#gemrb@irc.freenode.net logs for 8 Oct 2011 (GMT)

Archive Today Yesterday Tomorrow
GemRB homepage


[00:15:16] <-- Yoshimo has left IRC (Read error: Connection reset by peer)
[00:49:39] --> edheldil_ has joined #gemrb
[00:59:12] <-- lynxlynxlynx has left IRC (Remote host closed the connection)
[01:54:29] --> brad_a has joined #gemrb
[02:34:24] <-- edheldil_ has left IRC (Ping timeout: 248 seconds)
[02:38:32] <-- brad_a has left IRC (Quit: brad_a)
[02:40:10] --> joneirik has joined #gemrb
[03:39:10] <-- joneirik has left IRC (Remote host closed the connection)
[06:52:17] --> Yoshimo has joined #gemrb
[07:21:33] --> jschall has joined #gemrb
[07:46:15] <-- jschall has left IRC (Read error: Connection reset by peer)
[07:46:30] <-- Yoshimo has left IRC (Quit: Yoshimo)
[09:15:10] --> Yoshimo has joined #gemrb
[09:23:14] <-- Yoshimo has left IRC (Quit: Yoshimo)
[09:36:20] --> lynxlynxlynx has joined #gemrb
[09:36:20] <-- lynxlynxlynx has left IRC (Changing host)
[09:36:20] --> lynxlynxlynx has joined #gemrb
[09:36:20] --- ChanServ gives channel operator status to lynxlynxlynx
[09:45:54] <-- |Cable| has left IRC (Read error: Connection reset by peer)
[09:47:15] --> Yoshimo has joined #gemrb
[09:49:12] <-- Yoshimo has left IRC (Client Quit)
[10:32:12] --> Maighstir has joined #gemrb
[11:00:48] --> Yoshimo has joined #gemrb
[11:01:15] <Yoshimo> hi folks, i tried to build gemrb on windows , and it didnt work:
[11:01:18] <Yoshimo> C:\Users\Sam\Downloads\BaldursGate\GemRB_Source\gemrb\gemrb\core\System\Logging.cpp: In function 'void vprint(const char*, char*)':
[11:01:18] <Yoshimo> C:\Users\Sam\Downloads\BaldursGate\GemRB_Source\gemrb\gemrb\core\System\Logging.cpp:45:20: error: 'cprintf' was not declared in this scope
[11:03:50] <lynxlynxlynx> using what?
[11:05:06] <Yoshimo> cmake and Maighstirs batchfiles
[11:07:21] <Yoshimo> latest pulled diff was : Commit:bb69a5827122dd14b719ade806fef2a1b908df06
[11:07:21] <Yoshimo> * will hopefully fix building on mac os versions prior to 10.6
[11:10:11] <Yoshimo> anything else you need lynx?
[11:11:02] <lynxlynxlynx> so this is mingw?
[11:11:17] <Yoshimo> yes
[11:12:12] <lynxlynxlynx> then something serious is wrong
[11:12:27] <lynxlynxlynx> __MINGW32__ must be undefined to get you where you are
[11:12:58] <lynxlynxlynx> something is wrong on your end, since the build bot has no issues with mingw builds
[11:14:40] <Yoshimo> but shouldnt it fail very early then instead of failing at 66%?
[11:16:37] <fuzzie> not really, there's not much in there that fails like that
[11:16:50] <fuzzie> cmake will print the state of WIN32_USE_STDIO when you run it though..
[11:18:00] <Yoshimo> is it necessary to have that as "enabled"?
[11:18:50] <Yoshimo> -- WIN32_USE_STDIO: disabled , seems to be a new option
[11:20:04] <fuzzie> but the __MINGW32__ check is gone
[11:21:29] <fuzzie> and buildbot is explicitly messing with WIN32_USE_STDIO
[11:21:44] <fuzzie> so i imagine tomprince broke it in d26e8a20a56f99d4ba2b8e5e20048a9f943617ad.
[11:23:53] <fuzzie> which looks like it would be wrong anyway since it's trying to force NOCOLOR for all mingw builds
[11:24:01] <fuzzie> but i don't know enough to know what to check for mingw there instead
[11:27:12] <fuzzie> i guess it should be MINGW and the NOCOLOR bit should be first
[11:27:38] <Yoshimo> i have no idea either, we could wait for tom though
[11:28:01] <fuzzie> try editing CMakeLists.txt in the root and changing line 278 from MINGW32 to MINGW
[11:28:33] <lynxlynxlynx> oh, i had an old tree
[11:33:32] <Yoshimo> this is a quadcore and i think it is still slower at compiling gemrb than i exspect to
[11:37:57] <lynxlynxlynx> make -j5
[11:39:04] <Yoshimo> one more than available cores or how do you get that number?
[11:41:18] <lynxlynxlynx> empirically
[11:41:28] <lynxlynxlynx> it used to be 2n+1
[11:41:49] <lynxlynxlynx> depends on how much you want to stress it out
[11:42:05] <lynxlynxlynx> and sometimes buildsystems don't parallelise perfectly either
[11:42:44] <Yoshimo> fuzzie: with your cmakelists. change options get switched, nocolor and win32_use_stdio now read "yes" instead of "no" like before
[11:44:35] <Yoshimo> successfully built now
[11:45:31] <Yoshimo> btw why does gemrb need so many files in "override\gametype"?
[11:46:37] <lynxlynxlynx> because the original is a hack
[11:47:03] <Yoshimo> which we already noticed before, can you elaborate on that?
[11:47:45] <lynxlynxlynx> bad design shouldn't be copied if it can be avoided, so we dehardcoded stuff like projectiles
[11:48:26] <lynxlynxlynx> so instead of having to code them, we implemented the framework and the rest is done with data that resides in our override
[11:49:06] <Yoshimo> ah ok
[11:51:53] <fuzzie> re compile speed: you have a modern mingw?
[11:57:21] <Yoshimo> modern means?
[12:00:58] <Yoshimo> my av captured and quarantined mingw-get-inst-20110802.exe as suspicious
[12:02:19] <lynxlynxlynx> hehe
[12:03:46] <fuzzie> something from 2011 :)
[12:04:07] <Yoshimo> even better: somebody broke obfuscation on the skype binaries and released it together with some sourcecode, also suspicious and quarantined :D
[13:51:26] --> |Cable| has joined #gemrb
[13:59:21] <-- Drakkar has left IRC (Ping timeout: 255 seconds)
[14:00:34] --> Drakkar has joined #gemrb
[14:21:51] <-- gembot has left IRC (Ping timeout: 255 seconds)
[15:22:24] --> gembot has joined #gemrb
[15:29:25] --> brad_a has joined #gemrb
[16:47:43] <-- Yoshimo has left IRC (Quit: Yoshimo)
[17:28:49] --> Yoshimo has joined #gemrb
[18:31:03] <tomprince> MINGW32 should certainly be MINGW
[18:40:34] <lynxlynxlynx> yep
[19:57:45] <Yoshimo> why did you change this windows compile options to stdio? what is the use of that tom?
[19:58:08] <fuzzie> it was always stdio
[19:58:18] <fuzzie> was just hard-coded in the source before.
[19:58:24] <Yoshimo> ah
[20:34:04] <tomprince> and win32 with msvc was hard coded to use conio, which I couldn't (easiily) capture with my buildbot.
[20:34:28] <tomprince> Thus, when the test step was failing, (or succeeding) I couldn't see what was happening.
[20:38:49] <Yoshimo> thanks to fuzzie, i can built it again, which for me is all that matters , well almost, but my issues with the bigworldproject are not dealt with here^^
[20:40:44] <tomprince> Yoshimo: If you have a win32 machine, it would useful for you to test out the binaries i build on http://buildbot.gemrb.org/waterfall
[20:41:13] <tomprince> I don't run windows, so I don't have a good way to test them. (They are built in a remote vm)
[20:41:48] <Yoshimo> did the mingw32 problem get changed on the gittree?
[20:41:58] <tomprince> I set them up, partly for people who are on win32 don't need to setup up a dev environment to use gemrb, but I don't know if they work properly.
[20:42:48] <Yoshimo> this websites layout is a bit confusing
[20:43:50] <tomprince> Under nmake-msvc++6 and mingw32 there are [binary] links for each build.
[20:44:39] <tomprince> And http://buildbot.gemrb.org/binaries/GemRB-nmake-latest.zip and http://buildbot.gemrb.org/binaries/GemRB-win32-latest.zip should always point to the latest build.
[20:44:40] <Yoshimo> most recent files at the top i guess
[20:45:03] <tomprince> Yep.
[20:45:34] <CIA-26> GemRB: 03tom.prince * r01d0a023f60e 10gemrb/CMakeLists.txt:
[20:45:34] <CIA-26> GemRB: cmake: s/MINGW32/MINGW/
[20:45:34] <CIA-26> GemRB: Typo, cmake sets MINGW not MINGW32.
[20:45:37] <Yoshimo> why does it built some files with nmake-msvc and some with msvc and some with cmake and some with mingw32, shouldnt it built the source with all available tools?
[20:46:18] <tomprince> Also, if one or the other or both, work consistently, then we could add a directly link from the wiki.
[20:46:48] <tomprince> Each column represents one builder, that build in a different configuration
[20:47:34] <Yoshimo> so why are some columns partially empty?
[20:47:50] <tomprince> http://buildbot.gemrb.org/grid might make it more obvious what is going on.
[20:48:16] <tomprince> It only runs so many builds in parallel, so on the waterfall, they show up sequentially.
[20:51:03] <Yoshimo> can that system also do builts for lets say ios, android or mac osx?
[20:51:20] <tomprince> Yoshimo: So if you look at it now, you will see some builds have started, as you will see each builder triggerd in turn. autoupdating: http://buildbot.gemrb.org/waterfall?reload=60
[20:51:31] <tomprince> In theory, yes.
[20:51:47] <tomprince> I don't have access to mac.
[20:52:10] <tomprince> And I don't have the motivation to figure out how to set up an automated android build.
[20:53:01] <fuzzie> it is non-trivial.
[20:53:03] <Yoshimo> just curious , dont worry ;)
[20:53:56] <tomprince> If somebody set up a VM with all the necesary bits to do automated builds, I would be happy to run it.
[20:54:03] <fuzzie> i don't think you can
[20:54:41] <fuzzie> i mean, in the sense that it seems to need an alarming amount of manual intervention to keep it updated
[20:54:58] <fuzzie> i think iphone builds would be simpler to setup, even :)
[20:58:06] <Yoshimo> where does this buildbot run on, who owns the machine?
[20:59:02] <tomprince> On my machine. Actually, the master and front end is run from my desktop, and the builders are located on my server, which is located acrros the country.
[21:00:34] <tomprince> Which is silly, and why it was down most the summer (my desktop crashed while I was away, even though the server kept ticking, 618d uptime, currently)
[21:03:21] <Yoshimo> how about security updates then? :P
[21:03:49] <tomprince> No kerneal updates, but most of the rest of the stack I update periodically.
[21:05:10] <tomprince> I might reboot it moreo often, if I was closer, and could deal with any fallout from rebooting, but I am usually busy with other stuff when I am in town.
[21:06:01] <tomprince> speaking of osx/ios, if somebody had a machine they were willing to run a slave on, I would quite a happily hook it up.
[21:06:13] <tomprince> Maybe when my gf gets a mac. ;)
[21:06:29] <Yoshimo> hidden service: build_slave :P
[21:07:20] <Yoshimo> mhmm, how would a system like this work for our cellphone, tablet builts?
[21:09:35] <tomprince> Those aren't built on a phone/tablet, typically.
[21:10:52] <fuzzie> i'm sure you could build an iphone build with a free toolchain, it'd just be annoying work
[21:12:25] <fuzzie> Yoshimo: see http://buildbot.scummvm.org/builders for an idea of how many targets you can build on a single linux host.. :)
[21:13:04] <Yoshimo> stable-n64 ;), what is dingux? :P
[21:13:27] <fuzzie> linux for the dingoo. there are gemrb patches for it somewhere.
[21:13:42] <fuzzie> and i think master-ds is the craziest scummvm build :)
[21:13:44] <Yoshimo> never even heard of that system
[21:14:02] <Yoshimo> ds, very old videogame stuff, aint it?
[21:14:02] <fuzzie> the dingoo is a cheap (<$100 new) handheld console
[21:15:31] <lynxlynxlynx> gemrb runs on dingoo, but we haven't merged the downscaler code yet
[21:15:31] <fuzzie> and, well, the 3ds is rather new, but yes :)
[21:15:47] <fuzzie> brad_a's font code might help also
[21:38:56] --> yyz_ has joined #gemrb
[21:40:12] <-- yyz_ has left #gemrb
[22:05:32] <-- lynxlynxlynx has left IRC (Read error: Connection reset by peer)
[23:19:19] <-- brad_a has left IRC (Quit: brad_a)