Config
Log for #openttd.dev on 10th February 2014:
Times are UTC Toggle Colours
00:08:05  *** adf88 has quit IRC
00:08:18  *** adf88 has joined #openttd.dev
00:08:19  *** ChanServ sets mode: +v adf88
00:36:36  *** adf88 has quit IRC
00:36:48  *** adf88 has joined #openttd.dev
00:36:48  *** ChanServ sets mode: +v adf88
01:24:01  *** adf88 has quit IRC
01:24:13  *** adf88 has joined #openttd.dev
01:24:13  *** ChanServ sets mode: +v adf88
01:43:42  *** adf88 has quit IRC
01:43:54  *** adf88 has joined #openttd.dev
01:43:54  *** ChanServ sets mode: +v adf88
02:45:52  *** adf88 has quit IRC
02:46:04  *** adf88 has joined #openttd.dev
02:46:04  *** ChanServ sets mode: +v adf88
03:03:31  *** adf88 has quit IRC
07:46:20  *** Supercheese has quit IRC
08:55:41  *** Ristovski has joined #openttd.dev
09:14:53  *** adf88 has joined #openttd.dev
09:14:54  *** ChanServ sets mode: +v adf88
10:07:32  *** adf89 has joined #openttd.dev
10:07:32  *** adf88 has quit IRC
10:23:09  *** adf88 has joined #openttd.dev
10:23:09  *** ChanServ sets mode: +v adf88
10:23:09  *** adf89 has quit IRC
12:05:48  *** adf88 has quit IRC
12:05:48  *** Ristovski has quit IRC
12:05:48  *** tycoondemon has quit IRC
12:05:48  *** Belugas has quit IRC
12:05:48  *** dihedral has quit IRC
12:05:48  *** LordAro has quit IRC
12:05:48  *** Terkhen has quit IRC
12:05:48  *** Rubidium has quit IRC
12:06:11  *** Ristovski has joined #openttd.dev
12:07:03  *** LordAro has joined #openttd.dev
12:07:03  *** Terkhen has joined #openttd.dev
12:07:03  *** Rubidium has joined #openttd.dev
12:07:03  *** resistance.oftc.net sets mode: +vvv LordAro Terkhen Rubidium
12:08:27  *** Rubidium has quit IRC
12:08:27  *** Terkhen has quit IRC
12:08:27  *** LordAro has quit IRC
12:09:34  *** tycoondemon has joined #openttd.dev
12:09:57  *** Belugas has joined #openttd.dev
12:09:57  *** dihedral has joined #openttd.dev
12:09:57  *** graviton.oftc.net sets mode: +v Belugas
12:10:09  *** adf88 has joined #openttd.dev
12:10:09  *** oxygen.oftc.net sets mode: +v adf88
12:12:06  *** LordAro has joined #openttd.dev
12:12:06  *** Rubidium has joined #openttd.dev
12:12:06  *** Terkhen has joined #openttd.dev
12:12:06  *** weber.oftc.net sets mode: +vvv LordAro Rubidium Terkhen
12:48:51  *** Ristovski has quit IRC
12:50:08  *** Ristovski has joined #openttd.dev
12:59:56  *** Ristovski has quit IRC
13:28:09  *** Ristovski has joined #openttd.dev
16:37:42  *** adf88 has quit IRC
17:02:47  *** frosch123 has joined #openttd.dev
17:02:47  *** ChanServ sets mode: +v frosch123
17:13:54  *** DorpsGek changes topic to "OpenTTD Dev Channel || Latest SVN: r26328 || Logs: http://webster.openttdcoop.org/?channel=openttd.dev || Voice (talk-right) upon request via #openttd; make sure you are registered to NickServ before asking"
18:44:40  <fonsinchen> frosch123: the errors in 3082 warn you about a possible deadlock.
18:45:37  <frosch123> i am going to change it, so it checks whether the mutex is owned
18:45:57  <fonsinchen> If one thread first acquires lock A and then lock B and another one first lock B and then A, then they can deadlock if both acquire their first lock in short succession and then wait for the respective other thread on the second lock.
18:45:57  *** DorpsGek changes topic to "OpenTTD Dev Channel || Latest SVN: r26329 || Logs: http://webster.openttdcoop.org/?channel=openttd.dev || Voice (talk-right) upon request via #openttd; make sure you are registered to NickServ before asking"
18:46:13  <frosch123> hmm, oh, that kind of deadliock
18:46:21  <frosch123> are there even two mutexes involved?
18:47:51  <fonsinchen> Apparently there is one in gfx.cpp:1289 and one in sdl_v.cpp:844
18:48:34  <fonsinchen> or sdl_v.cpp:778 and gfx.cpp:1289 for that matter
18:48:51  <frosch123> oh, model progress
18:48:56  <frosch123> *modal
18:49:40  <fonsinchen> That's a very nice tool you've discovered there
18:49:45  * fonsinchen makes a note
18:50:13  <frosch123> well, it need patience
18:50:26  <frosch123> you cannot run ottd with -v sdl:no_threads obviously :p
18:50:34  <frosch123> so it becomes very hard to use
18:50:53  <frosch123> if i would use it longer i would have to add keyboard shortcuts for everything :p
18:50:56  * fonsinchen spent the entire last week manually debugging thread synchronization at work ...
18:55:03  *** DorpsGek changes topic to "OpenTTD Dev Channel || Latest SVN: r26330 || Logs: http://webster.openttdcoop.org/?channel=openttd.dev || Voice (talk-right) upon request via #openttd; make sure you are registered to NickServ before asking"
19:37:17  <fonsinchen> https://github.com/ulfhermann/openttd/commit/617abc888ea563e92ca6879d9b77ef562b026fca
19:37:39  <fonsinchen> Can this be the cause of FS#5898?
19:37:51  <fonsinchen> It seems I have to finish my windows install after all...
19:38:01  <planetmaker> interesting diff :)
19:38:34  <LordAro> is the constant not used anywhere?
19:38:46  <fonsinchen> No, it's just there for completeness
19:39:20  <fonsinchen> The interesting part is the other half of the patch. If you have > 64k link graph jobs, something interesting should happen
19:39:28  <fonsinchen> But what? no idea
19:39:59  <fonsinchen> And I can't imagine that guy routinely produces more than 64k link graph jobs.
19:42:57  <Rubidium> if the INVALID_LINK_GRAPH_JOB isn't used, then that bit won't do much... which is why I'd ponder why the decrease would improve something... except that LinkGraphJobID is uint16, so *if* for whatever reason those jobs are leaking or so, you'd get into trouble
19:43:56  <frosch123> having more than uint16 entries in the pool makes it impossible to address them
19:44:09  <frosch123> once created you willl access other ones
19:45:43  <fonsinchen> There's no offline way of installing VC++, is there?
19:45:47  <fonsinchen> I hate it ...
19:45:58  <frosch123> you can copy the compilefarm vm
19:46:03  <frosch123> only 5gb :p
19:46:33  <fonsinchen> Would be fine for me. Is it 64bit?
19:52:43  <frosch123> that savegame is not particulary big
19:52:48  <frosch123> it's a small map
19:52:50  <frosch123> only few lines
19:52:54  <fonsinchen> I know
19:53:18  <fonsinchen> The only way this could affect it is if it leaks link graph jobs in a nasty way
19:53:44  <fonsinchen> There's no obvious way to make it do that.
19:57:46  <Rubidium> the same happens with LinkGraph; 16 bit variable, 24 bit pools size
20:03:56  <Rubidium> http://rbijker.net/openttd/max_size_validation.diff
20:06:34  <fonsinchen> Yes, good idea.
20:07:02  <fonsinchen> I'll only fix the typo then.
20:07:32  <Rubidium> pff, just merge this with it ;)
20:13:09  *** DorpsGek changes topic to "OpenTTD Dev Channel || Latest SVN: r26331 || Logs: http://webster.openttdcoop.org/?channel=openttd.dev || Voice (talk-right) upon request via #openttd; make sure you are registered to NickServ before asking"
21:53:36  *** Ristovski has quit IRC
22:38:48  *** frosch123 has quit IRC

Powered by YARRSTE version: svn-trunk