T2 IRC Log: 2010-10-16
This is the log as captured by an IRC bot in the channel. The statements are those of the individual people and might not neccessarily reflect the policy and legal rules as set forth by the T2 SDE Project.
--- Log opened Sat Oct 16 00:00:58 2010
11:18 < rxr> re
11:52 -!- arete74 [~arete74@151.56.83.90] has joined #t2
12:23 -!- orici [~quassel@2001:470:9a41:107:20c:29ff:fea4:8189] has joined #t2
12:45 < rogermason> hello.
12:47 < rogermason> I'm still interested in bulding root on my G5. I don't care if it is 32 or 64 bit. Any hints on how to proceed?
14:02 -!- Kladiin [~kladiin@host86-152-184-184.range86-152.btcentralplus.com] has joined #t2
14:04 < orici> patch
14:04 < orici> http://pastebin.com/pQBi1xBJ
15:26 < rxr> re
15:28 < rxr> rogermason: I do not know how many tricks (assembly and such) root is using
15:28 < CIA-27> rene * r37834 /trunk/package/database/postgresql/postgresql.desc: * updated postgresql (9.0.0 -> 9.0.1)
15:28 < rxr> I would try getting it built for 64bit
15:29 < rxr> we could, however, also add support to build ppc64 with 32bit userspace, like we do on sparc64
15:29 < rxr> or building selective packages for 32 vs 64 bit so you can bulid root and it's deps for 32bit, ...
15:29 < rxr> but that is more complrex
15:29 < rxr> complex
15:38 -!- arete74 [~arete74@151.56.83.90] has quit [Ping timeout: 245 seconds]
15:48 < rogermason> rxr: OK. I'll try building the 64 bit version. The first thing I guess is to find out why it segfaults. Any ideas?
15:51 < rxr> nope, but I will power on my g5 and give it a spin
15:51 < rxr> btw. what else failed for you, atk?
15:55 < Kladiin> hi, sorry for bothering you again, but I'm trying to build a generic minimal Linux T2 on a fresh T2 chroot based on the downloadable pre-built ISO, and I'm getting an error when building develop/ppl -
15:55 < Kladiin> !> checking for GNU M4 that supports accurate traces... configure: error: no ..
15:55 < Kladiin> !> GNU M4 1.4.5 or later is required; 1.4.11 is recommended
15:57 < rogermason> rxr: yes, atk failed. Here is what I did to get it built:
15:57 < rogermason> Rebuild perl. libffi gtkdoc(docbookx) gobject-introspection fails: aclocal: coul\
15:57 < rogermason> dn't open directory `//share/aclocal': No such file or directory Change line in atk.conf.
15:59 < Kladiin> (both the version of T2 I'm building and the T2 I'm building on are 8.0, btw)
16:13 < rxr> Kladiin: hm
16:15 < Kladiin> m4 --version says "m4 (GNU M4) 1.4.14" >.>
17:00 < rxr> Kladiin: no idea why that happens, think I have not seen that yet, can you look in the configure / config.log to see what those scripts are doing ?
17:01 < Kladiin> sure, in a minute - the machine's stuck doing something else atm
17:01 < rxr> rogermason: good news is that the root configure even at least has a linuxppc64gcc conditional :-)
17:02 < rxr> I'll try it over the next minutes on my g5 when a test build on x8664 finished
17:10 < rogermason> rxr: yes, I explicitly used the linuxppc64gcc Makefile in a manual build of root (which fails exactly the same way as when using the T2 machinery).
17:29 < Kladiin> in the configure.ac:
17:29 < Kladiin> if test x$build_c_interface = xyes \
17:29 < Kladiin> || test x$build_java_interface = xyes \
17:29 < Kladiin> || test x$build_ocaml_interface = xyes \
17:29 < Kladiin> || test x$build_some_prolog_interfaces = xyes
17:29 < Kladiin> then
17:29 < Kladiin> # Look for a recent enough version of GNU M4.
17:29 < Kladiin> AC_PROG_GNU_M4
17:29 < Kladiin> fi
17:30 < rxr> for me root on the g5 does not even find libx11, ...
17:31 < rxr> I guess you did some more tweaks to it already =
17:40 < rogermason> rxr: yes, the same thing happened for me. I edited root.conf around line 59 to add support for ppc64.
17:45 < Kladiin> definition of AC_PROG_GNU_M4 is: http://pastebin.ca/1964045
17:56 < CIA-27> rene * r37835 /trunk/package/base/sysfiles/etc_protocols.txt: * updated etc/protocols
18:03 < Kladiin> "m4 -F conftest.m4f &1" outputs /, and breaks the test that it's blank. it returns blank under GNU M4 1.4.5 at least.
18:15 < Kladiin> rxr: I'm not familiar with m4, so I couldn't tell you precisely which version of m4 this changed in, but that seems to be the crux of the problem, and it happened somewhere between 1.4.11 and 1.4.14 assuming that the comment given during the build process is correct
18:21 < Kladiin> I'm going to try setting up a 7.0 chroot and building 8.0 inside that as a workaround
18:23 < Kladiin> hmm. 7.0 rc2, as a 7.0 ISO is missing from the download site it seems.
18:23 < Kladiin> nope, 7.0 RC2s don't exist either.
18:24 < Kladiin> >.> sorry, there's one on ibiblio
18:30 < CIA-27> rene * r37836 /trunk/package/scientific/root/root.conf: * cleaned up root.conf, added a couple of more architecture configs
18:33 < Kladiin> ouch, I'm wrong
18:33 < Kladiin> /dev/null outputs a /. wtf.
18:33 < rxr> oops
18:33 < rxr> rogermason: segfaults here, too, needs to be really debugged :-(
18:40 < rogermason> rxr: i presume it is gcc that is segfaulting?
18:42 < Kladiin> there, that should work >.>
19:17 < rxr> rogermason: hm, may be the temporary root executable: rootcint_tmp
19:28 < CIA-27> rene * r37837 /trunk/package/multimedia/live/live.desc: * updated live (2010.10.06 -> 2010.10.15)
19:45 < rogermason> rxr: yes. I'm instrumenting it now. tricky because there are lots of ifdefs.
20:13 -!- Kladiin [~kladiin@host86-152-184-184.range86-152.btcentralplus.com] has quit [Remote host closed the connection]
20:15 < rogermason> I think the problem is in cint/cint/main/G__main.c. Calling G__main like this: if (G__main(argcc, argvv) < 0) { around line 4844 in rootcint.cxx seems to cause the crash. I tried instrumenting G__main.c but can't get anything out of it.
23:07 -!- Kladiin [~kladiin@host86-152-184-184.range86-152.btcentralplus.com] has joined #t2
23:12 < Kladiin> http://pastebin.ca/1964330
23:12 < Kladiin> any ideas?
--- Log closed Sun Oct 17 00:00:58 2010
11:18 < rxr> re
11:52 -!- arete74 [~arete74@151.56.83.90] has joined #t2
12:23 -!- orici [~quassel@2001:470:9a41:107:20c:29ff:fea4:8189] has joined #t2
12:45 < rogermason> hello.
12:47 < rogermason> I'm still interested in bulding root on my G5. I don't care if it is 32 or 64 bit. Any hints on how to proceed?
14:02 -!- Kladiin [~kladiin@host86-152-184-184.range86-152.btcentralplus.com] has joined #t2
14:04 < orici> patch
14:04 < orici> http://pastebin.com/pQBi1xBJ
15:26 < rxr> re
15:28 < rxr> rogermason: I do not know how many tricks (assembly and such) root is using
15:28 < CIA-27> rene * r37834 /trunk/package/database/postgresql/postgresql.desc: * updated postgresql (9.0.0 -> 9.0.1)
15:28 < rxr> I would try getting it built for 64bit
15:29 < rxr> we could, however, also add support to build ppc64 with 32bit userspace, like we do on sparc64
15:29 < rxr> or building selective packages for 32 vs 64 bit so you can bulid root and it's deps for 32bit, ...
15:29 < rxr> but that is more complrex
15:29 < rxr> complex
15:38 -!- arete74 [~arete74@151.56.83.90] has quit [Ping timeout: 245 seconds]
15:48 < rogermason> rxr: OK. I'll try building the 64 bit version. The first thing I guess is to find out why it segfaults. Any ideas?
15:51 < rxr> nope, but I will power on my g5 and give it a spin
15:51 < rxr> btw. what else failed for you, atk?
15:55 < Kladiin> hi, sorry for bothering you again, but I'm trying to build a generic minimal Linux T2 on a fresh T2 chroot based on the downloadable pre-built ISO, and I'm getting an error when building develop/ppl -
15:55 < Kladiin> !> checking for GNU M4 that supports accurate traces... configure: error: no ..
15:55 < Kladiin> !> GNU M4 1.4.5 or later is required; 1.4.11 is recommended
15:57 < rogermason> rxr: yes, atk failed. Here is what I did to get it built:
15:57 < rogermason> Rebuild perl. libffi gtkdoc(docbookx) gobject-introspection fails: aclocal: coul\
15:57 < rogermason> dn't open directory `//share/aclocal': No such file or directory Change line in atk.conf.
15:59 < Kladiin> (both the version of T2 I'm building and the T2 I'm building on are 8.0, btw)
16:13 < rxr> Kladiin: hm
16:15 < Kladiin> m4 --version says "m4 (GNU M4) 1.4.14" >.>
17:00 < rxr> Kladiin: no idea why that happens, think I have not seen that yet, can you look in the configure / config.log to see what those scripts are doing ?
17:01 < Kladiin> sure, in a minute - the machine's stuck doing something else atm
17:01 < rxr> rogermason: good news is that the root configure even at least has a linuxppc64gcc conditional :-)
17:02 < rxr> I'll try it over the next minutes on my g5 when a test build on x8664 finished
17:10 < rogermason> rxr: yes, I explicitly used the linuxppc64gcc Makefile in a manual build of root (which fails exactly the same way as when using the T2 machinery).
17:29 < Kladiin> in the configure.ac:
17:29 < Kladiin> if test x$build_c_interface = xyes \
17:29 < Kladiin> || test x$build_java_interface = xyes \
17:29 < Kladiin> || test x$build_ocaml_interface = xyes \
17:29 < Kladiin> || test x$build_some_prolog_interfaces = xyes
17:29 < Kladiin> then
17:29 < Kladiin> # Look for a recent enough version of GNU M4.
17:29 < Kladiin> AC_PROG_GNU_M4
17:29 < Kladiin> fi
17:30 < rxr> for me root on the g5 does not even find libx11, ...
17:31 < rxr> I guess you did some more tweaks to it already =
17:40 < rogermason> rxr: yes, the same thing happened for me. I edited root.conf around line 59 to add support for ppc64.
17:45 < Kladiin> definition of AC_PROG_GNU_M4 is: http://pastebin.ca/1964045
17:56 < CIA-27> rene * r37835 /trunk/package/base/sysfiles/etc_protocols.txt: * updated etc/protocols
18:03 < Kladiin> "m4 -F conftest.m4f &1" outputs /, and breaks the test that it's blank. it returns blank under GNU M4 1.4.5 at least.
18:15 < Kladiin> rxr: I'm not familiar with m4, so I couldn't tell you precisely which version of m4 this changed in, but that seems to be the crux of the problem, and it happened somewhere between 1.4.11 and 1.4.14 assuming that the comment given during the build process is correct
18:21 < Kladiin> I'm going to try setting up a 7.0 chroot and building 8.0 inside that as a workaround
18:23 < Kladiin> hmm. 7.0 rc2, as a 7.0 ISO is missing from the download site it seems.
18:23 < Kladiin> nope, 7.0 RC2s don't exist either.
18:24 < Kladiin> >.> sorry, there's one on ibiblio
18:30 < CIA-27> rene * r37836 /trunk/package/scientific/root/root.conf: * cleaned up root.conf, added a couple of more architecture configs
18:33 < Kladiin> ouch, I'm wrong
18:33 < Kladiin> /dev/null outputs a /. wtf.
18:33 < rxr> oops
18:33 < rxr> rogermason: segfaults here, too, needs to be really debugged :-(
18:40 < rogermason> rxr: i presume it is gcc that is segfaulting?
18:42 < Kladiin> there, that should work >.>
19:17 < rxr> rogermason: hm, may be the temporary root executable: rootcint_tmp
19:28 < CIA-27> rene * r37837 /trunk/package/multimedia/live/live.desc: * updated live (2010.10.06 -> 2010.10.15)
19:45 < rogermason> rxr: yes. I'm instrumenting it now. tricky because there are lots of ifdefs.
20:13 -!- Kladiin [~kladiin@host86-152-184-184.range86-152.btcentralplus.com] has quit [Remote host closed the connection]
20:15 < rogermason> I think the problem is in cint/cint/main/G__main.c. Calling G__main like this: if (G__main(argcc, argvv) < 0) { around line 4844 in rootcint.cxx seems to cause the crash. I tried instrumenting G__main.c but can't get anything out of it.
23:07 -!- Kladiin [~kladiin@host86-152-184-184.range86-152.btcentralplus.com] has joined #t2
23:12 < Kladiin> http://pastebin.ca/1964330
23:12 < Kladiin> any ideas?
--- Log closed Sun Oct 17 00:00:58 2010