*** tpb has joined #freeorion | 00:00 | |
-jordan.freenode.net- [freenode-info] channel flooding and no channel staff around to help? Please check with freenode support: http://freenode.net/faq.shtml#gettinghelp | 00:00 | |
*** mithro has quit IRC | 00:19 | |
kroddnoMobil | GeofftheMedio: I think i managed to link Ogre 1.7.0 and boost 1.42 | 00:22 |
---|---|---|
GeofftheMedio | woo | 00:22 |
kroddnoMobil | I'll test that now on my ati system | 00:22 |
kroddnoMobil | What was that crash related to Hazamonzo was reporting | 00:24 |
kroddnoMobil | that problem with autosaves | 00:25 |
kroddnoMobil | http://www.ntronium.com/Armada2526%20Home.html <- looks interesting | 00:27 |
tpb | Title: Armada 2526 Home (at www.ntronium.com) | 00:27 |
GeofftheMedio | sent you email | 00:28 |
kroddnoMobil | http://www.youtube.com/watch?v=UVYafoBvycE&feature=related | 00:34 |
tpb | <http://ln-s.net/5yr2> (at www.youtube.com) | 00:34 |
kroddnoMobil | GeofftheMedio: bad news. Still the same crash on my ATI notebook | 00:37 |
*** mithro has joined #freeorion | 00:38 | |
kroddnoMobil | the call that leads to the crash is inside GG, not inside ogre | 00:38 |
GeofftheMedio | can you relink to the details? | 00:38 |
kroddnoMobil | pardon? | 00:38 |
GeofftheMedio | post logs or error messages realted to the crash | 00:38 |
kroddnoMobil | I'll try to | 00:39 |
kroddnoMobil | The crash is now boost related... | 00:45 |
kroddnoMobil | strange ^^ | 00:45 |
kroddnoMobil | I have to check if it runs on my nvidia system before I post any errors | 00:46 |
kroddn | okay, seems to be my fault | 00:58 |
kroddn | there are boost 1.37 related errors inside the binary | 00:58 |
GeofftheMedio | never works the first time... | 00:59 |
kroddn | i hate boost ^^ | 01:01 |
GeofftheMedio | since you're packaging rather than programming, that's understandable | 01:01 |
kroddn | But i have to say something positive about ogre | 01:04 |
GeofftheMedio | really? | 01:05 |
GeofftheMedio | I'm shocked | 01:05 |
kroddn | with version 1.7 the support for building static versions has improved | 01:05 |
kroddn | hehe | 01:05 |
kroddn | It was easy to build ogre 1.7 static libs | 01:05 |
kroddn | no manual work needed | 01:05 |
kroddn | after that horror work I will try to release a build-enviroment for Linux-versions | 01:07 |
kroddn | based on a debian testing chroot | 01:07 |
kroddn | in case that i die it would be great if my legacy will be continued ^^^ | 01:08 |
kroddn | TATA | 01:08 |
kroddn | works! | 01:08 |
GeofftheMedio | an ati? | 01:08 |
kroddn | on nvidia now | 01:09 |
kroddn | not crashing any more | 01:09 |
kroddn | I'll go to my notebook in some minutes | 01:09 |
GeofftheMedio | hmm. I'm not too hopeful | 01:10 |
kroddn | me too | 01:11 |
kroddn | but i saw that tzlainze has changed that code that crashed before | 01:11 |
kroddn | damn, to early | 01:15 |
kroddn | crash | 01:15 |
kroddn | main() caught exception(std::exception): unregistered class | 01:16 |
GeofftheMedio | hmm... quite similar to some issues I was having earlier | 01:16 |
kroddn | ClientNetworking::NetworkingThread() : Networking thread will be terminated due to disconnect exception "End of file" | 01:16 |
GeofftheMedio | check out serializeuniverse.cpp | 01:17 |
GeofftheMedio | remove the ifdef around #include "SerializeUniverseExports.ipp" | 01:17 |
GeofftheMedio | is the networking error from the same execution? | 01:17 |
kroddn | yes | 01:19 |
kroddn | the "main() caught exception(std::exception): unregistered class" is on terminal | 01:19 |
kroddn | the other from logfile | 01:19 |
GeofftheMedio | ooh. makes sense | 01:19 |
GeofftheMedio | check out that #include | 01:19 |
kroddn | doesnt compile | 01:20 |
kroddn | the ifdef is for non-win32, so what should it help to remove the ifndef? | 01:20 |
GeofftheMedio | oh... hmm. thought it was ifdef not ifndef | 01:21 |
GeofftheMedio | ok, also look at serialize.ipp | 01:21 |
GeofftheMedio | there's an ifdef in that | 01:22 |
kroddn | but theres an ifdef in Serialize.ipp | 01:22 |
GeofftheMedio | yeah | 01:22 |
GeofftheMedio | try removing both? | 01:22 |
kroddn | you want to have the same as on win32 buld? | 01:22 |
GeofftheMedio | yes | 01:22 |
GeofftheMedio | because I had a very similar crash | 01:23 |
GeofftheMedio | and tzlaine did some rearranging of the code including adding those ifdefs so that it would work on win32 | 01:23 |
GeofftheMedio | but maybe your complier is having similar issues? | 01:23 |
kroddn | compiles | 01:24 |
GeofftheMedio | hmm. I think this might be a boost version issue | 01:26 |
GeofftheMedio | .Id reported something very similar in email when using boost 1.42 | 01:27 |
kroddn | same error | 01:28 |
kroddn | Did anybody get it to work then? | 01:28 |
GeofftheMedio | .Id was planning to build v0.3.14 with boost 1.40 | 01:28 |
GeofftheMedio | but I'm thinking the #ifdef cases in the code now aren't necessary for WIN32, but are necessary for boost 1.42 | 01:28 |
kroddn | thats what i hate about boost - there's no "stable" version | 01:29 |
GeofftheMedio | well, in principle we could just stick with the same version... but there are improvements regularly | 01:29 |
GeofftheMedio | and the various libraries within don't always have breaking changes for each boost release | 01:30 |
kroddn | the exceptions seems to be coming from Serialize.o | 01:31 |
kroddn | okay wait | 01:32 |
kroddn | I'll try a "clean" build | 01:32 |
kroddn | most boost errors in the past where to be fixed doing scons -c :-) | 01:32 |
kroddn | With 4-core rootserver its now not a problem to recompile ^^ | 01:33 |
GeofftheMedio | For me it was happening in GameStartMessage, which calls Serialize for the Universe classs. The problem seems to be that various UniverseObject derived classes weren't being exported properly, so Universe serialization would fail | 01:33 |
kroddn | okay, now that i built freeoriond again and clean, the whole binary has changed a lot | 01:34 |
GeofftheMedio | were you, and do you still, get the crash when trying to start a new game? | 01:35 |
kroddn | Yes, it crashes while showing "Creating universe..." | 01:36 |
kroddn | and still crashing | 01:36 |
GeofftheMedio | did you remove those ifdefs? | 01:37 |
kroddn | yes | 01:37 |
GeofftheMedio | did you rebuild freeorion as well? | 01:37 |
GeofftheMedio | (not just freeoriond) | 01:37 |
kroddn | yes | 01:38 |
GeofftheMedio | hmm. | 01:40 |
GeofftheMedio | maybe try putting back the #ifndef FREEORION_WIN32 in SerializeUniverse.cpp | 01:40 |
kroddn | but that does not work either | 01:41 |
GeofftheMedio | hmm | 01:41 |
GeofftheMedio | you've got all the serialize cpp files in the build? | 01:42 |
kroddn | hm | 01:42 |
kroddn | theres one export missing? | 01:42 |
GeofftheMedio | multiplayercommon, orderset, pathingengine, universe, empire | 01:42 |
kroddn | SerializeUniverseExports.ipp includes 6 files | 01:42 |
kroddn | but exports only 5 | 01:42 |
GeofftheMedio | yeah, it doesn't export shipdesign | 01:42 |
kroddn | ShipDesign ? | 01:42 |
kroddn | you fixed some shipdesign related issues in the last versions, didnt you? | 01:43 |
GeofftheMedio | in a recent commit yes. players weren't being sent designs they had created if there wasn't a ship in existance with that design, which meant stuff on the build queue using the design was crashing | 01:44 |
GeofftheMedio | shouldn't be relevant to this | 01:44 |
GeofftheMedio | can you comment out these lines in Universe::serialize (in SerializeUniverse.cpp): | 01:44 |
GeofftheMedio | & BOOST_SERIALIZATION_NVP(objects) | 01:44 |
GeofftheMedio | & BOOST_SERIALIZATION_NVP(empire_latest_known_objects) | 01:44 |
GeofftheMedio | and rebuild | 01:44 |
GeofftheMedio | ? | 01:44 |
kroddn | anything else? | 01:45 |
kroddn | export ShipDesign? | 01:45 |
GeofftheMedio | no | 01:45 |
GeofftheMedio | test one thing at a time | 01:45 |
GeofftheMedio | though feel free to try your idea first | 01:45 |
kroddn | i am compiling your suggestion now | 01:47 |
kroddn | My problem atm is that debugging a version that I build on the server takes a long time :-) | 01:48 |
kroddn | If I would have known that this much problems arise, I would have downloaded the whole chroot for building local | 01:48 |
kroddn | GeofftheMedio: game runs | 01:53 |
GeofftheMedio | w/o any objects in the universe, though | 01:53 |
kroddn | of course without any objs | 01:53 |
GeofftheMedio | right. so it seems to be the same problem I was having | 01:53 |
GeofftheMedio | feel free to uncomment and instead try exporting shipdesign, though I doubt it will help | 01:54 |
kroddn | no change | 02:02 |
kroddn | don't tell me to use boost 1.40 now... I will not sleep very will if you do so | 02:02 |
kroddn | its 4am here now, time to go to bed | 02:02 |
GeofftheMedio | no change means it still throws the exception? | 02:02 |
kroddn | yes | 02:03 |
GeofftheMedio | ok. I'll email tzlaine | 02:04 |
*** tomboy64 has quit IRC | 03:17 | |
*** mithro has quit IRC | 03:18 | |
*** mithro has joined #freeorion | 03:39 | |
*** kr0ddnoMobil has joined #freeorion | 03:40 | |
*** ChanServ sets mode: +o kr0ddnoMobil | 03:40 | |
*** kroddnoMobil has quit IRC | 03:44 | |
*** STalKer-X has joined #freeorion | 03:57 | |
*** STalKer-Y has quit IRC | 04:00 | |
*** bernardh has joined #freeorion | 04:46 | |
*** mithro has quit IRC | 07:12 | |
*** mithro has joined #freeorion | 07:31 | |
*** STalKer-X has quit IRC | 09:12 | |
*** STalKer-Y has joined #freeorion | 09:12 | |
*** bernardh has quit IRC | 09:58 | |
*** enig_ has quit IRC | 10:01 | |
*** bernardh has joined #freeorion | 11:42 | |
*** bernardh has quit IRC | 15:08 | |
*** tomboy64 has joined #freeorion | 15:27 | |
*** tomboy64 has quit IRC | 16:13 | |
*** enigmatic has joined #freeorion | 17:14 | |
*** tomboy64 has joined #freeorion | 18:04 | |
GeofftheMedio | kroddn: tzlaine says he's working on the serialization exception | 18:42 |
*** aleksil has quit IRC | 20:09 | |
*** tomboy64 has quit IRC | 21:42 | |
*** bernardh has joined #freeorion | 22:55 |
Generated by irclog2html.py 2.5 by Marius Gedminas - find it at mg.pov.lt!