#exult@irc.freenode.net logs for 11 Jun 2016 (GMT)

Archive Today Yesterday Tomorrow
Exult homepage


[01:40:12] <-- frob has left IRC (Ping timeout: 276 seconds)
[01:52:54] --> DominusExult has joined #exult
[01:52:54] --- ChanServ gives channel operator status to DominusExult
[01:56:21] --- Knexus is now known as ParuNexus
[01:56:30] <-- Dominus has left IRC (Ping timeout: 246 seconds)
[01:56:30] --- DominusExult is now known as Dominus
[04:39:04] --> frob has joined #exult
[04:39:19] <-- frob has left IRC (Remote host closed the connection)
[04:39:35] --> frob has joined #exult
[04:46:43] <-- frob has left IRC (Remote host closed the connection)
[06:26:43] --> frob has joined #exult
[07:02:10] <-- frob has left IRC (Remote host closed the connection)
[09:14:07] --> frob has joined #exult
[09:18:40] <-- frob has left IRC (Ping timeout: 260 seconds)
[10:19:44] <Dominus> hmm, Malignant's commit b80349f broke real MT32 support for some reason on OS X and I can't figure out why...
[10:43:24] <Dominus> I really can't see why that commit crashes coremidi in OS X. But I checked, the commit before worked...
[15:23:29] <Dominus> Marzo: any idea why the placement of miscoptions.shp in flx.in matters? https://github.com/exult/exult/issues/1
[15:24:31] <Dominus> when it is added/placed in the middle Coremidi crashes when using a real MT32 device. When placed last, everything works
[15:25:02] <Marzo> It ought to make a difference only if something, somewhere, is using a raw numeric constant instead of the symbolic constants from the header
[15:25:29] <Dominus> yeah, that's what I thought, too
[15:25:43] <Dominus> just couldn't find the culprit...
[15:26:49] <Dominus> seems mtgm.mid is referenced somewhere raw
[15:27:14] <Dominus> placing miscoptions.shp before crashes, after is fine
[16:54:58] --> Malignant_Manor has joined #exult
[16:55:50] <Malignant_Manor> Dominus: gdb or valgrind could get a better idea of what is going on.
[17:23:54] <Dominus> Gdb seems to go into system libs right away...
[17:24:26] <Dominus> Valgrind I don't have any idea about how to use
[17:25:00] <Dominus> Any idea where the midi file is referenced? I couldn't find it...
[17:43:14] <Malignant_Manor> audio\Midi.cc has EXULT_FLX_MTGM_MID
[17:47:06] <Dominus> Oh, I tried gdb with a non-debug build...
[17:47:22] <Dominus> Will try again...
[17:55:14] <Malignant_Manor> Dominus: is it possible that the compiler didn't update midi.o?
[17:56:21] <Malignant_Manor> That would be an easy cause of the bug.
[17:57:22] <Malignant_Manor> Exult should rebuild flx dependencies but maybe that file doesn't get rebuilt.
[18:07:07] <Dominus> I don't think that's possible, since I made clean and made again since you added that shape
[18:21:07] <Dominus> that is my bt http://pastebin.com/u06ifQm3
[18:29:20] <Dominus> wjp, marzo: any idea why what how?
[18:50:55] <Dominus> valgrind is as annoying as ever...
[19:01:14] <-- ParuNexus has left IRC (Ping timeout: 258 seconds)
[19:09:33] <Malignant_Manor> Does changing index = EXULT_FLX_MTGM_MID;
[19:09:48] <Malignant_Manor> to index = 27; produce the same result
[19:19:08] <Dominus> yes
[19:20:19] <Dominus> so it's probably like Colourless wrote, it's somehow triggering a memory corruption... odd
[19:44:45] --> ParuNexus has joined #exult
[20:19:48] <-- Malignant_Manor has left IRC (Quit: ChatZilla 0.9.92 [Firefox 47.0/20160604131506])
[23:50:06] --> DominusExult has joined #exult
[23:50:06] --- ChanServ gives channel operator status to DominusExult
[23:52:42] <-- Dominus has left IRC (Ping timeout: 276 seconds)
[23:52:42] --- DominusExult is now known as Dominus