| *** tpb has joined #freeorion | 00:00 | |
| *** mithro has quit IRC | 07:16 | |
| *** STalKer-X has joined #freeorion | 08:23 | |
| *** Carmina has quit IRC | 08:39 | |
| *** tyreth has quit IRC | 08:39 | |
| *** theHive has joined #freeorion | 08:59 | |
| *** theHive has left #freeorion | 08:59 | |
| *** qk-gen has joined #freeorion | 09:19 | |
| qk-gen | Hi, little question, does anyone try compile freeorion with boost-1.34.0 ? | 09:20 |
|---|---|---|
| GeofftheMedio | I think someone tried with a release candidate | 09:20 |
| GeofftheMedio | and it fixed some boost signals related bugs | 09:21 |
| *** mithro has joined #freeorion | 09:21 | |
| qk-gen | now I try it and cannot compile it, some filesystem error like filesystem_error doesn`t have error only system_error and also doesn`t know 'security_error' is not a member of 'fs' and some other | 09:26 |
| qk-gen | I have stable boost not rc | 09:26 |
| GeofftheMedio | I guess they've changed how that works then | 09:33 |
| qk-gen | OK, I try edit it and post as patch to sf | 09:34 |
| GeofftheMedio | posting the patch on the forums would be preferable | 09:34 |
| qk-gen | OK, no problem....I find what change...It write, that they change errors to have all defined in POSIX cerror and only added some others | 09:37 |
| GeofftheMedio | k, thanks | 09:37 |
| qk-gen | OK, changed...it's only little few change, I post it on forum | 09:50 |
| GeofftheMedio | ok | 09:50 |
| GeofftheMedio | here's the post about the previous attempt: http://freeorion.org/forum/viewtopic.php?p=26719#p26719 | 09:56 |
| tpb | <http://ln-s.net/dyY> (at freeorion.org) | 09:56 |
| qk-gen | He describe same error, but I cannot open his patch to look, If he have same patch | 09:59 |
| GeofftheMedio | if yours is with the stable release, then it's probably more useful than his version | 10:00 |
| qk-gen | OK, when I get activation on forum I post it...is better open new topic or append to that? | 10:01 |
| GeofftheMedio | either way is fine | 10:01 |
| GeofftheMedio | most of those programming threads start out with someone posting a problem, and then they go off on some topic that isn't very well described by the thread title | 10:02 |
| GeofftheMedio | (they = the threads) | 10:03 |
| qk-gen | OK, I see that I repair only gigi, so I take same time also change orion :) | 10:16 |
| GeofftheMedio | that would be useful | 10:17 |
| qk-gen | another problem - it write me Checking for FSOUND_GetVersion() in C library fmod-3.75... no, but I have fmod 3.75 and this function is also defined in fmod.h DLL_API float F_API FSOUND_GetVersion(); | 10:44 |
| GeofftheMedio | did this stop the configuration? It can fail to find the file in one location and find it in another | 10:46 |
| GeofftheMedio | also, there's a bunch of stuff on the board about switching to open al in the last few days | 10:46 |
| GeofftheMedio | http://freeorion.org/forum/viewtopic.php?p=27141#p27141' | 10:47 |
| tpb | <http://ln-s.net/dyq> (at freeorion.org) | 10:47 |
| GeofftheMedio | http://freeorion.org/forum/viewtopic.php?p=27141#p27141 | 10:47 |
| tpb | <http://ln-s.net/dlT> (at freeorion.org) | 10:47 |
| GeofftheMedio | (without the trailing ') | 10:47 |
| qk-gen | OK, thanks it works | 10:56 |
| GeofftheMedio | which patch specifically did you use? | 10:57 |
| qk-gen | first | 11:09 |
| qk-gen | and some problem with code....I have quite restrict gcc and it write me that in Universe.h in serialize is in function Universe::serializable called method of ship, but ship is only forward declarated...is posible move this method to cpp file? | 11:10 |
| GeofftheMedio | I've been trying to fix this for the last few hours | 11:11 |
| GeofftheMedio | http://freeorion.org/forum/viewtopic.php?p=27317#p27317 | 11:12 |
| tpb | <http://ln-s.net/dyz> (at freeorion.org) | 11:12 |
| qk-gen | my recommend is move this method from header to source code ;) | 11:12 |
| GeofftheMedio | it's a templated function | 11:12 |
| GeofftheMedio | it has to be in the header | 11:12 |
| GeofftheMedio | make sure you get my post after that one too... with the forward decalre of UniverseObject | 11:14 |
| GeofftheMedio | any luck? | 11:20 |
| qk-gen | sorry, some work now I look on it | 11:22 |
| GeofftheMedio | k | 11:22 |
| qk-gen | hmm, this include looks like really hell :(..Empire/../universe/UniverseObject.h:183: error: 'Universe' has not been declared | 11:31 |
| GeofftheMedio | UniverseObject.h includes Universe.h | 11:32 |
| GeofftheMedio | stick class Universe; on line 35 of UniverseObject.h | 11:33 |
| qk-gen | I think problem is that Universe include Ship.h and Ship.h include UniverseObject :( | 11:33 |
| GeofftheMedio | yes, there is circular #includes, but everything has include guards so should be included only once per file | 11:34 |
| qk-gen | yes I know, but sequence is important | 11:35 |
| qk-gen | addind Universe doesn't help, because is used some nested items...I repair it and add it to whole path...important is declarate only what you really need | 11:38 |
| GeofftheMedio | I don't follow what you just said | 11:39 |
| qk-gen | in code is used Universe::somethink and forward declaration of Universe does't help.....so I think better is that this include and forward declaration hell is forward declarate only what is really needed in header like Universe::somethink and have constant in another header to better manage | 11:41 |
| qk-gen | Hmm, it not good idea | 11:47 |
| GeofftheMedio | I'm working on a solution | 11:47 |
| GeofftheMedio | just the parts of thet serialize function that use the Ship, Empire and EmpireManager classes are being put into a new function that isn't templated so can be put into the cpp | 11:48 |
| qk-gen | better what do you think about change serialize from templated function to normal function and Archive is abstract class | 11:48 |
| qk-gen | you are quicker :) | 11:48 |
| GeofftheMedio | I'm not going to muck around with how serialization works | 11:49 |
| qk-gen | hmm, it looks like it is boost serialization and it work only when function is templated...hmm, another idea is have universe variables like ALL_VISIBLE or s_encoding_empire in other class then universe containers | 11:55 |
| GeofftheMedio | http://freeorion.org/forum/viewtopic.php?p=27322#p27322 | 11:58 |
| tpb | <http://ln-s.net/dzG> (at freeorion.org) | 11:58 |
| GeofftheMedio | try that | 11:58 |
| qk-gen | OK, only one mistake void Universe::GetShipDesignsToSerialize in header must be only void GetShip... without Universe | 12:07 |
| qk-gen | and work (ehm can compile, other I don't test :) | 12:07 |
| GeofftheMedio | fixed in post | 12:08 |
| GeofftheMedio | does the compilation succeed after that change? | 12:10 |
| qk-gen | another little error util/Order.h:612 missing ; | 12:10 |
| qk-gen | and also one little change in boost I do it and when I have all I post it | 12:11 |
| GeofftheMedio | fixed the order thing | 12:11 |
| GeofftheMedio | ... why does MSVC let me compile these things? | 12:11 |
| qk-gen | Uhm...I don't know :) | 12:12 |
| qk-gen | and bad problem I have 64bit system and that assert fail BOOST_STATIC_ASSERT(sizeof(long) == 4); in serialize.cpp | 12:13 |
| GeofftheMedio | yeah, that's been a recurring complaint | 12:13 |
| GeofftheMedio | you can change it to == 8 for your local compile | 12:13 |
| GeofftheMedio | but it's not really a portable solution, since i think that would make it impossible for you to connect to 32 bit server | 12:14 |
| qk-gen | OK, but It's not nice solution, better is define own long type which have always 4 bytes on all architecture | 12:14 |
| GeofftheMedio | is that the boost change you referred to above, or is it something else? | 12:16 |
| qk-gen | no, boost change I mean change in filesystem error handle | 12:18 |
| GeofftheMedio | are you using boost 1.34.0 ? | 12:18 |
| GeofftheMedio | ... did we discuss this a few hours ago? | 12:19 |
| GeofftheMedio | yes, it was the first thing you asked about... | 12:19 |
| qk-gen | yes :) | 12:27 |
| GeofftheMedio | anyway, if you know how to set that up (portable fixed 4-byte integer) then I'd appreciate help doing it | 12:28 |
| qk-gen | hmm, problem with AL :) | 12:30 |
| qk-gen | client/human/HumanClientAppSoundOpenAL.cpp:87: undefined reference to `ov_read' | 12:30 |
| qk-gen | client/human/chmain.cpp:63: undefined reference to `HumanClientAppSoundFMOD::HumanClientAppSoundFMOD()' | 12:30 |
| GeofftheMedio | ov_read is part of vorbisfile | 12:34 |
| GeofftheMedio | if you're switching from fmod to openal, you'll need to exclude HumanClientAppsoundFmod from the build | 12:34 |
| GeofftheMedio | I guess the scons scripts don't check for vorbisfile ? | 12:35 |
| qk-gen | Hmm, my mistake, sorry....about long type...ideal is define for exaple mylong __int64 | 12:36 |
| qk-gen | __int64 is allways 6 bytes | 12:37 |
| GeofftheMedio | you mean 8 bytes? | 12:37 |
| GeofftheMedio | will that work nice with boost::serialization ? | 12:37 |
| qk-gen | yes 8 :)...I look | 12:37 |
| qk-gen | I exclude HumanClientAppsoundFmod problem is that client/human/chmain.cpp missing it | 12:38 |
| GeofftheMedio | you need to change chmain.cpp to replace the two occurances of Fmod with OpenAL | 12:38 |
| GeofftheMedio | the #include and the instantiaion of app | 12:39 |
| GeofftheMedio | lines 1 and 63 | 12:39 |
| qk-gen | OK, thanks....I think boost serializate work, because it is native types...in amanith it do that http://www.amanith.org/include2/gglobal.h | 12:40 |
| qk-gen | around half of deocument | 12:40 |
| GeofftheMedio | there's nothing about boost serialization in there, is there? | 12:42 |
| qk-gen | no, it how they do own types. about vorbis serialization is here http://www.boost.org/libs/serialization/doc/index.html | 12:42 |
| tpb | <http://ln-s.net/FvW> (at www.boost.org) | 12:43 |
| qk-gen | vorbis = boost :) | 12:43 |
| GeofftheMedio | isn't vorbis for ogg music files? | 12:43 |
| qk-gen | is my mistake I change words :) | 12:46 |
| GeofftheMedio | the scons script here | 12:49 |
| GeofftheMedio | http://freeorion.org/forum/viewtopic.php?p=27303#p27303 | 12:49 |
| tpb | <http://ln-s.net/dzf> (at freeorion.org) | 12:49 |
| GeofftheMedio | checks for vorbis file | 12:49 |
| qk-gen | I haven't problem with version, I have problem with function, now I finding where it is | 12:58 |
| qk-gen | I found it, it doesn't link with libvorbis | 13:00 |
| GeofftheMedio | why not? | 13:03 |
| qk-gen | in thread when I edit SConstruct it isn`t...but in your yes | 13:05 |
| qk-gen | otherwise why using scons and not autotools? | 13:07 |
| GeofftheMedio | you lost me... | 13:07 |
| qk-gen | when I edit first time SConstruct I read another thread on forum, when vorbis isn't in SConstruct | 13:08 |
| GeofftheMedio | sorry, I can't figure out what you're asking... start over from the beginning? | 13:09 |
| qk-gen | OK, when I switch from FMOD to OpenAL I change SConstruct and in forum thread where I read it isn`t write about adding vorbis check...and in last forum thread which you post it is and now it works | 13:11 |
| GeofftheMedio | ok... so is there a question? | 13:12 |
| GeofftheMedio | those links were from the same thread, with different people posting their scons scripts... one person's checked for vorbis, but the other's didn't | 13:13 |
| GeofftheMedio | I'd like to figure out one good script to use and commit it to SVN, but I don't use linux myself and scons doesn't work for me on WinXP, so I'm having a hard time doing it | 13:13 |
| qk-gen | last is good....but scons have another problem when i run scons install it do this | 13:15 |
| qk-gen | Install file: "freeoriond" as "$prefix/bin/freeoriond" | 13:15 |
| qk-gen | Install file: "freeorionca" as "$prefix/bin/freeorionca" | 13:15 |
| qk-gen | Install file: "freeorion" as "$prefix/bin/freeorion" | 13:15 |
| qk-gen | and install to directory $prefix !!!! | 13:15 |
| GeofftheMedio | I think you needed to set the value of prefix | 13:16 |
| GeofftheMedio | scons prefix=/bin | 13:17 |
| GeofftheMedio | or somesuch | 13:17 |
| GeofftheMedio | run scons -h and near the top of the list of parameters is prefix | 13:17 |
| qk-gen | but it isn't expected....bindir default is $prefix/bin and prefix default is /usr/local so bindir must be /usr/local/bin | 13:18 |
| GeofftheMedio | hmm... maybe the script is broken then? | 13:19 |
| qk-gen | maybe...I am not python specialist and don't know how substitute $prefix with prefix | 13:21 |
| GeofftheMedio | it seems odd that nobody's noticed this before... | 13:22 |
| qk-gen | maybe everyone change this directory...I try change only prefix what happen :) | 13:23 |
| qk-gen | still broken | 13:25 |
| qk-gen | I can change it to to allways $prefix/$bindir | 13:25 |
| qk-gen | Hmmm, I solve it :) on windows it isn't often that executables is on another place then data, so I add it only on posix system and add default posix variable | 13:30 |
| GeofftheMedio | mkay | 13:31 |
| qk-gen | Hmm, now I try play and complete instalation is broken, no data is installed and it doesn`t start...i looks like that noone install it with scons :) | 13:56 |
| qk-gen | I post boost patches http://freeorion.org/forum/viewtopic.php?f=9&t=1600&p=27327#p27327 | 14:04 |
| tpb | <http://ln-s.net/e0B> (at freeorion.org) | 14:04 |
| GeofftheMedio | I think you have to do scons intall after building | 14:08 |
| qk-gen | I do it and it doesn't work....I look at python script and there missing any data instalation | 14:12 |
| GeofftheMedio | you messed around with $prefix and other variables... maybe that broke something | 14:20 |
| GeofftheMedio | try running from in the data directory | 14:25 |
| GeofftheMedio | or wherever you checked out from svn into | 14:26 |
| GeofftheMedio | did you post your boost filesystem changes? | 15:24 |
| qk-gen | yes few post above is link | 15:24 |
| qk-gen | game can run from directory where is compiled, but it's everythink | 15:24 |
| GeofftheMedio | I don't know what you mean by "but it's everythink" | 15:25 |
| qk-gen | on unix system is convention, that binaries is installed somewhere, data somewhere and user profile like savegame to another directory, but this game couldn`t work this | 15:26 |
| qk-gen | so it run only when everythink is at same place | 15:26 |
| GeofftheMedio | fyi: everything is spelt with a g not a k | 15:27 |
| GeofftheMedio | I think it was intended that it work correctly, but either it's broken, or you broke it while playing around with the path stuff in SCons... | 15:28 |
| qk-gen | I think that program doesn`t have support for loading data from another directory | 15:34 |
| GeofftheMedio | FreeOrion? | 15:34 |
| GeofftheMedio | it does. you can set the data directory in options | 15:34 |
| qk-gen | data like fonts? | 15:35 |
| GeofftheMedio | art, game content (buildings, techs, specials), sounds | 15:36 |
| GeofftheMedio | not sure how fonts are treated | 15:36 |
| qk-gen | OK, I look how it exactly work ;) | 15:37 |
| qk-gen | GeofftheMedio: How can I try battle? I have fleet at same planet as enemy and no battle | 16:52 |
| GeofftheMedio | you have to build a combat ship | 16:52 |
| GeofftheMedio | go to the production screen of your home system, scroll down in the buildable items list and build a Mark I | 16:52 |
| GeofftheMedio | then move it to an enemy fleet | 16:52 |
| GeofftheMedio | but don't expect much... combat is rudimentary | 16:53 |
| qk-gen | OK, thanks...I try fight with scout and nothink happen, only I discovery bug http://freeorion.org/forum/viewtopic.php?f=9&t=1780 | 16:54 |
| tpb | <http://ln-s.net/e0v> (at freeorion.org) | 16:54 |
| GeofftheMedio | you should post bugs on sourceforge | 16:55 |
| qk-gen | I post it on forum | 16:57 |
| qk-gen | but other which i found i post to sf | 16:57 |
| GeofftheMedio | all on sf is preferable | 16:57 |
| qk-gen | another question...what is this code -> if (!fs::path::default_name_check()(*it)) { it seg faulted on it | 17:07 |
| qk-gen | in FileDlg.cpp line 524 | 17:07 |
| GeofftheMedio | it's part of GG, not FreeOrion itself | 17:09 |
| GeofftheMedio | I don't know what default_name_check is supposed to do | 17:09 |
| GeofftheMedio | my guess is that the transition to 1.34 is the cause | 17:10 |
| qk-gen | I know...so post on gigi sf page? it looks like some bad generated code and double braces | 17:10 |
| qk-gen | because I don't think that default_name_check return functor | 17:11 |
| GeofftheMedio | path.hpp | 17:11 |
| GeofftheMedio | static name_check default_name_check(); | 17:11 |
| GeofftheMedio | typedef bool (*name_check)( const std::string & name ); | 17:12 |
| GeofftheMedio | I'm looking at 1.33.1 | 17:13 |
| GeofftheMedio | so it might have changed in 1.34 | 17:13 |
| qk-gen | yes, you are absolutely right | 17:14 |
| qk-gen | now it is | 17:14 |
| qk-gen | static void default_name_check( name_check ) {} static name_check default_name_check() { return 0; } | 17:15 |
| qk-gen | it is depredecated and not to use | 17:15 |
| qk-gen | I look how it solve new version | 17:15 |
| GeofftheMedio | it's not really a bug, since GG is still using 1.33.1, but I'd add another post to your patch post about how to fix | 17:15 |
| qk-gen | OK | 17:16 |
| qk-gen | hmm I find alternative, it is native function | 17:29 |
| qk-gen | yes it help, now I can load game :) | 17:35 |
| qk-gen | I added new diff to forum | 17:35 |
| *** qk-gen has quit IRC | 18:50 | |
| *** tyreth has joined #freeorion | 19:05 | |
Generated by irclog2html.py 2.5 by Marius Gedminas - find it at mg.pov.lt!