19:27:11< Soliton> note that your current moves can be higher then max_moves.
20090902 19:28:00< alink> yes i was worried about special cases, but if it can be detected at init time of pathfinding, it should be ok
20090902 19:29:51< CIA-62> esr * r38331 /trunk/data/tools/wmllint: More code simplification.
20090902 19:29:59< alink> Crab_: btw the AI seems to ignore his current MP of unit for reachability. Is it something wanted ? (I suppose not, but not sure about the phases thing)
20090902 19:30:29< Crab_> alink: where this is happening?
20090902 19:30:33< alink> * reachability of target for rating them
20090902 19:31:11< alink> always the same crude move_cost_calculator.cost
20090902 19:31:44< Crab_> alink: it has plenty of bugs and assumptions
20090902 19:32:08< alink> Crab_: to give you an idea of more complex cost calculator, you may check shortest_path_calculator
20090902 19:32:13< Crab_> alink: for example, it thinks that the unit, which is moved, has either 0 MP left, or full MP left.
20090902 19:32:28< alink> (but it is much slower)
20090902 19:32:52< alink> Crab_: seems weird
20090902 19:33:26< Crab_> (usually this is right assumption)
20090902 19:35:10-!- yamokidzu-it [n=rainy-it@host104-69-dynamic.4-87-r.retail.telecomitalia.it] has joined #wesnoth-dev
20090902 19:35:45< alink> ok seems cheap to add max_MP checking and MP wasting (unit with 5MP moving on 2MP hexes), but I check
20090902 19:37:47-!- Crab_ [n=Crab_@wesnoth/developer/crab] has quit ["Leaving."]
20090902 19:39:58-!- Crab_ [n=Crab_@wesnoth/developer/crab] has joined #wesnoth-dev
20090902 19:41:06< Crab_> alink: it is interesting to know, that removing 'movement and targeting phase' altogether  (and replacing it with simple 'get enemy leader' movement phase) has little impact on overall ai power
20090902 19:42:01< alink> ai power = being how good it is to win against the normal ai?
20090902 19:42:04< Crab_> yes
20090902 19:42:53< Crab_> it becomes something like 45-55%, but mostly because the simplified ai forgets to take some villages.
20090902 19:44:13< alink> but does it means that the current AI has a poor 'movement and targeting phase', or that rushing leaders and ignoring villages is not a bad tactic
20090902 19:45:11< Crab_> imo, it means that current ai has inherently flawed (and slow) movement and targeting phase
20090902 19:45:12< alink> or that "rushing a leader who has sent his units grabbing villages" is a good tactic
20090902 19:46:20< Crab_> the problem is that the current AI decided on attacks before deciding on movement
20090902 19:46:43< alink> (btw i agree the targetting phase has big flaws, we talked about some of them)
20090902 19:47:15< Crab_> so, firstly, ai does 'good enough attacks' but then it's movement options are limited, so it loses badly to counterattacks next turn, because its units are exposed
20090902 19:47:46< Crab_> so, in MP games, a gold advantage 2:1 or higher is needed to offset the fact that the ai is not using troops efficiently.
20090902 19:48:30< Crab_> alink: so, the targeting phase works only when the ai is forced to target the enemy, bringing troops to battlefield as soon as possible
20090902 19:50:23-!- noy [n=Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20090902 19:51:25< alink> about "bringing troops to battlefield", i have a pathfinding feature/question. If there is several equivalent paths to reach a target, do we want some randomness to choose them, which will create a "spreading when moving towards target", or do we prefer "all move near the same path, causing more grouping"
20090902 19:51:49< Crab_> alink: no, no need to randomize anything
20090902 19:51:50< alink> well speed difference will also spread units along the path(s)
20090902 19:52:14< Crab_> alink: if you want to, you can try the following:
20090902 19:52:24< Crab_> alink: launch NR, :cl to Showdown,
20090902 19:52:29< Crab_> (in debug mode)
20090902 19:52:40< Crab_> then, do:
20090902 19:52:52< Crab_> !replace_ai 2 ai/dev/testing_ai_default.cfg
20090902 19:52:52< Crab_> !replace_ai 3 ai/dev/testing_ai_default.cfg 
20090902 19:52:52< Crab_> !replace_ai 4 ai/dev/testing_ai_default.cfg 
20090902 19:52:52< Crab_> !replace_ai 5 ai/dev/testing_ai_default.cfg 
20090902 19:52:52< Crab_> !replace_ai 6 ai/dev/testing_ai_default.cfg 
20090902 19:53:00< Crab_> from formula ai console, one at a time
20090902 19:53:23< Crab_> (it should say 'added [testing_ai_default] for side N from file...)
20090902 19:54:31< Crab_> and note the speed of targeting selection
20090902 19:55:14< alink> trying, but trunk crashes here, checking if it's my local changes
20090902 19:55:18< Crab_> ok
20090902 19:55:26< alink> (crash before your AI stuff)
20090902 19:56:02< alink> yeah loading savegame seems to cause config::error
20090902 19:56:25< CIA-62> crab * r38332 /trunk/data/ai/dev/testing_ai_default.cfg: removed fai recruitment from testing_ai_default
20090902 19:56:32< Crab_> alink: ^ for the ai to recruit properly (DKs recruitment is still beta and doesn't work in all cases)
20090902 19:56:35< alink> wait, didn't svn up
20090902 19:58:16-!- mordante [n=mordante@wesnoth/developer/mordante] has joined #wesnoth-dev
20090902 19:58:25-!- knotwork_ [n=markm@hlfxns01bbg-142177235097.pppoe-dynamic.ns.aliant.net] has joined #wesnoth-dev
20090902 19:58:31< mordante> servus
20090902 19:58:43< alink> hi mordante
20090902 19:58:49< mordante> hi alink 
20090902 19:58:50< Crab_> hi, mordante
20090902 19:58:54< mordante> hi Crab_ 
20090902 19:59:01-!- knotwork [n=markm@hlfxns01bbg-142177235097.pppoe-dynamic.ns.aliant.net] has quit [Read error: 104 (Connection reset by peer)]
20090902 19:59:19< Crab_> alink: if you update/try, I'll also ask you about a particular problem in ai target selection)
20090902 20:00:32-!- MikeJB [n=Michael@wesnoth/umc-dev/developer/aethaeryn] has joined #Wesnoth-dev
20090902 20:01:15< mordante> wesbot seen euschn
20090902 20:01:15< wesbot> mordante: The person with the nick euschn last spoke 8h 13m ago.  5h 18m ago they left with the message: Read error: 110 (Connection timed out)
20090902 20:02:36-!- yamekidzu-it [n=rainy-it@host4-65-dynamic.25-79-r.retail.telecomitalia.it] has joined #wesnoth-dev
20090902 20:04:57-!- dfranke [i=dfranke@wesnoth/developer/dfranke] has quit [Read error: 110 (Connection timed out)]
20090902 20:05:13< alink> Crab_: ok fai stuff done, how do I know the time spent in targetting selection (logs ?)
20090902 20:05:42< alink> btw "skip AI moves" seems broken
20090902 20:06:57< Crab_> alink: known issue
20090902 20:07:03-!- yamokidzu-it [n=rainy-it@host104-69-dynamic.4-87-r.retail.telecomitalia.it] has quit [Read error: 60 (Operation timed out)]
20090902 20:07:16< alink> ok loading savegame seems broken too, if someone can confirm, it's a serious bug
20090902 20:07:27< Soliton> i thought corn fixed that?
20090902 20:07:41< Soliton> (skip ai moves)
20090902 20:08:32< zookeeper> fendrin, how stable is DM in trunk atm? any notable bugs?
20090902 20:08:32< Crab_> Soliton: no, i tried to fix this but run into graphics invalidation issues and boucman was not around, so i postponed it.
20090902 20:09:02< Crab_> alink: you can use --log-debug=ai/manager to witness ai turn time. but, time difference of targeting is well-noticeable with unaided eye.
20090902 20:09:09< Soliton> ahh, why not get rid of it then?
20090902 20:09:36< zookeeper> fendrin, or if you know any other trunk bugs which would be annoying when playing trunk DM?
20090902 20:11:05< Crab_> alink: please watch the ai for 2-3 turns, and then I'll ask a question
20090902 20:11:53-!- dfranke [i=dfranke@wesnoth/developer/dfranke] has joined #wesnoth-dev
20090902 20:11:59< alink> ok i try again, but without savegames, these fai lines are annoying to enter each times:-/ 
20090902 20:12:10< Crab_> alink: copypaste them :)
20090902 20:12:45< alink> that i know, but i can copypaste all of them at once?
20090902 20:13:01< Crab_> no, one by one
20090902 20:13:15< Soliton> alink: loading works fine here.
20090902 20:14:23< alink> Soliton: ok thanks for the check, but no idea why mine don't work
20090902 20:14:41< Soliton> what doesn't work?
20090902 20:14:56< alink> well it crash when reloading
20090902 20:15:27< Soliton> oh, got a bt?
20090902 20:16:02< alink> ok will do, I finish the fai test of Crab_
20090902 20:16:12< Crab_> alink: no fai involved, actually :)
20090902 20:16:39< Crab_> alink: only 'ai_default implemented as RCA ai with simpler & faster targeting phase "go for side 1 leader"'
20090902 20:16:42< Soliton> maybe you need a clean build since you had issues before and it's make...
20090902 20:16:42< alink> Soliton: but i already says:  terminate called after throwing an instance of 'config::error'
20090902 20:17:11< alink> Soliton: ah yes make sense, I try that later, because it will take time
20090902 20:17:29-!- YogiHH [n=chatzill@d052081.adsl.hansenet.de] has joined #wesnoth-dev
20090902 20:17:44 * Soliton is happily using scons to never have such annoying issues again.
20090902 20:18:25-!- busfahrer [n=busfahre@unixboard/user/busfahrer] has quit ["leaving"]
20090902 20:19:02-!- giusef [n=giusef@unaffiliated/giusef] has quit [Remote closed the connection]
20090902 20:19:33< YogiHH> For all the C++ knowing people: I have a unit vector and i want to find a unit by its id. Surely there is a more elegant way then manually coding a foreach? And i could use an iterator as afterwards i need to erase that unit from the vector. Any help is appreciated.
20090902 20:20:03< alink> std::find ?
20090902 20:20:08< mordante> YogiHH, std::find
20090902 20:20:41< mordante> if you want to remove it std::erase and std::remove
20090902 20:21:09< YogiHH> ah, nice, thanks
20090902 20:21:33< mordante> sorry vector::erase
20090902 20:21:49< Crab_> YogiHH: also see src/ai/default/ai.cpp:1142
20090902 20:22:22< Crab_> if you want to remove not one unit, but a couple of units, by criteria
20090902 20:23:26< YogiHH> Crab_: ah, good to know as well :)
20090902 20:23:33< mordante> YogiHH, and  ai/testing/ca.cpp:1223 if you want to remove one unit
20090902 20:23:47< mordante> notice the difference between std::remove and std::remove_if
20090902 20:23:53-!- MikeJB [n=Michael@wesnoth/umc-dev/developer/aethaeryn] has quit ["bbl"]
20090902 20:28:24< Crab_> alink: so, how it's going ? have you reached turn 4-5 ?
20090902 20:28:38-!- DDR_away [n=chatzill@66.183.125.196] has joined #wesnoth-dev
20090902 20:29:18-!- DDR_away is now known as DDR
20090902 20:29:19< YogiHH> I am not sure what to provide for the third parameter of std::find. This is probably not right (unit_id is a string)?
20090902 20:29:20< alink> Crab_: tested 2 or 3 turns after the fai lines, currently trying the normal version
20090902 20:29:21< YogiHH> std::find(current_team.recall_list().begin(), current_team.recall_list().end(), unit_id)
20090902 20:36:09< Soliton> YogiHH: you need an actual unit or use find_if and then you need a function that takes a unit and returns true or false based on that unit_id.
20090902 20:36:44< Soliton> boost bind helps you to make such a function.
20090902 20:37:00< YogiHH> i see
20090902 20:37:50< Crab_> YogiHH: or, you can make a small class which overloads operator(), if you need a 'stateful' predicate to find_if
20090902 20:38:32< YogiHH> hmm, actually, i tried to find a simpler solution than foreach ;-)
20090902 20:38:43< alink> Crab_: mmh the blue side is certainly different
20090902 20:39:08< Soliton> isn't there already a function to get a unit by id?
20090902 20:39:23< Crab_> alink: so, how is the speed of the original ai_default ?
20090902 20:39:34< alink> Crab_:  btw several Error #2007 in full move by side 2 from location 30,30 to location 30,29
20090902 20:40:10< Crab_> alink: 2007 is E_NOT_REACHED_DESTINATION, it happens on ambush, for example
20090902 20:40:31-!- grzywacz [n=grzywacz@wesnoth/developer/grzywacz] has joined #wesnoth-dev
20090902 20:40:52< alink> I think testing_ai_default.cfg was faster to move towards me
20090902 20:40:53< wesbot> alink: Sometimes we are fast
20090902 20:41:31< alink> but without savegames it's hard to recheck a turn, i probably must fix that first
20090902 20:41:38< Ivanovic> AI0867, alink, Chusslove, cib0, corn, Crab_, crimson_penguin, dfranke, esr, fendrin, grzywacz, ilor_, loonycyborg, mordante, nonsensical, Rhonda, shadowmaster, Soliton, thespaceinvader, YogiHH, zookeeper, everyone else who cares: i plan to release tomorrow and i plan to do so earlier than normally, so basically in 18h i would like to tag things
20090902 20:41:46< Ivanovic> is there any reason why i should not do so?
20090902 20:41:50< Crab_> alink: no reason to recheck
20090902 20:41:55< Crab_> alink: but, a question
20090902 20:42:07< YogiHH> Ivanovic: not around here
20090902 20:42:20-!- wajimba [n=Andrew_A@24-158-30-63.dhcp.dlth.mn.charter.com] has joined #wesnoth-dev
20090902 20:42:22< thespaceinvader> Ivanovic: ok, i'll try to have the Drake Blademaster in before then
20090902 20:42:28< thespaceinvader> and, of course, not break anything this time
20090902 20:42:32< Crab_> alink: testing_ai_default.cfg is faster because it has exactly one target - side 1 leader
20090902 20:42:34< thespaceinvader> i'll check diffs before i commit etc
20090902 20:42:43< esr> Ivanovic: No blockers here.
20090902 20:43:08< alink> Crab_: which seems to visibly affect the AI behavior
20090902 20:43:18< Crab_> alink: if I want the ai to attack all three human sides, I need something to 'divide' the forces of the ai
20090902 20:43:37< mordante> Ivanovic, I've no blockers and nothing pending
20090902 20:43:40< Crab_> alink: I can do this 'by hand' (i.e. 'side 2 attacks west, side 3 attacks east, side 4 attacks north, etc)
20090902 20:43:45< Soliton> YogiHH: i guess you don't want find_unit_type()?
20090902 20:44:07< Crab_> alink: but, imagine that I want to let the ai sort it out
20090902 20:44:29< YogiHH> Soliton: no, i need to deal with the recall list, given as a unit vector.
20090902 20:44:58< alink> Crab_: wait, you want that the AI coordinate the 3 sides ?
20090902 20:45:02< Crab_> alink: the problem with ai_default is that it does it 'for each my unit, for each target, A* from unit to target and select 'best' unit-target pair, move"
20090902 20:45:22< Crab_> alink: I want the ai to attack all three 'player-allied' sides
20090902 20:45:27< Crab_> alink: not just 'side 1'
20090902 20:46:40< Soliton> YogiHH: oh recall list.. all the time i read recruit list and was wondering why that's not unit types... :-P
20090902 20:46:42< Crab_> alink: so, the orcs need to divide their forces to move to tagets
20090902 20:46:42< loonycyborg> IIvanovic: Who's 'nonsensical'? :P
20090902 20:46:45< alink> Crab_: so something like, pick different 'best' for each unit?
20090902 20:46:46< ilor_> Ivanovic: out of curiosity, why release earlier this time?
20090902 20:47:02< Crab_> alink: something like 'do something not stupid and do it fast'
20090902 20:47:07< Ivanovic> ilor_: some things got mingled a little with the drakes
20090902 20:47:21< Soliton> YogiHH: getting rid of index based logic?
20090902 20:47:21< Ivanovic> as in crashes when advancing to inferno drake
20090902 20:47:29< Crab_> alink: I don't really care if it's 'optimal', 'best', etc
20090902 20:47:31< YogiHH> Soliton: exactly
20090902 20:47:33< ilor_> Ivanovic: ah, fair enough
20090902 20:48:10< Ivanovic> plus screwed resistances for some drakes...
20090902 20:48:11< Crab_> alink: I want the ai to do something like 'oh, third part of an army goes west, third part goes east, third part goes north'
20090902 20:48:48< alink> Crab_: but is always a good idea ?
20090902 20:49:00< alink> * is it
20090902 20:49:06< Soliton> YogiHH: ok, a similar example is on src/server/ban.cpp:672
20090902 20:49:08< Ivanovic> loonycyborg: ???
20090902 20:49:18< Ivanovic> why do you ask this question?
20090902 20:49:19< YogiHH> Soliton: The savegame generating part works already, now i need to adapt the replay functionality. Not that big a deal, just some annoying STL stuff ;-)
20090902 20:49:22< Crab_> alink: well, when the ai is far from enemy,it doesn't matter much how exactly we go
20090902 20:49:27-!- fabi [n=fabi@e180234139.adsl.alicedsl.de] has joined #wesnoth-dev
20090902 20:49:34 * YogiHH checks
20090902 20:49:39< loonycyborg> You accidentally listed 'nonsensical' in the list of people
20090902 20:49:53< loonycyborg> who you asked about releasing tomorrow.
20090902 20:49:55< Crab_> alink: but, it should be terrain-aware
20090902 20:50:09< Ivanovic> esr: looks like i currently whistle the pofix.py song...
20090902 20:50:10< Ivanovic> ;)
20090902 20:50:18< Ivanovic> loonycyborg: he is the one that is too close to noy
20090902 20:50:43< Crab_> alink: i.e., if there was large pool of water to the north of orc keep, I'd expect all the ai sides to recruit enough mermen and move them north, and use other forces in other locations.
20090902 20:50:44< Soliton> YogiHH: you just need to write a very simple function that gets a unit and an id and checks whether they match and then you can use boost::bind to bind it with unit_id as one argument.
20090902 20:51:03< Crab_> alink: but, doing A* 'for each unit for each target' seems like an overkill
20090902 20:51:08< YogiHH> Soliton: mhm, ok
20090902 20:51:08< alink> Crab_: and we need to compare forces, to avoid splitting the AI army and having 3 losing "1/3 against 1" instead of a fair "1 vs 1"
20090902 20:51:14< Ivanovic> it really s**** when the users in the chan are not able to choose nicknames so that the first 2 letters are unique
20090902 20:51:17< Ivanovic> ;)
20090902 20:51:19< Crab_> alink: yes
20090902 20:51:33< Crab_> Ivanovic: write a script to list dev's names and aliases in a msg :)
20090902 20:51:55< alink> Crab_ aaah ok the "A* for each unit for each target" seems indeed overkill 
20090902 20:52:12< Crab_> alink: so, I want to get rid of it, somehow
20090902 20:52:32< Soliton> YogiHH: if you don't want to wrap your head around that a foreach is just fine. i think there is no real performance difference either it just looks nicer. :-)
20090902 20:52:48 * alink is thinking (but has nothing yet)
20090902 20:52:48< Crab_> one idea is 'if we have 5 mermen near each other (and they can reach each other), then one A* for all of them should be enough to get a good enough approximation of the distance to target'
20090902 20:53:12< YogiHH> Soliton: yes, i guessed so already ;-)
20090902 20:54:03< fabi> Ivanovic: I will do some testing but don't see any problem for the release.
20090902 20:54:11< noy> Ivanovic: !?!?!
20090902 20:54:13< alink> Crab_: or even use movement_type, so various unit_type can be combined
20090902 20:54:15< Ivanovic> fabi: okay
20090902 20:54:21< Crab_> alink: yes
20090902 20:54:28< Ivanovic> noy: i plan to release tomorrow and i plan to do so earlier than normally, so basically in 18h i would like to tag things
20090902 20:54:40< noy> no you said that one that is too close to me?
20090902 20:54:43< noy> whats that about?
20090902 20:54:54< Ivanovic> the first two letters are identical
20090902 20:54:54< Crab_> alink: also note that for 'get there and kill them all'' targets it makes little sense to reach them at different time
20090902 20:54:59< noy> ah
20090902 20:55:11< Ivanovic> (the same problem with silene and sirp as well as CP and Crab_)
20090902 20:55:34< Crab_> alink: maybe some 'group distance appropriator' can be created ?
20090902 20:55:51< alink> Crab_: do you really always use the distance (or move_cost) or do you only need a reachable in less than X MP answer ?
20090902 20:56:24< Crab_> alink: even 'reachable in less than X turns' is good enough
20090902 20:56:44< Crab_> alink: 'village' targets are perfectly fine with A*, since there's not a big amount of villages and village grabbers
20090902 20:56:46< alink> the later is more a connectivity problem and may have more way to combine results
20090902 20:57:48< Crab_> alink: I basically need 'if I'll move that group of units north, how many (approximate) turns it'll take them (providing that they wait for each other) to reach the enemy at X,Y ?
20090902 20:58:37< Crab_> the group is assumed 'connected'
20090902 20:59:52-!- wajimba [n=Andrew_A@24-158-30-63.dhcp.dlth.mn.charter.com] has quit [Read error: 54 (Connection reset by peer)]
20090902 21:00:23< Crab_> alink: note http://wesnoth.pastebin.com/m4af2ae7a
20090902 21:00:27< alink> Crab_: i will need some thinking, all ideas that i see are defeated by special terrain geometry cases
20090902 21:00:33< Crab_> it's a list of movement types of units
20090902 21:02:50< Crab_> alink: if we assume the group having movement_type's T1,T2,T3 is 'directly connected' (i.e. no impassable terrain between the group units), and if we check A* to (X,Y) from 'central unit', treating it as 'T1', then as T2, then as 'T3', will we be able to get an upper bound on A* result of other units to X,Y ?
20090902 21:04:03-!- fendrin [n=fabi@wesnoth/developer/fendrin] has quit [Read error: 110 (Connection timed out)]
20090902 21:04:11< Crab_> since, if we know distance from unit U1 to (x,y), then 'distance from unit U2 to (x,y) is <= distance from U2 to U1, and then from U1 to (x,y) '
20090902 21:05:32< alink> Crab_: with 3 A*, yes seems possible, but not sure how easy to implement simply 
20090902 21:07:19-!- lizard_r [n=Miranda@wesnoth/umc-dev/developer/lizard] has joined #wesnoth-dev
20090902 21:07:37< Crab_> alink: what about some kind of caching. 'if we have A* result for the same movement_type from the map_location near enough (by pathfinding) to us, use the cached result)
20090902 21:08:13< Crab_> alink: if 'near enough (by pathfinding) to us' is 'location which is reachable by us in 1 turn' then the cached result will be at most 1 turn wrong.
20090902 21:09:41< alink> yes that is also possible, but you also need to have a stable cost function (not affected if there is unit changes between A* calls)
20090902 21:09:54< Crab_> alink: 1) we can ignore ZOC
20090902 21:10:07< Crab_> since if we meet enemy, we'll fight him anyway
20090902 21:10:27< Crab_> and to avoid 'unable to target enemy leader fully shielded by friendly troops' issue
20090902 21:10:59< alink> so we ignore units too ?
20090902 21:11:01< Crab_> yes
20090902 21:11:08< Crab_> we do care for terrain
20090902 21:12:04< Crab_> those A* calls are basically 'how fast we can relocate near to (x,y) ?' questions
20090902 21:13:01< Crab_> and they're for 'mass movement' targets, not precision targets like villages
20090902 21:13:28< Crab_> also, 'quick' trait can be ignored
20090902 21:13:42< fabi> zookeeper: In a attack event, how can I interrupt the attack?
20090902 21:14:02< Crab_> it makes a huge difference on some terrain types, but we don't want to 'make some units reach there fast and die even faster' problem
20090902 21:14:30< Crab_> if we (later) decide to care for 'quick', we can always do that (by considering quick and non-quick units to have a different movement type)
20090902 21:14:33< alink> Crab_: well, an A* on terrain only can be very fast and even made faster, maybe a fast brute force way is enough
20090902 21:14:54-!- Chusslove [n=Chusslov@212.124.170.166] has quit [Read error: 145 (Connection timed out)]
20090902 21:15:14< Crab_> alink: yes, 'on terrain only' is enough to, for example, divide units into 'closer to enemy side 1' and 'closer to enemy side 2'
20090902 21:16:16< Crab_> alink: imagine a map with a lake to the north, 'team 1' mermen at 12 o'clock, 'team 1' undead at 6 o'clock, 'team 2' orcs in the center
20090902 21:16:35< Soliton> alink: oh, seems like i tested the wrong version. i can reproduce the crash.
20090902 21:16:46< Crab_> alink: let's say that orcs recruit 3 nagas, and 3 grunts
20090902 21:17:06< alink> Soliton : ok thanks, i try the make clean, just to be sure
20090902 21:17:23< Soliton> hmm, first load worked though.
20090902 21:18:00< Crab_> alink: 'terrain-only A*' will give us (in 2 A* runs only) the answer that 'to get to loyalists (mermen) to the north, use nagas, to get to undead leader to the south, use grunts'
20090902 21:18:03< Soliton> and exiting a game gives me: "Mandatory WML child missing yet untested for. Please report."
20090902 21:18:54< Crab_> so, the fishes will happily swim north, and grunts will go south.
20090902 21:19:11< Crab_> oops, not 2, 4 A* runs (2x target x 2 movement type)
20090902 21:19:18< alink> Crab_: mmh no you need to test each group against 2 targets to see which one is better, so 4 A*
20090902 21:19:22< alink> :)
20090902 21:19:25< Crab_> yes
20090902 21:19:39< Crab_> but today that's 12 A* runs
20090902 21:19:56< Crab_> since there are 6 units
20090902 21:20:19< alink> Crab_: but does such trick stay useful when units start to spread around the map ?
20090902 21:20:58< Crab_> alink: yes, since units tend to support each other
20090902 21:21:07< Crab_> alink: most importantly, it scales better.
20090902 21:21:14-!- yamekidzu-it [n=rainy-it@host4-65-dynamic.25-79-r.retail.telecomitalia.it] has quit [Read error: 104 (Connection reset by peer)]
20090902 21:22:44< Crab_> alink: also, current ai adds some targets from time to time. this increases the total number of targets on the map and makes things even slower.
20090902 21:22:54< Soliton> alink: looks like upload log stuff that throws the uncaught config error.
20090902 21:22:58< Soliton> corn: ^
20090902 21:23:13< Crab_> alink: i.e. it considers all villages a target, it considers all 'attacks done this turn where the enemy was wounded' as targets, etc
20090902 21:23:22-!- mjs-de [n=mjs-de@vpw.wh.uni-dortmund.de] has quit ["On the road again"]
20090902 21:23:55< alink> so we can't assume less targets than units ?
20090902 21:24:15< Crab_> alink: I think that we must use only 'scenario objectives' as targets
20090902 21:24:19< zookeeper> fabi, at least by removing either unit and then unstoring it back in the attack end event
20090902 21:24:42< zookeeper> fabi, btw, looks like mage leader in DM doesn't have a leadership animation
20090902 21:24:45< Crab_> alink: i.e. 'kill this enemy, protect this location, escort this unit, etc'
20090902 21:25:17< Crab_> alink: those smaller targets do nothing that is really important
20090902 21:25:34< alink> because i was going to propose some pathfinding *from* targets, looking for interested units 
20090902 21:26:12< Crab_> alink: yes, this is possible wrt villages, but, it's not that necessary
20090902 21:26:15< alink> but only good if we a limited number of target (like leaders only)
20090902 21:26:29< alink> *have
20090902 21:26:49< Crab_> its not that necessary because moves which require only 1 unit to do can be handled by candidate actions / specialized stages (such as 'get_villages')
20090902 21:27:24< Crab_> alink: so, I suggest rewriting a 'move to targets' phase from scratch, to allow easier division of forces towards a small number of target areas
20090902 21:27:50< fabi> zookeeper: right, where should it come from?
20090902 21:27:56< Rhonda> corn: around now
20090902 21:28:33< alink> Crab_: "target area", does that means that some A* towards an area (instead of a location) could be useful ?
20090902 21:28:35< Crab_> input will be std::vector<strategic_target>, where strategic_target is 'map_location + value'
20090902 21:28:55< Crab_> alink: yes, 'camp around this location' is very useful for campaigns.
20090902 21:29:08 * alink is considering allowing different heuristic (as we do for cost) in A*
20090902 21:29:11< Crab_> alink: right now, there's now way to make the ai intelligently defend an area
20090902 21:30:05-!- Chusslove [n=Chusslov@adsl-ppp-2063.yubc.net] has joined #wesnoth-dev
20090902 21:30:27< Crab_> current movement_and_targeting phase babysits units too much - it brings them 'right to the spot' and keeps them there
20090902 21:30:35< alink> Go to an area centered in X,Y having a N hexes radius seems doable
20090902 21:30:52< Crab_> alink: yes. 'and once you get there, you're free to do anything you want'
20090902 21:31:38< alink> just to be clear: I mean the A* give you the distance+path to the nearer hex of that area
20090902 21:32:56< zookeeper> fabi, uh, what? the leadership anim? just use some of the magic casting frames for that i guess?
20090902 21:33:25< Crab_> alink: I need two things: 1) assign 'available' units to targets (this is easily done if we have approximate distances for each unit for each target)
20090902 21:34:05< Crab_> 2) assign a good enough path to each unit to go 'somewhere nearer to the selected target'
20090902 21:34:33< Crab_> (if we have that path for the 'group leader', we can just make other units follow it)
20090902 21:35:27< Ivanovic> is it only me or is www.wesnoth.org (with all the other subparts like g.w.o) damn slow atm?
20090902 21:35:28< fabi> zookeeper: Hmmm, why not.
20090902 21:35:53< alink> Crab_: 'group' being unit having similar movement_type ?
20090902 21:36:41< Crab_> alink: yes
20090902 21:37:13< Crab_> alink: since we spoke about caching the result by 'near location and same movement type'
20090902 21:37:28< Crab_> alink: so, the 'leader which we follow' will be 'near enough' and will have same movement type
20090902 21:38:24-!- YogiHH [n=chatzill@d052081.adsl.hansenet.de] has left #wesnoth-dev []
20090902 21:38:27< Ivanovic> Soliton: is currently anything heavy on IO running?
20090902 21:38:36< Ivanovic> since top shows me rather low cpu usage
20090902 21:38:52< Crab_> main aim of that movement phase is to 'divide' units towards 'suitable enough' targets, and bring them closer to those targets
20090902 21:39:51< Soliton> Ivanovic: nothing heavy should be running and wesnoth.org works fine for me.
20090902 21:40:26< Ivanovic> okay, then it probably has to do with the routing
20090902 21:40:38< Crab_> alink: it will also work via 'evaluate just once, and voila - all moves are known and ready to execute' way
20090902 21:40:45< loonycyborg> Ivanovic: low cpu usage is normal unless you compile something.
20090902 21:40:58< alink> Crab_: the movement_type code is messy but if we look only unit_type it may be doable (but special units may have problems)
20090902 21:41:32< Crab_> alink: if the ai uses special units not efficiently, it's the problem of scenario creator. usually they have their own agenda, anyway.
20090902 21:41:51< Ivanovic> wow, got rather high package loss
20090902 21:45:59< alink> Crab_: ok, i am about to go eat. So you want some faster or less often called A*, by using caching or approximation (to turn or movement_type).
20090902 21:46:22< Crab_> alink: yes
20090902 21:46:23< alink>  I will think about that, because there is various other possibilities
20090902 21:47:03< Crab_> alink: I'll create a test scenario, with 'ai in the center, two enemies on sides, one enemy reachable via water, other via land, ai has 3 grunts and 3 mermen'
20090902 21:47:13< alink> for example, for one movement_type, the map is a just a grid of 1,2,3,99, less various that we you look terrains
20090902 21:47:30< Crab_> alink: yes, you're right...
20090902 21:47:31< alink> s/we.when
20090902 21:48:03< Crab_> and for 100x100 map, the content-addressable table of all those movement costs is very small in size...
20090902 21:48:30< Crab_> and, the number of movement types is not high
20090902 21:51:11< alink> ok, bbl
20090902 22:09:07< mordante> I'm off night
20090902 22:09:30-!- mordante [n=mordante@wesnoth/developer/mordante] has quit ["Leaving"]
20090902 22:33:04-!- yamokidzu-it [n=rainy-it@host13-75-dynamic.1-79-r.retail.telecomitalia.it] has joined #wesnoth-dev
20090902 22:34:28< stikonas> wesbot needs evil pingall :) command, so that Ivanovic will not have to write all names when pinging for release
20090902 22:35:03-!- yamokidzu-it is now known as KenN-hoOb-IT
20090902 22:35:51< Crab_> stikonas: why 'evil' ping-all ?
20090902 22:36:01< Crab_> stikonas: what's the difference between 'good' and 'evil' ping-alls ?
20090902 22:37:33< stikonas> #amarok's bot has pingall command and when somebody executes it, about 180 people are pinged (that's evil :))
20090902 22:37:41< fabi> Somewhere in data is a suf with attribute team= but that attribute isn't recognized by a suf. Anyone an idea how to find it?
20090902 22:38:05< Crab_> fabi: fgrep ?
20090902 22:38:18< Crab_> fabi: or abuse of wmllint ?
20090902 22:38:59< fabi> Crab_: you can use fgrep to find team= inside [filter] ?
20090902 22:39:26< fabi> esr: can wmllint find a team= in a [filter]?
20090902 22:40:32< Crab_> fabi: ./campaigns/The_Rise_Of_Wesnoth/scenarios/17d_Cursed_Isle.cfg:427
20090902 22:41:31< fabi> Crab_: how did you find it?
20090902 22:41:34< Crab_> fgrep -Rn -a6 -b6 "team=" ./ | grep filter
20090902 22:41:49< fabi> black magic
20090902 22:42:21< esr> fabi: If I teach it to, it can.
20090902 22:42:55< fabi> esr: Crab_ did it with some fgrep commandline black magic.
20090902 22:43:00< Crab_> fabi: SUFs are usually small, so I found all the snippets of 'team=' with 6 lines of context, and then, done an additional search to find any occurrences of 'filter' in those snippets
20090902 22:43:50< fabi> who maintains RoW?
20090902 22:44:17< zookeeper> me
20090902 22:45:09< fabi> zookeeper: it's not a big issue since this is only some extra speech. But I am sure it's not working as it is now.
20090902 22:45:24< zookeeper> i'll fix it
20090902 22:45:36< fabi> zookeeper: replace team=Undead with side=whatever, more whatever
20090902 22:45:37< zookeeper> however i'd expect that event to fire when he kills any unit
20090902 22:45:44< zookeeper> and IMO it should
20090902 22:46:24< fabi> zookeeper: right, the problem is: the wesnoth emacs mode learns from wrong wml. That is how i found the issue.
20090902 22:46:26-!- yamokidzu-it [n=rainy-it@host164-69-dynamic.4-87-r.retail.telecomitalia.it] has joined #wesnoth-dev
20090902 22:47:06< CIA-62> zookeeper * r38333 /trunk/data/campaigns/The_Rise_Of_Wesnoth/scenarios/17d_Cursed_Isle.cfg: Fixed a broken unit filter.
20090902 22:49:42< fabi> zookeeper: why is 
20090902 22:49:43< fabi>   [set_variable]
20090902 22:49:44< fabi>             name=betraying_side
20090902 22:49:46< fabi>             rand=2..3
20090902 22:49:52< fabi> always 2?
20090902 22:51:22< zookeeper> bad luck? :P
20090902 22:52:33< fabi> Tried it for 20 times. Can't have such much "bad luck".
20090902 22:52:53< zookeeper> tried random= instead?
20090902 22:53:06< fabi> no
20090902 22:53:51< zookeeper> well, that's the only thing that comes to mind
20090902 22:54:21< zookeeper> and/or then trying different kinds of random strings to see if something else works
20090902 22:54:44< fabi> replacing rand with random does the trick
20090902 23:03:15-!- Zarel [i=Zarel@warzone2100/developer/Zarel] has quit [Read error: 110 (Connection timed out)]
20090902 23:03:19-!- KenN-hoOb-IT [n=rainy-it@host13-75-dynamic.1-79-r.retail.telecomitalia.it] has quit [Read error: 110 (Connection timed out)]
20090902 23:04:05< Ivanovic> stikonas: in fact i doN't want to have *everyone* pinged
20090902 23:04:20< Ivanovic> just those in where that are likely to have something to commit (as in people with commit access)
20090902 23:05:31< zookeeper> anyone want to play some trunk MP?
20090902 23:05:46-!- lizard_r [n=Miranda@wesnoth/umc-dev/developer/lizard] has quit ["Saurian Augur - I'll heal you by 4 hp if you leave next to me"]
20090902 23:06:15< Crab_> Ivanovic: then the script can get 'who has committed since previous release, do an intersection with people currently on IRC, and prepare that message :)
20090902 23:06:38< Ivanovic> Crab_: commited since last release is not enough
20090902 23:06:46< Ivanovic> Crab_: eg see the time since the last release
20090902 23:06:54< Crab_> ok )
20090902 23:07:03< Ivanovic> i'd guess that especially those that have *not* commited yet might have something left they want to get in
20090902 23:07:05< Ivanovic> ;)
20090902 23:07:22< Crab_> true
20090902 23:07:53< zookeeper> no MP? crud.
20090902 23:11:09< Crab_> zookeeper: I can try, but i'm not sure if my trunk is recent enough
20090902 23:11:16< Crab_> (yesterdays)
20090902 23:12:18-!- boucman [n=rosen@wesnoth/developer/boucman] has joined #wesnoth-dev
20090902 23:12:29< zookeeper> Crab_, well, probably just fine
20090902 23:15:21< boucman> hey all
20090902 23:15:27< Crab_> hi, boucman
20090902 23:15:33< Ivanovic> hi boucman
20090902 23:19:00< Soliton> fabi: want to play?
20090902 23:19:22< fabi> Soliton: yepp
20090902 23:19:26< zookeeper> excellent
20090902 23:19:29< zookeeper> get on the trunk server
20090902 23:19:43 * fabi compiles first
20090902 23:19:55< Ivanovic> okay, heading off to bed now, got to get some sleep so that i can maybe get out 1.7.5 tomorrow this early in the day...
20090902 23:19:57< Ivanovic> n8
20090902 23:21:09< CIA-62> crab * r38334 /trunk/data/ (4 files in 4 dirs): new test scenario: move_to_targets, for AI improvement
20090902 23:21:24< Crab_> alink: ^  ./wesnoth -t move_to_targets -d 
20090902 23:21:51< Crab_> alink: it's the test scenario for 'fish should go left, grunts should go right'
20090902 23:21:57< alink> Crab_: ok i will try
20090902 23:21:58< Crab_> alink: current ai fails badly.
20090902 23:23:33< alink> Ivanovic: mmh about releasing, the reload savegame crash need to be fixed (and tested) first
20090902 23:24:07< Crab_> boucman, anyone else interested in 'fishes on land' feature present in default ai : you can also check that test scenario :)
20090902 23:24:07< Ivanovic> alink: uhm, ah, which "reload savegame crash"?
20090902 23:24:17< fabi> esr: Is there a chance for reprosing some more strings in low 21 before the release?
20090902 23:24:35< boucman> Crab_: not tonight, I try to get LTO anims working for thespaceinvader before tomorow's release
20090902 23:24:40< Crab_> ok
20090902 23:24:41< esr> Probably.  What do you want done?
20090902 23:24:52< zookeeper> fabi, how long will your compile take?
20090902 23:24:53< alink> Ivanovic: loading savegame crashes here and I think Soliton confirmed it
20090902 23:25:03< Ivanovic> OUTSCH
20090902 23:25:10< Ivanovic> yogihh is aware of this?
20090902 23:25:11< fabi> zookeeper: it's finished , installing ...
20090902 23:25:16< Ivanovic> and also euschn?
20090902 23:25:27< Soliton> i think corn will probably fix it tonight.
20090902 23:25:38< Soliton> it's something in the upload log.
20090902 23:25:42< fabi> esr: There are TODO ESR in the file. One of the allies of kalenz is changing the side during the combat...
20090902 23:25:47< zookeeper> fabi, ok, great
20090902 23:25:51< Ivanovic> Soliton: ah, okay
20090902 23:27:08< esr> fabi:  I'll take a look.  Note, I'm doing your request about detecting units speaking in their death events now.
20090902 23:27:32< alink> it is blocker (but seems to test)
20090902 23:27:43< alink> *seems easy to test
20090902 23:27:55< Crab_> Ivanovic: if it's corn's, then it's probably r38311
20090902 23:28:04-!- grzywacz [n=grzywacz@wesnoth/developer/grzywacz] has quit [Remote closed the connection]
20090902 23:28:42< Soliton> oops need to recompile all campaingd since the boost update...
20090902 23:29:19< fabi> esr: you need to match at least id in the filter with speaker=id_from_filter or id=id_from_filter or speaker=unit
20090902 23:29:26< Crab_> there's untested  ' game->child("upload_log").child("ai_log") ' operation there, it should fail if game has no upload_log child
20090902 23:29:49< Ivanovic> ugh, looks like the typo fixes esr commited in r38316 is a huge load of string changes, will be "fun" to add to pofix...
20090902 23:30:02< esr> fabi: I know.  There's also the case where speaker is "unit"
20090902 23:30:28< esr> Ivanovic: Have fun doing it, I'm backed up wit three other requests right now.
20090902 23:32:12< Ivanovic> argh and for this stuff the gna webinterface *really* sucks regarding diffs sicne it just shows the changed lines which means complete strings
20090902 23:32:37< Ivanovic> is there some good and easy way to get the *exact* changes (as in changed letters) marked nicely
20090902 23:32:44< Ivanovic> the changes between two svn revisions that is
20090902 23:34:20< alink> Crab_: nice test case, indeed ,the AI started well, sendind 2 nagas and 3 orcs in the good directions, but after that, it's less good
20090902 23:34:56< Crab_> alink: so, we want to ai to behave correctly. it values both leaders the same.
20090902 23:35:23< alink> seems to do that, but badly uses terrains
20090902 23:35:38-!- yann [n=dwitch@nan92-1-81-57-214-146.fbx.proxad.net] has quit [Read error: 110 (Connection timed out)]
20090902 23:35:47< Crab_> it should send grunts only east, and nagas only west
20090902 23:35:59< Crab_> and it should *stick* to the choices unless something changes
20090902 23:36:19< shadowmaster> Ivanovic: eh, why 1.7.5 tomorrow?
20090902 23:36:35< loonycyborg> shadowmaster: Due to bugs.
20090902 23:36:37< shadowmaster> (sorry, a certain problem with the house's heating has kept me away)
20090902 23:36:38< Ivanovic> shadowmaster: because the drakes are screwed badly in 1.7.4 (even leading to a crash)
20090902 23:37:27< alink> Crab_: the good news is that A* should be ~2x faster on such map now :)
20090902 23:37:36< Crab_> alink: that's good
20090902 23:38:14< boucman> thespaceinvader: commited, do you have a second to discuss how this works  ?
20090902 23:38:21< CIA-62> boucman * r38335 /trunk/ (changelog src/unit_animation.cpp src/unit_display.cpp): new animations : pre_movement_anim and post_movement_anim for drake take-off and landing
20090902 23:40:24< Crab_> alink: so, I want to create a new movement phase from scratch, which will not 1) send fishes on land 2) will not send a western fish east at the same time as sending an eastern fish west
20090902 23:41:08< thespaceinvader> boucman: sure
20090902 23:41:27< thespaceinvader> but i'll probably wait to test until after Ivanovic releases 1.7.5 if that's OK by you
20090902 23:41:40< boucman> ok, so basically you can declare [pre_movement_anim] and [post_movement_anim] that will be player before/after each moe
20090902 23:41:48< boucman> that's fine
20090902 23:42:04< boucman> with that and terrain filter, you should be able to do all the movement+standing animations
20090902 23:42:11< thespaceinvader> OK, that seems logical
20090902 23:42:24< thespaceinvader> thanks very much, i'll look into implementing and testing it after 1.7.5
20090902 23:42:32< shadowmaster> although/exec free -m
20090902 23:42:35< boucman> (my advice: do a macro ALL_FLYING_TERRAIN and ALL_LANDING_TERRAIN for easy update)
20090902 23:42:36< shadowmaster> buh.
20090902 23:42:43< alink> Crab_: but ridiculous AI are so funny <:o) 
20090902 23:42:44< thespaceinvader> hoefully, that should allow us to streamline the drakes a MASSIVE amount with macros
20090902 23:42:59< boucman> thespaceinvader: you would also be able to do the fighting part but it's a bit long to explain so i'll do that another day
20090902 23:43:24< Crab_> alink: yes, fishes on land never fail to deliver :)
20090902 23:43:27< thespaceinvader> given that in terms of the flight animation (and fire animations, which i'm planning on macro-ing too), all the frame names are the same apart from the name of the actual drake
20090902 23:43:49< boucman> thespaceinvader: yes that should make things much easier indeed
20090902 23:44:19< thespaceinvader> boucman: how does that fit with the glider line units that want to land only to attack and defend?
20090902 23:44:25< thespaceinvader> or will you explain that later on?
20090902 23:44:31< boucman> later
20090902 23:45:23< stikonas> Ivanovic: if you have a diff, kompare can show *exact* changes
20090902 23:46:07< shadowmaster> too bad kompare doesn't have line-warpping
20090902 23:46:11< shadowmaster> *wrapping
20090902 23:46:17< Ivanovic> stikonas: i plain diff file is enough?
20090902 23:46:58< stikonas> Ivanovic: yes
20090902 23:47:11-!- Baufo [n=thomas@wesnoth/developer/baufo] has quit ["Leaving"]
20090902 23:47:20< stikonas> the characters that are changed, will be shown in bold
20090902 23:49:31< zookeeper> err, did everyone else drop too, or just me?
20090902 23:49:44< Crab_> zookeeper: I'm ok
20090902 23:49:59< zookeeper> right, rejoining..
20090902 23:53:11-!- Blueblaze [n=nick@adsl-76-202-22-1.dsl.hstntx.sbcglobal.net] has joined #wesnoth-dev
20090902 23:54:41< Soliton> campaign servers are up again.
20090902 23:55:25< Soliton> Blueblaze: fix the spawn labels in survival fusion!
20090902 23:55:32< Blueblaze> !
20090902 23:55:40< Blueblaze> 3rd person to tell me that in 1 minute
20090902 23:56:10< Soliton> we're playing survival fusion on the trunk server atm.
20090902 23:56:16< Blueblaze> oh
20090902 23:56:23 * Blueblaze didn't upload that version
20090902 23:56:33< fabi> Blueblaze: And the fusion units are much to overpowered.
20090902 23:59:54< zookeeper> you didn't?
--- Log closed Thu Sep 03 00:00:48 2009