#bitfighter IRC Log

Index Search ←Prev date Next date→

IRC Log for 2013-06-13

Timestamps are in GMT/BST.

00:01:40raptor Quit ()
00:03:41watusimoto has joined
00:03:41ChanServ sets mode +o watusimoto
00:14:59watusimoto Quit (Quit: Leaving.)
01:55:48koda has joined
04:59:26koda Quit (Read error: Connection reset by peer)
06:07:26koda has joined
06:48:27raptor has joined
06:48:27ChanServ sets mode +o raptor
07:50:55Nothing_Much Quit (Ping timeout: 260 seconds)
08:04:00Nothing_Much has joined
08:12:09Nothing_Much Quit (Read error: Connection reset by peer)
08:27:51Nothing_Much has joined
08:29:46Nothing_Much Quit (Changing host)
08:29:47Nothing_Much has joined
08:32:01raptoronly 3 servers up!
08:37:06watusimoto has joined
08:37:06ChanServ sets mode +o watusimoto
08:51:11watusimotohello!
08:51:15raptorhi
08:51:30watusimotosaw your note re: luajit
08:51:49watusimotoI agree with your judgement
08:52:20raptorthe stars need to align before we decide on it..
08:52:22watusimotoif we do move foward with luajit, I also agree that we should abandon luavec, for the reasons we discussed earlier (and that you touched on in your note)
08:52:49watusimotothough I thought we were now builidng with cmake on osx
08:53:04raptorwell
08:53:09raptorI got it so it can build the binary
08:53:19raptorbut not layout the shared frameworks and create the bundle yet
08:53:28raptor*lay out
08:53:29watusimotoI see
08:54:13watusimotoI have a checkin I am trying to make, but the hg repo I have on this computer is totally screwed up
08:54:41watusimotoI either need to reset it, or figure out how to prune a whole branch that was never merged
08:54:41raptorrevert, revert!
08:54:46raptoroh easy
08:54:49raptorhg strip
08:55:14raptorhg strip abcdef123
08:55:19watusimotoI don't want to revert as I have my good changes!
08:55:28raptorwill prune every revision from that one on to the end of the branch
08:55:29watusimotostrip is probably what OI'm looking for
08:56:12watusimotois there any way to find all branches that would be candidates for prune? I've been through our enitre tree at least 3 times and I don't want to do it again if this next prune doesn't fix the problem
08:56:45watusimotobut I'll try stripping and see if I found the last branch
08:57:00raptorwell, you can look at the 'heads' with 'hg heads'
08:57:27raptorwhen you find one, i look at it in a tree-UI and follow it until i find the base where it split
08:57:49watusimotogood idea
08:57:57raptorour repo should only have 2 heads
08:58:00watusimotoI need to enable an extension to enable prune, no?
08:58:03raptortip and lua52
08:58:08watusimotomine has several
08:58:09raptorstrip
08:58:11raptoruhh
08:58:15raptorlet me see..
08:59:13watusimotoI got my colleageus to play a bitfighter match a few days ago
08:59:23watusimotoit's great to have a group in the same room playing
08:59:24raptormq=
08:59:30raptorthe mq extensions
08:59:32watusimotomq, that's wht iI thought
08:59:33raptorgreat!
08:59:44raptoryeah, bitfighter is very fun with a group in the same room
09:00:15watusimotoit's a great lan game
09:01:16raptoryeah - i consider that one of it's strengths
09:01:18raptor*its
09:01:52raptorwe need to infiltrate more school parking lots with USB flash drives..
09:02:00watusimotoindeed
09:02:14watusimotoThey were all somewhat symied by the complexity of the controls
09:02:19watusimotounfortunately
09:02:22raptor:(
09:02:33raptoranything in particular? or just the dual-axis, multi-module aspect?
09:02:38watusimotothat is what I regard as bf's greatest weakness... it is too complex to pick up and play
09:03:32watusimotoI'm not sure I can describe it at the moment... I need to debrief them next week. move and fire were ok, with a little handholding... but modules were complex, and loadouts were very complex.
09:03:44raptorhmm
09:03:48watusimotothey're not gamers per se, but are not game illiterate, either
09:04:11watusimotoone kept trying to capture a loadout zone in a ZC game
09:04:15watusimoto(I think)
09:04:19raptorheh
09:05:08raptormaybe we need to separate zone rendering by more than just letters..
09:05:15raptorlike change the outline somehow
09:05:45watusimotojust what I was thinking
09:05:51watusimotoperhaps a heavy dashed line
09:07:21watusimotoso... I added mq = to my hgrc file, but hg still does not recongnize prune
09:07:53raptor'strip'
09:08:07raptornot 'prune'
09:08:31raptori remember my first usage of the technique of 'prune', it was in an old DOS application called XTree
09:08:55raptori loved it because I could destroy huge junks of data on the hard drive and have to reinstall the system
09:09:37kodai did that too but without xtree :p
09:10:32watusimotoxtree!
09:11:08watusimotostrip does something!
09:11:33bobdaduck has joined
09:12:55BFLogBot Commit: adb9eddad2e8 | Author: watusimoto | Message: Add "Goal" text effect when a goal is scored in soccer
09:13:15raptordoes that mean your repo is cleaned up now?
09:15:11watusimotoit worked!!
09:15:17watusimotomaaaybe
09:16:54raptorok
09:16:56raptorso
09:17:00raptori'm working with LuaW again...
09:17:23watusimotolucky you!
09:17:25raptorand just so I get this straight - our version has two major differences to upstream:
09:17:43raptor1. we use proxy objects (and have decided to always do so)
09:17:50watusimotoyes
09:17:55watusimotoabsolutely essential
09:18:03raptor2. we have a whole bunch of registerCLass logic..
09:18:07watusimotoyes
09:18:22watusimotonot essential but makes things cleaner
09:18:30watusimoto(and I suspect are not a problem)
09:19:54raptoroh
09:19:58raptorexcept
09:20:30raptoryou added your own caching
09:20:41raptorand then upstream took your idea of caching and added it
09:20:51raptorexcept I think we're still using your implementation
09:21:27watusimotoI think so too
09:21:43watusimotocaching has nothing to do with the registration
09:23:49raptorhuh
09:24:02raptorit looks like we implemented the upstream caching at one point (since 018a)
09:24:11raptorbut we had to revert because of a bot crash..
09:24:22watusimotoyes
09:24:24raptormaybe I should start from there..
09:24:31raptori don't remember the specific problem
09:26:24kodab has joined
09:26:24watusimotoI don't either... oh wait
09:26:36watusimotoI vaugely do remember
09:26:51watusimotowe were getting reused addresses back that broke the caching system
09:27:07watusimotothough we might have fixed that
09:28:04watusimotojust to clarify (if clarification is needed) registration just tells luaw what classes we have and what methods they offer
09:28:18watusimotoit is not used during the game, only during initialization
09:30:17koda Quit (Ping timeout: 245 seconds)
09:31:00watusimotobtw, I'll be heading out in about 30 mins and probably won't be back online before sunday
09:34:32watusimotoso if I remmber correclty, currently we don't get back the same proxy when we grab an object twice from a script
09:34:52watusimotoand what we want to do is to cache the proxies so that we do get back a consistent object
09:35:05watusimotoand, if I recall, it was this caching that went wrong
09:35:14watusimotoor something related to it
09:40:27raptorhmm
09:40:30raptorok
09:40:56raptorwell, see you Sunday - i'll probably hit some wall soon.. :)
09:42:09watusimoto:-)
09:42:38watusimotoI have to work until 2AM tonight
09:42:52watusimotoI started at 8:30 this morning
09:43:08raptoroh man
09:43:24watusimototonight won't be too too bad, as the setting is a dinner party
09:43:45watusimotobut I'm really tired and not really in the mood for this event
09:44:19raptorsee this more recent upstream commit (since we've toyed with LuaW): https://bitbucket.org/alexames/luawrapper/commits/82f54ba7da4a617b5c6f5bcab3d4cd1b46934eeb
09:44:38raptorbooo... evil events
09:45:48watusimotoI'm not positive that will fix the problem, as the resued memory addresses were for the same object type
09:45:58watusimotobut... maybe
09:46:36raptorthere have been a couple more fixes to the upstream caching since...
09:46:50raptorI'm thinking about doing as much as I can with upstream and see what happens
09:47:02watusimotothat would be my strategy as well
09:53:18raptorwell, i'm off for a bit - thanks for joggin my memory on the LuaW stuff..
09:54:27raptor has left
09:58:28watusimoto Quit (Quit: Leaving.)
10:35:42kodab Quit (Ping timeout: 256 seconds)
10:43:10raptor has joined
10:43:10ChanServ sets mode +o raptor
10:52:27Nothing_Much Quit (Ping timeout: 264 seconds)
11:05:27Nothing_Much has joined
11:06:05Nothing_Much Quit (Changing host)
11:06:05Nothing_Much has joined
13:09:28bobdaduck Quit (Remote host closed the connection)
13:24:58LordDVG has joined
13:43:47koda has joined
13:51:48raptoroh my goodness
13:52:08raptorI finally finished pulling in upstream changes for LuaW... and... no crashes
14:06:58bobdaduck has joined
14:17:22raptorok, test with old DnD on 018a for 2 min
14:17:39raptorRAM when to 165 MB
14:17:45raptor*went
14:17:50raptornow for the new test..
14:30:32raptorhmm... still inflates..
14:38:13BFLogBot Commit: e20c8828f30e | Author: buckyballreaction | Message: LuaWrapper changes: - Update LuaW to upstream 3c54015. This let me remove the workaround for the script crashes (commenting out luaW_hold) and hopefully removing the memory leak that was introduced - Revert 01414236b4ed for now, LuaW should handle the proxy now
14:53:22koda pokes raptor to see if he can test the experimental repo
14:53:28raptorhi koda
14:53:35raptori tested yesterday i think..
14:53:37kodaalso Y U NO LURK in #hedgewars :(
14:53:43raptoruhh
14:53:43koda waves
14:56:28Nothing_Much Quit (Read error: Connection reset by peer)
14:57:46koda kidnapped raptor and nobody noticed *evil laugh*
15:06:56LordDVG Quit (Remote host closed the connection)
15:08:46raptor has left
15:13:00Nothing_Much has joined
15:14:32Nothing_Much Quit (Changing host)
15:14:33Nothing_Much has joined
15:37:16raptor has joined
15:37:16ChanServ sets mode +o raptor
17:11:13fordcars has joined
17:19:24bobdaduck Quit (Remote host closed the connection)
18:25:14koda Quit (Quit: koda)
19:22:33raptorok kaen, do we have a good Lua memory leak testing script?
19:25:36fordcars Quit (Ping timeout: 250 seconds)
21:23:26bobdaduck has joined
21:59:00bobdaduckhard crash
21:59:11raptor:(
22:20:57bobdaduckraptor you thought you fixed the memory leak?
22:26:24raptoryeah, but I think I made everything worse...
22:26:35raptorlike I think there is still a leak, but not as bad
22:26:52raptorand now objects of a single type all act in unison
22:28:31bobdaducklol
22:28:50bobdaduckI kind of want to see this
22:31:51raptorit's not really in a state to show at the moment - i've been doing surgery
22:54:46raptorheading to bed. night!
22:54:49raptor Quit ()
22:55:11bobdaduck Quit (Remote host closed the connection)

Index Search ←Prev date Next date→

These logs were automatically created by BFLogBot on irc.freenode.net.