#exult@irc.freenode.net logs for 26 Jul 2010 (GMT)

Archive Today Yesterday Tomorrow
Exult homepage


[00:01:02] <Dominus> same result
[00:01:52] <Marzo> Out of curiosity: does rolling back configure.ac to the previous version make it work?
[00:02:11] <Dominus> no, tested that right away :)
[00:02:48] <Marzo> Hm. So it is not that
[00:03:50] <Marzo> Just to be on the safe side, run `make clean distclean; ./autogen.sh` then re-run configure.sh with the options you normally use
[00:04:49] <Marzo> (I received a warning today about a different version of autoconf IIRC; something similar might be afoot here)
[00:06:11] <Dominus> hmm, now it worked and I think the problem was a patch that the macports mainteiner put into their autoconf install
[00:06:28] <Dominus> they replaced some stuff in types.am and functions.am
[00:06:46] <Marzo> Aha
[00:07:10] <Dominus> I mean *.m4
[00:07:17] <Marzo> (I thought so)
[00:07:37] <Marzo> When m4 files get modified, it is usually a good policy to re-run ./autogen.sh
[00:08:00] <-- Cahaan has left IRC ()
[00:08:07] <Marzo> (there is a cache for them, and it gets cleaned up by autogen.sh)
[00:08:20] <Marzo> The mismatch was probably causing the issue
[00:08:28] <Dominus> strange because I *did* run -/autogen.sh
[00:08:41] <Marzo> Hm
[00:08:50] <Marzo> So maybe it was the distclean?
[00:08:58] <Marzo> Now I am lost
[00:09:13] <Marzo> Well, it is working now so that is good :-)
[00:09:33] <Dominus> got to fix that ppc machine now so it works there, too :)
[00:13:50] <Dominus> there I can't get over that problem :(
[01:15:09] <-- Dominus has left IRC (Quit: Leaving.)
[03:04:06] --> julien_ has joined #exult
[03:04:44] <-- julien has left IRC (Ping timeout: 265 seconds)
[05:19:52] <-- Marzo has left IRC (Ping timeout: 240 seconds)
[07:58:40] --> Dominus has joined #exult
[07:58:40] --- ChanServ gives channel operator status to Dominus
[08:50:44] <-- jvlee has left IRC (Quit: jvlee)
[09:18:40] --> SiENcE has joined #exult
[09:21:40] <Dominus> hmm, on the mini something is borked. the configure script is missing the continous stuff about amdep that my intel machine has
[09:22:03] <Dominus> also libtool is not put into the exult folder by autogen
[09:22:23] <Dominus> further investigating :)
[10:44:34] <Dominus> grrrr
[10:45:11] <Dominus> still not working. I redid the whole built environment with macports because with that it *should* work
[10:45:21] <Dominus> but it doesn't
[10:50:38] <wjp> other projects' configure scripts don't have the same problem, I guess?
[10:51:07] <Dominus> hmm, I'd need to test. right now I have only exult on that machine
[10:51:23] <Dominus> any particular project you want me to test? scummvm?
[10:51:35] <wjp> scummvm doesn't use autoconf
[10:51:50] <wjp> you could try pentagram
[10:52:04] <wjp> no, wait
[10:52:20] <wjp> that likely won't trigger the depcomp stuff either
[10:53:48] <wjp> you could try libpng for example
[10:53:50] --> jvlee has joined #exult
[10:54:47] <Dominus> k, just a moment
[10:54:58] <wjp> e.g. this one, https://sourceforge.net/projects/libpng/files/01-libpng-master/1.4.3/libpng-1.4.3.tar.bz2/download
[10:55:32] <Dominus> i have the 1.4.4 around here
[10:56:05] <Dominus> oh well, trying that one :)
[11:00:43] <Dominus> yes, that one worked
[11:01:03] <wjp> what does config.log say about the dependency style?
[11:01:37] <Dominus> configure:3633: checking dependency style of gcc
[11:01:37] <Dominus> configure:3743: result: gcc3
[11:02:12] <wjp> gcc 3? exult was using /usr/bin/gcc-4.0, wasn't it?
[11:02:56] <Dominus> yes it does, but the dependency style gives gcc3
[11:03:02] <Dominus> on intel os x as well
[11:03:30] <Dominus> what I notice is that libtool isn't created in exulkt's folder
[11:03:39] <Dominus> it did get created in libpng
[11:03:46] <Dominus> and also on intel os x
[11:05:01] <wjp> that shouldn't be a problem. I assume ltmain.sh is created?
[11:05:52] <Dominus> yes
[11:26:17] <wjp> after changing the build env, did you try with a completely fresh exult svn checkout, just to be sure nothing from the old setup is interfering?
[11:27:05] <wjp> (and for future reference, the 'gcc3' style turns out to be gcc >= 3)
[11:29:55] <Dominus> trying with a fresh checkout now - didn't do that before
[11:34:07] <Dominus> grrr, so much for hoping...
[11:34:18] <Dominus> still the same error
[11:36:20] <wjp> what were the exact dependency style lines again?
[11:38:33] <Dominus> configure:5981: checking dependency style of /usr/bin/gcc-4.0
[11:38:33] <Dominus> configure:6091: result: none
[11:39:13] <wjp> I wonder if the /usr/bin/gcc-4.0 vs gcc is relevant
[11:39:21] <wjp> does /usr/bin/gcc-4.0 exist?
[11:39:37] <wjp> and is it a different compiler than gcc?
[11:40:40] <Dominus> it does exist and /usr/bin/gcc is a symlink to that
[11:42:30] <wjp> what does "sed -n 's/^#*\([a-zA-Z0-9]*\))$/\1/p' < ./depcomp" return? (It should return a list of about 14 names, one of which is gcc3)
[11:43:57] <Dominus> http://pastebin.com/vs4wBgAE
[11:45:25] <Dominus> on the intel machine (where configure works again, btw) the same output, except msvcmsys is missing
[11:55:08] <wjp> can you add "echo $depmode >> /tmp/depcomp.run" as line 3 of depcomp, and run configure again?
[11:55:24] <wjp> switching to printf-style debugging :-)
[11:56:23] <wjp> (and look at the contents of /tmp/depcomp.run afterwards)
[11:59:19] <Dominus> hmm, the file doesn't get created
[11:59:27] <wjp> ok, interesting
[12:00:15] <Dominus> i double checked whether I have no writing permissions in /tmp and echoed it in my user home as well in another try
[12:09:47] <wjp> argh
[12:10:11] <Dominus> (doing the same with the depcomp in the libpng folder makes the file appear with gcc3 in it)
[12:10:21] <Dominus> what is it?
[12:10:26] <wjp> Marzo broke configure yesterday
[12:11:15] <Dominus> he did it after all?
[12:11:24] <wjp> please try r6317
[12:13:28] <wjp> in fact, he broke it in exactly the same way as in 2008
[12:14:38] <wjp> just for the logs: it's very bad to put a macro that checks stuff (such as AC_PATH_XTRA) in a case/if in configure.ac
[12:17:34] <Dominus> yeah, that works
[12:17:50] <Dominus> grr and I had tested that on my intel machine yesterday
[12:18:20] <Dominus> but there it was probably a bad cache or some other circumstances made a play, so I never tested it on the ppc machine :)
[12:18:39] <Dominus> configure did run through on the ppc machine now
[12:21:14] <Dominus> he he and the best part of it was that it broke just on the day I finally set up my ppc mac for compiling :)
[12:21:37] <wjp> it was too much of a coincidence for me to even consider checking it :-(
[12:21:57] <Dominus> it also works nicely on my non-macports-built-environment (less overhead)
[12:22:56] <wjp> could you test the new version I just committed too?
[12:24:13] <Dominus> yes, one moment
[12:29:03] <Dominus> yes that works as well
[12:29:18] <Dominus> now actually building that thing on the ppc
[12:29:47] <Dominus> thanks wjp for fixing this.
[12:47:15] <Dominus> hmm, might have another problem on ppc but need to make clean, make first so I don't cry wolf...
[12:47:18] <Dominus> Making all in compiler
[12:47:18] <Dominus> /bin/sh ../../ylwrap ucparse.yy y.tab.c ucparse.cc y.tab.h ucparse.h y.output ucparse.output -- bison -y -d -v
[12:47:18] <Dominus> /Users/hansdampf/code/exult/usecode/compiler/ucparse.yy:274: type clash (`' `stmt') on default action
[12:47:18] <Dominus> /Users/hansdampf/code/exult/usecode/compiler/ucparse.yy:279: type clash (`' `stmt') on default action
[12:47:18] <Dominus> make[3]: *** [ucparse.cc] Error 1
[13:01:28] <wjp> hm, you'll have to complain to Marzo :-)
[13:01:43] <Dominus> he he, will do :)
[13:02:01] <Dominus> going to test it now on the macports environment, who knows... :)
[13:25:46] --> Marzo has joined #exult
[13:39:49] --> shazza has joined #exult
[14:05:54] <-- Kirben has left IRC ()
[14:35:54] --> Marzo1 has joined #exult
[14:35:54] <-- Marzo has left IRC (Disconnected by services)
[14:35:56] --- Marzo1 is now known as Marzo
[14:44:12] <Dominus> Marzo: I've run into another problem with compilng the usecode compiler on PPC (see the logs).
[14:44:31] <Dominus> it's not really important since I don't need to build that on PPC
[14:44:59] <Dominus> the PPC built is just for snapshots anyway and I can safely disable tools and usecode compiler.
[14:48:18] <Dominus> and that compiled (though with a strange timidity warning even though I had disabled timidity: /usr/libexec/gcc/powerpc-apple-darwin8/4.0.1/ld: warning empty table of contents: audio/midi_drivers/timidity/.libs/libtimidity.a (can't load from it))
[14:48:32] <Marzo> Dominus: what version of bison do you have on the ppc?
[14:48:52] <Dominus> 1.28
[14:49:06] <wjp> Dominus: that linker warning looks fairly harmless
[14:49:16] <Marzo> Can you try upgrading it to a less ancient version? :-)
[14:49:22] <Marzo> (mine is 2.4.1)
[14:50:11] <Dominus> Marzo: yeah, I'll give that a try, even though I remember bison being a bull to compile (could be wrong though).
[15:03:14] <Dominus> (stupid osx tiger doesn't automatically recognize binary files...)
[15:21:47] <-- Marzo has left IRC (Ping timeout: 264 seconds)
[15:24:59] --> Marzo has joined #exult
[15:25:28] <Dominus> hmm, is --disable-tools also disabling building the usecode compiler? I'm not complaining since it makes sense but I thought it wouldn't be like that :)
[15:25:47] <Dominus> testing right now compiling with nee bison installed
[15:32:04] <Dominus> compiling on that machine takes some time... PPC G4 1.25 GHz, 1GB RAM...
[15:34:25] <Dominus> Marzo: with bison 2.4.2 it got past that hurdle. thanks
[15:38:38] <Marzo> (I actually already knew it would; Kirben also had some issues in the past with old MinGW bison which were solved by an upgrade)
[15:51:55] <Dominus> soooo, finally.... got a working ppc built environment for exult. quickly tested and it works...
[15:52:16] <Dominus> later today or tomorrow I'll make a new universal snapshot
[16:22:42] <Dominus> wjp, marzo, colourless, regarding http://exult.sourceforge.net/forum/read.php?f=1&i=337514&t=337514, I mailed that guy wanting to put exult on their game magazine DVD (I know the magazine, it is an established old one)
[16:23:25] <Marzo> You did remind him that GPL requires making the source code available too, right?
[16:23:35] <-- shazza has left IRC (Ping timeout: 260 seconds)
[16:24:19] <Dominus> I wrote him about the grey copyright area concerning sfx and digital music, also that we consider snapshot much better these days.
[16:24:28] <Dominus> he still wants to publish it :)
[16:25:10] <Dominus> he sent me a document where I should sign as project representative that they are allowed to publish it on their DVD along with other stuff
[16:25:31] <Dominus> is that ok for you that I sign this? I think it is ok :)
[16:25:42] <Dominus> I'll remind him of the GPL, source thing
[16:26:21] <Dominus> I'll be gone for a while now, let me know here :)
[16:35:08] <wjp> I wouldn't sign anything
[16:35:33] <wjp> anything we could give permission for, we already have
[16:43:05] <-- SiENcE has left IRC (Quit: @all: cya)
[20:19:44] <Dominus> right