--- Log opened Mon Aug 09 00:00:49 2010
01:50 < rconstruct> atk fails at stage 5
01:50 < rconstruct> strange
01:50 < rconstruct> configure.in:41: warning: AC_INIT: not a literal: http://bugzilla.gnome.org/enter_bug.cgi?product=atk
02:02 -!- soopurman [~L7-loser@75-1-216-170.lightspeed.snantx.sbcglobal.net] has joined #t2
02:19 -!- soopurman [~L7-loser@75-1-216-170.lightspeed.snantx.sbcglobal.net] has quit [Ping timeout: 276 seconds]
02:57 < rconstruct> gawk: /cache hit/ { hit=$4-; }
02:57 < rconstruct> syntax error
02:58 < rconstruct> minor typo though
03:31 < rconstruct> hmm, wtf, orbit2 being built at stage 0?
03:31 < rconstruct> am i missing anything?
03:32 < rconstruct> lol
03:34 -!- old_gaffer [~quassel@211.238.240.87] has joined #t2
05:51 -!- rconstruct [~cgwork@bl17-185-231.dsl.telepac.pt] has quit [Quit: leaving]
09:01 -!- sparc-kly [~mubex@adsl-64-237-248-6.prtc.net] has quit [Ping timeout: 264 seconds]
09:13 -!- sparc-kly [~mubex@adsl-72-50-28-199.prtc.net] has joined #t2
11:10 <@rxr> re
11:15 < rogermason> Hello.
11:16 < rogermason> I have a question about the dependency cache for a package. When and how is it generated?
11:22 <@rxr> when building a package it is create automatically from other package's files the package used during the build and written to var/adm/cache/
11:40 < rogermason> Thanks. I have a follow-up question. Yesterday i made a patch to add Grass GIS to T2 (just posted on ML). I installed grass using the desc file in the patch and Build-Target ...... -job 5-grass. There were dpendencies on fftw, tcl and tk, which I had to add to pkgsel, and build using Build-Target before grass would install. Does -job ignore dependencies?
11:42 <@rxr> jobs currently only uses this static number for sorting in the priority
11:42 <@rxr> we want to eventually remove this and also use the deps graph for Target scheduling, too
11:42 <@rxr> currently the deps are mostly used for end-user emerge
11:46 < rogermason> OK. Thanks. Sounds like I'm doing the right thing.
11:55 < rogermason> I just posted a replacement patch for cgal. I fixed the license and changed the maintainer to me. Hope that is OK.
12:01 <@rxr> cgal does not download here
12:05 < rogermason> That is b/c it is required to agree to the license by selection a button on the cgal website. I have no idea how to handle that.
12:07 <@rxr> for one thare are ways to annocate the Url as manual
12:07 <@rxr> maybe it was - prefix, ...
12:07 <@rxr> but this is GPL, free distributable, and it comes out of their FTP server anyway, ... so just using the FTP server URI should be sufficient
12:10 < rogermason> I'll look at it again.
12:10 <@rxr> I do right now
12:10 <@rxr> but for some reason the ftp URI does not work with curl as invoken by the T2 scripts, will look at that, too
12:14 <@rxr> ah, because I used the wrong URL :-) need to go to that "forge", ...
12:22 -!- Baldzius [~aldas@87-198-133-94.static.ptr.magnet.ie] has joined #t2
12:22 < Baldzius> moin
12:22 < CIA-46> rene * r37118 /trunk/package/scientific/cgal/ (. cgal.cache cgal.desc):
12:22 < CIA-46> Roger Mason :
12:22 < CIA-46> * added cgal (3.6.1) - Computational Geometry Algorithms Library
12:26 < Baldzius> so my laptop's wireless got working out of box with 2.6.35, yay :)
12:26 <@rxr> congrats
12:27 < Baldzius> was waiting ages for it to happen
12:27 <@rxr> looking forward to even more t2 use then :-) !!!
12:27 <@rxr> which chip?
12:27 < Baldzius> yeah, used for whole saturday
12:27 < Baldzius> have some changes for 8.0
12:27 < Baldzius> broadcom lower power
12:27 < Baldzius> 4312
12:27 < Baldzius> it was suppose to work with 2.6.32
12:27 < Baldzius> but somehow it didn't for me :/
12:28 <@rxr> I have my problems with just the atheros one alone
12:28 <@rxr> my netbook (wireless) works better as Hackintosh under OS X than with this ath9k (or so) driver
12:28 <@rxr> stupid driver oopses every now and then and looses the AP connection
12:28 <@rxr> needs and unload / reload cycle or reboot then, ...
12:29 <@rxr> (under OSX is works well, ...)
12:33 < CIA-46> aldas * r37119 /trunk/package/office/openoffice-org/ (configure.patch lbnames.patch):
12:33 < CIA-46> Urs Pfister :
12:33 < CIA-46> * fixed openoffice-org compilation with gcc-4.5.1 and bdb-4.8.x
12:33 < Baldzius> btw, found out one very stupid regression over weekend
12:34 < Baldzius> wonder why nobody else complained
12:34 <@rxr> which regression?
12:34 < Baldzius> it's with intel graphics
12:34 < Baldzius> black screen on boot, fb related
12:34 <@rxr> oh go away with those X and dri drivers
12:34 < Baldzius> +CONFIG_FB_I810 2010-08-08 2.6.34 # collides with Intel's KMS leading to black screen
12:34 <@rxr> I need to start my micro kernel with sane GFX sub-system
12:34 < Baldzius> works fine then
12:35 < Baldzius> I spoted this myself long time ago
12:35 <@rxr> did you commit already ?
12:35 < Baldzius> but it was only occasional
12:35 <@rxr> with my Poulsbo Netbook I have unaccelerated graphics in Linux and OS X anyway
12:36 < Baldzius> but with >= 2.6.34 it was constant
12:36 <@rxr> and the Windows 7 I rarely boot driver has screen flickering artifacts
12:36 < Baldzius> I will commit later
12:36 <@rxr> I wonder how Intel get's away with the mess that is the GMA500
12:36 < Baldzius> have some other patches
12:36 < Baldzius> overall spent 3-4 days trying to get t2 compiled on Ubuntu
12:36 < Baldzius> when hit 2 years old fecking bug still not resolved
12:36 < Baldzius> fecking Ubuntu
12:37 <@rxr> damn, vtk needs time to built even on an X6
12:37 <@rxr> == 12:28:41 =[5]=> Building scientific/vtk [5.6.0 9.0-trunk].
12:37 <@rxr> == 08/09/10 12:37:02 =[5]=> Finished building package vtk.
12:37 < Baldzius> damn , I lost so much time
12:37 < Baldzius> but it was good experience though :)
12:37 <@rxr> I can imagnie, and I love patches :-)
12:38 < Baldzius> tried wicd on T2
12:38 < Baldzius> this little thing rocks!
12:38 <@rxr> wicd ?
12:38 < Baldzius> network manager
12:38 < CIA-46> rene * r37120 /trunk/package/scientific/vtk/ (. vtk.cache vtk.desc):
12:38 < CIA-46> Roger Mason :
12:38 < CIA-46> * added vtk (5.6.0) - The Visualization Tool Kit
12:38 < Baldzius> will commit some improvements later
12:40 <@rxr> looking forward to them :-)!
12:40 * rxr gotta go lucnh
12:46 < CIA-46> aldas * r37121 /trunk/package/java/ (3 files in 3 dirs):
12:46 < CIA-46> Urs Pfister :
12:46 < CIA-46> * updated sun-jdk-160 (6u18 -> 6u20)
12:46 < CIA-46> * updated sun-jre-160 (6u18 -> 6u20)
13:04 < rogermason> What's this CIA stuff? Whey do they care?
13:05 < CIA-46> aldas * r37122 /trunk/package/develop/urwid/ (. urwid.cache urwid.desc): * added urwid (0.9.9.1) - A console user interface library for Python
13:05 < koan> rogermason: that's a bot, he gives info about repository iupdates
13:05 < CIA-46> aldas * r37123 /trunk/package/gnome2/libbonoboui/libbonoboui.cache: * removed gtk1.x dependency from libbonoboui
13:05 < rogermason> Ah! I was worried for a moment.
13:07 < CIA-46> aldas * r37124 /trunk/package/network/wicd/ (wicd.cache wicd.desc): * wicd needs some runtime packages, like dbus-python, pygobject and urwid
13:09 < CIA-46> aldas * r37125 /trunk/package/base/linux26/disable-broken.lst: * fix regression, disable FB_I810 it colides with Intel's KMS which is needed for xorg-server
13:13 < Baldzius> rogermason: it's not a _real_ CIA :)
13:14 < rogermason> Phew!
13:22 <@rxr> re
13:22 <@rxr> http://cia.vc/
13:23 <@rxr> http://cia.vc/stats/project/t2
13:33 <@rxr> oh, that vtk build time I posted earlier wasn't froman X6, was from a dual-socket Xeon, ... 8-way
13:33 <@rxr> with sucking, read poor VT performance :-(
13:33 < rogermason> OK. Mystery solved. Thanks. :-)
13:33 <@rxr> rogermason: we already had "proj4" as just "proj"
13:33 <@rxr> http://t2-project.org/packages/proj.html
13:34 <@rxr> for http://t2-project.org/packages/mapnik.html
13:34 < rogermason> Sorry. I should have checked more carefully. I've only ever seen it referenced as proj4.
13:35 <@rxr> therefore we got peer review :-)
13:35 <@rxr> I noticed mostly because I got shared files on my test build
13:35 <@rxr> Found shared files with other packages:
13:35 <@rxr> usr/bin/cs2cs: proj proj4
13:35 <@rxr> usr/bin/geod: proj proj4
13:36 <@rxr> ...
13:36 <@rxr> "shared files" is a T2 term for files already installed by another package
13:36 <@rxr> -- unrelated google feeding:
13:36 <@rxr> http://t2-project.org/packages/sdcc.html
13:36 < koan> why is the generic target installing dietlibc as part of the 00-minimal package set?
13:37 < Baldzius> I _hate_ shared files, so always ignore them
13:37 <@rxr> Baldzius: I always fix them :_)
13:37 <@rxr> koan: because we link initrd stuff statically for size optimizatoin
13:37 <@rxr> so udev, pdksh, module-init-tools are all statically dietlibc linked in T2 by default
13:37 < koan> rxr: thanks, makes sense
13:38 < rogermason> rxr: Didn't you get a message about the package being duplicated in the tree? That's how I found gdal was already present yesterday.
13:38 <@rxr> koan: of course one can customize that as needed in a target, ...
13:38 <@rxr> rogermason: you only get that message if you have the other package installed
13:38 < koan> rxr, well I'm thinking about creating a netbook target, I started by copying the generic target and now looking at what to add and remove
13:39 < rogermason> rxr: OK.
13:39 <@rxr> some more google teaching:
13:39 <@rxr> http://www.t2-project.org/packages/tcc.html
13:41 < CIA-46> rene * r37126 /trunk/package/scientific/grass/ (. grass.cache grass.desc):
13:41 < CIA-46> Roger Mason :
13:41 < CIA-46> * added grass (6.4.0RC6) - Geographic Resources Analysis Support System
13:42 < CIA-46> aldas * r37127 /trunk/package/base/ (2 files in 2 dirs): * /etc/modprobe.conf is deprecated and will we droped soon , all configurations should go to /etc/modprobe.d/
13:44 <@rxr> http://t2-project.org/packages/grass.html
13:48 < koan> compiling 32-bit x86 on x86_64 is cross-build, right?
13:53 <@rxr> well, yes, and no
13:54 <@rxr> while scriptly it may be, your kernel can certainly execute the 32bit binaries, you you do not need the cross compile option in t2
13:54 <@rxr> t2 does a clean, cross compiling boot strap in either case
13:54 <@rxr> the option just disables the native build phase
13:54 <@rxr> which you do not need to disable as your kernel can run the resulting executales
13:55 <@rxr> with the native build phase you can build all packages, not only the 50% we setup and patched to actually cross compile
13:55 < koan> so for building a 32-bit target on a 64-bit machine, I should not enable "This is a cross-build between..."
13:56 <@rxr> nope
13:56 <@rxr> less fragile build then, with more packages to choose from
14:06 < rogermason> Is there a way to get a new user added and their home dir set up during the install?
14:10 <@rxr> you could add a custom stne module that does just this sigle thing
14:10 <@rxr> e.g. to the 99-final package we we added for such custom modes
14:22 < CIA-46> aldas * r37128 /trunk/package/network/net-tools/net-tools.conf: * force off config.h rerun, we supply preconfigured config files anyway. fixes build on Ubuntu
15:06 < CIA-46> aldas * r37129 /trunk/package/network/bacula/bacula.desc: * updated bacula (5.0.2 -> 5.0.3)
15:06 < CIA-46> aldas * r37130 /trunk/package/filesystem/ntfs-3g/ntfs-3g.desc: * updated ntfs-3g (2010.5.16 -> 2010.8.8)
15:12 < Baldzius> http://www.freebsd.org/cgi/cvsweb.cgi/ports/graphics/exact-image/
15:12 < Baldzius> didn't know there's freebsd port
15:12 <@rxr> yeah, ... :-)
15:13 <@rxr> even gentoo
15:13 <@rxr> sorry, I'm currently a bit busy, ...
15:13 <@rxr> so IRC respose is sluggish :-)
15:13 <@rxr> http://www.exactcode.de/site/open_source/exactimage/releases/
15:14 <@rxr> ^- Debian, Ubuntu, too
15:14 <@rxr> it even was recently slashdotted
15:15 <@rxr> http://linux.slashdot.org/story/10/07/22/1852234/Open-Source-OCR-That-Makes-Searchable-PDFs?from=rss
15:15 <@rxr> thanks god that was after the "high traffic" to ext. server relocation :-) !!!
15:16 < Baldzius> :)
15:17 <@rxr> I just hack to make a tiny function "portable"
15:17 <@rxr> I got 3 (!!!) #ifdef WIN32 in there
15:17 <@rxr> freaking OS that is
15:17 <@rxr> and no, it's not T2 related :-)
15:20 <@rxr> just happen to debug alien code
15:20 <@rxr> and one where one can not set breakpoints in 99% of the code due to the alien way it is comiled together, sigh!
16:09 < CIA-46> aldas * r37131 /branches/8.0/package/develop/libcoyotl/ (gcc43.patch libcoyotl.cache):
16:09 < CIA-46> * merged r37114 from trunk to branches/8.0:
16:09 < CIA-46> fixed libcoyotl for gcc >= 4.3 (or so)
16:09 < CIA-46> aldas * r37132 /branches/8.0/package/develop/libevocosm/ (gcc43.patch.libevocosm libevocosm.cache):
16:09 < CIA-46> * merged r37115 from trunk to branches/8.0:
16:09 < CIA-46> fixed libevocosm for gcc >= 4.3
16:42 < rogermason> rxr: 'you could add a custom stne module'. OK. I'll look into that. Thanks.
16:49 < CIA-46> aldas * r37133 /branches/8.0/package/develop/urwid/ (. urwid.cache urwid.cache urwid.desc urwid.desc):
16:49 < CIA-46> * merged r37122 from trunk to branches/8.0:
16:49 < CIA-46> added urwid (0.9.9.1) - A console user interface library for Python
16:49 < CIA-46> aldas * r37134 /branches/8.0/package/gnome2/libbonoboui/libbonoboui.cache:
16:49 < CIA-46> * merged r37123 from trunk to branches/8.0:
16:49 < CIA-46> removed gtk1.x dependency from libbonoboui
16:51 < CIA-46> aldas * r37135 /branches/8.0/package/network/wicd/ (wicd.cache wicd.desc):
16:51 < CIA-46> * merged r37124 from trunk to branches/8.0:
16:51 < CIA-46> wicd needs some runtime packages, like dbus-python, pygobject and urwid
16:51 < CIA-46> aldas * r37136 /branches/8.0/package/base/linux26/disable-broken.lst:
16:51 < CIA-46> * merged r37125 from trunk to branches/8.0:
16:51 < CIA-46> fix regression, disable FB_I810 it colides with Intel's KMS which is needed for xorg-server
17:03 -!- rconstruct [~cgwork@bl17-178-150.dsl.telepac.pt] has joined #t2
17:03 < rconstruct> hi
17:05 < Baldzius> hey rconstruct
17:05 < Baldzius> welcome back
17:06 < Baldzius> not sure when I saw you last time :)
17:06 < rconstruct> lol
17:06 < rconstruct> some time, but i'm always around
17:07 < rconstruct> i'm too lazy to use other distros
17:07 < rconstruct> tried ubuntu on a workstation couple of weeks ago
17:07 < rconstruct> and... well...
17:07 < rconstruct> ok... i suppose there's a target audience for that, just not me
17:09 < Baldzius> :)
17:09 < rconstruct> i noticed some changes though
17:09 < Baldzius> yeah, I need to move myself from Ubuntu back to T2 :)
17:09 < rconstruct> well, i have ubuntu in a laptop
17:09 < rconstruct> that's sort of ok, i don't wnat to fiddle with all that wifi crap and such
17:09 < rconstruct> but not on a workstation
17:09 < rconstruct> i don't want gnome
17:10 < rconstruct> evolution
17:10 < rconstruct> kde
17:10 < rconstruct> or any of that crap
17:10 < rconstruct> i noticed orbit2 is now built in stage 0 as well (?!)
17:10 < rconstruct> my last target was a strange breed between 6.0 and trunk/7.0 with some custom pkgs thrown to the mix
17:11 < rconstruct> i've noticed a couple of issues with automake, AC_INIT: not a literal
17:11 < rconstruct> now in poppler
17:11 < rconstruct> i might be missing some .m4
17:11 < rconstruct> not sure which, either that, or some dev in the linux world was bored and decided to reshuffle the way things are done yet again
17:11 < Baldzius> hm
17:11 < Baldzius> what autoconf version you have?
17:12 < rconstruct> let me check, i installed a 8.0 minimal iso build environment
17:12 < rconstruct> and doing a 9.0/trunk target build
17:12 < rconstruct> autoconf 2.67
17:12 < rconstruct> automake 1.11.1
17:12 < rconstruct> m4 1.4.14
17:13 < Baldzius> 2.67 on host?
17:13 <@rxr> re
17:13 < rconstruct> yes
17:13 < Baldzius> that's strange
17:13 < Baldzius> 8.0 comes with 2.66
17:13 < rconstruct> i updated some pkgs, was having some problems with the build
17:13 < rconstruct> btw, my patches went through?
17:14 < Baldzius> which ones?
17:14 < rconstruct> 1m
17:14 < Baldzius> as for orbit2, that's been a while, just mv .err to .log
17:14 < Baldzius> I would disable it in stage 0
17:14 < Baldzius> rxr keeps it for cross compile
17:14 < rconstruct> well i removed that pkg atm, and do a orbit2, libbonobo, etc later, if i eventually need it
17:15 <@rxr> yes,
17:15 <@rxr> you can disable orbit and glib or what other thing was in stage 0
17:15 < Baldzius> glib is fine
17:15 < Baldzius> stupid orbit2 only :/
17:15 <@rxr> in trunk I want to fix it the real way, namely by introducing a new toolchain mark to only enable the orbit2 stage 0 build if orbit2 and stuff is cross build
17:15 < rconstruct> patches for netkit-rsh, scribus, live, libmikmod, foomatic-db, foomatic-filters
17:15 <@rxr> (we need the native executable around for the "generators")
17:16 < rconstruct> going to mail them to t2@t2-project.org
17:16 < rconstruct> right?
17:16 < Baldzius> yep
17:16 <@rxr> I think glib or what it was is only in stage 0 for orbit2, anyway, ...
17:16 < rconstruct> strange, i have glib, and was built
17:17 < Baldzius> when you sent email?
17:17 < rconstruct> but hmm ah i know
17:17 < rconstruct> yes, i built orbit at stage 0 iirc
17:17 < rconstruct> it was protesting about libidl
17:17 < rconstruct> no, cairo
17:17 < rconstruct> the .conf had pkginstalled cairo || var_append confopt ' ' '--disable-tests'
17:17 < rconstruct> either cairo or libidl, it was late and i was sleepy lol
17:18 < rconstruct> but removed the test and forced var_append confopt, that fixed it lol
17:18 < rconstruct> ok, let me mail the patches
17:18 < Baldzius> could you upload you error somewhere
17:18 < Baldzius> like pastebin or so
17:18 * rxr we need a t2 pastebin :-)!
17:19 <@rxr> clipboard.t2-proejct.org?
17:19 < rconstruct> http://old.nabble.com/Fixes-for-netkit-rsh%2C-scribus%2C-foomatic-db%2C-foomatic-filters%2C-libmikmod%2C-live-to29381754.html
17:19 < Baldzius> maybe paste.
17:19 < rconstruct> going to mail to t2 anyway, shouldn't have used nabble
17:20 < rconstruct> k sent
17:21 -!- mqueiros_ [~mqueiros@a83-132-248-197.cpe.netcabo.pt] has joined #t2
17:22 < rconstruct> the orbit2 error? need to do another build, i'm at stage 5 with poppler atm
17:22 < Baldzius> rconstruct: as for autoconf: could you revert last commit? and recompile autoconf on all stages
17:22 < Baldzius> and see if you get errors
17:22 < Baldzius> 2.66+fix should work fine
17:22 < Baldzius> I am sure about it
17:22 < rconstruct> ok, 1m
17:23 < Baldzius> thanks for patches, either me or rxr will apply them
17:23 <@rxr> which package? I still got some autoconf regen errors in the last night's trunk, too
17:23 < Baldzius> really?
17:23 < Baldzius> ok I will check myself then
17:23 <@rxr> yes, on x86-64
17:23 < rconstruct> there will be more, i already noticed some pkgs that need update or that no doubt will hate gcc 4.5
17:23 < Baldzius> autoconf 2.67?
17:23 <@rxr> some of the gnome things that have this if x86-64 then regen fluff, ...
17:24 <@rxr> not all just some
17:24 <@rxr> IIRC before the update
17:24 < Baldzius> oh for gnome, that's fine
17:24 < Baldzius> happens
17:24 <@rxr> yeah, not fine if the crap does not build, ...
17:24 <@rxr> but I have to urgently get back to my unrelated "portability hackery"
17:24 < rconstruct> ok, so want me to enable orbit2, libbonobo, libglade, etc
17:24 < Baldzius> that's why it's crap :)
17:24 <@rxr> yesh, ..
17:24 < rconstruct> to try and see where things explode (if they do) ?
17:24 < rconstruct> ok
17:24 <@rxr> IIRC in my build even gtksourceview and thus some other packgaes failed :-(
17:25 <@rxr> well, not that I need the personally, but they still should build, :_)))
17:25 < Baldzius> might be autoconf 2.67
17:25 <@rxr> it was before the update
17:25 < Baldzius> could you try revert last change?
17:25 < Baldzius> doh
17:25 < rconstruct> Baldzius: going to 2.66 yes
17:25 < Baldzius> that's trange
17:25 < Baldzius> last time I built trunk it was mostly fine
17:25 <@rxr> Baldzius: btw. in the future those checks should always be libdir != lib or so, as also other archs have lib64
17:26 <@rxr> I suppose this gnome configure regen on x86-64 are due to lib64, right?
17:26 <@rxr> anyway, we should get this sporadic auto* fluff fixed
17:26 < Baldzius> yep
17:26 <@rxr> but I'm now looking to my other's window program code, now, again, cu :-)
17:26 < Baldzius> it's ridiculous now
17:26 < rconstruct> hmm, i haven't had issues with lib64 with gnome stuff, and i keep the other (t2) machines up to date, but they're a bit old though
17:26 <@rxr> autoconf* junk always was
17:26 <@rxr> guess why exact-image and all other new stuff from me does not use it
17:27 <@rxr> got my hands and head bruned back in the gsmp days
17:27 <@rxr> not to mention all the regular breakage I have to deal in T2
17:27 <@rxr> really sick of it
17:27 < rconstruct> lol
17:27 <@rxr> http://rene.rebe.de/2010-02-16/darn-how-i-hate-this-gnuauto-junk/
17:28 <@rxr> all this generation of generated input of unreadable and undebugable macro fluff really is just insane
17:28 < rconstruct> yeah, llvm/clang, that will be fun
17:28 < rconstruct> or open64, the amd compiler
17:28 < rconstruct> btw not sure if that's OSS; i know it's free, just not sure if it's OSS
17:28 < rconstruct> they have a set of patches for gcc for AMD64 though
17:28 <@rxr> we have llvm/clan in t2
17:28 < rconstruct> besides some linear algebra libraries
17:29 < rconstruct> yeah, it's in the config
17:29 < rconstruct> need it for OSL
17:29 < rconstruct> OSL @ googlecode
17:30 < rconstruct> Baldzius: btw, seems it's present in 2.66 as well
17:30 < rconstruct> the error i had with 5-poppler was
17:30 < rconstruct> configure:12439: error: possibly undefined macro: AS_MESSAGE_LOG_FDdnl If this token and others are legitimate, please use m4_pattern_allow. See the Autoconf documentation.
17:31 < rconstruct> so, autoconf 2.66 + hot-fix.patch
17:31 < Baldzius> what's the earliest crash?
17:31 < Baldzius> so I could test myself pretty easy
17:31 < Baldzius> ok never mind will check on my current build
17:31 < rconstruct> earliest i had was orbit2 at stage 0
17:31 <@rxr> that does not count
17:31 <@rxr> it's using host tools
17:32 < rconstruct> ok, other than that, everything goes ok
17:32 < rconstruct> until poppler at stage 5
17:32 < Baldzius> ok
17:33 < rconstruct> http://thread.gmane.org/gmane.comp.sysutils.autoconf.general/13133/focus=13153
17:34 < rconstruct> what a mess
17:36 < Baldzius> 2.66 was a big F*
17:37 < Baldzius> ok, popple builds fine on 8.0 and trunk for me
17:37 <@rxr> 64bit ?
17:37 < Baldzius> but my trunk is July 16th
17:37 < Baldzius> yes 64 bits
17:37 <@rxr> rconstruct: do you build for x86-64 ?
17:37 < rconstruct> yes
17:37 <@rxr> hm
17:37 < Baldzius> I don't do 32 bits anymore
17:37 * rxr should focus elsewhere
17:37 < Baldzius> only for my laptop
17:37 * rxr does
17:38 < Baldzius> fecking flesh
17:38 < Baldzius> flash
17:38 < rconstruct> would love to get a power7 though
17:38 <@rxr> hahaha
17:38 < rconstruct> lol
17:39 < rconstruct> i don't think they sell power workstations anymore though
17:40 < rconstruct> hmm, host has 2.66 + hotfix, target had autoconf 2.66+hotfix rebuilt at stage 0, 1, 3, 5
17:40 < rconstruct> i have the same error in poppler, 1m
17:40 < rconstruct> going try a thing, brb
17:43 < Baldzius> I need to get my trunk from scratch
17:43 < rconstruct> hmm, seems gmx.de doesn't accepts mails from gmx.com
17:43 < rconstruct> my mail to the ml was bounced
17:43 < rconstruct> lol
17:44 < rconstruct> http://t2.pastebin.com/fKRez6qR
17:44 < rconstruct> netkit-rsh
17:44 < rconstruct> missing stdlib.h in rlogind/auth.c
17:44 < Baldzius> it went through
17:44 < Baldzius> http://news.gmane.org/gmane.comp.t2.devel
17:45 < rconstruct> ah ok, thanks :)
17:45 < rconstruct> btw, added geeqie, but let me do a build from scratch
17:45 < rconstruct> been using it for a year or so
17:45 < Baldzius> what's that?
17:45 < rconstruct> replacement for gqview
17:46 < rconstruct> gqview = dead
17:46 < Baldzius> ic
17:46 < rconstruct> all bugfixes for gqview = went to the fork, geeqie
17:46 < rconstruct> new development also went to geeqie
17:46 < rconstruct> they released a stable 1.0 some time ago
17:46 < rconstruct> so it makes sense to get rid of gqview and replace it by geeqie, but first things first
17:47 < rconstruct> lots of updates and gcc 4.5 fun ahead surely, specially in some older slightly less maintained code for sure
17:47 <@rxr> most stuff builds in trunk and 8.0 with gcc 4.5
17:47 < Baldzius> host autoconf doesn't matter for 5-poppler so it must be autoconf 2.67
17:47 < rconstruct> happens with 2.66 too
17:47 < Baldzius> 2.66 is crap
17:48 < Baldzius> you can downgrade to 2.65
17:48 < rconstruct> going to try a thing, i'll let you know if it works, give me 5m
17:48 < Baldzius> hm but why it compiles for me on 8.0 then
17:48 < rconstruct> i'm too stubborn to just let this go, it's a good reason to send flamemails to the autoconf devs
17:48 < Baldzius> :)
17:48 < rconstruct> god knows they deserve it... lol
17:48 < Baldzius> (nod)
17:49 < Baldzius> or we could send some friendly email with lots of hugs :)
17:50 < rconstruct> lol, a chain mail offering compassion and wishes for a speedy recovery from the mental breakdown that lead to 2.66
17:50 < Baldzius> heh
17:50 < rconstruct> well, it's always a bitch though, cmake has its woes as well, scons....
17:51 < rconstruct> scons scons scons...
17:51 < rconstruct> ....
17:51 < rconstruct> hmm, ok, i'll b back in 5m, doing evil things with poppler, m4 and autoconf
17:51 < Baldzius> good luck
17:52 < rconstruct> thanks )
17:52 < rconstruct> ;)
17:54 < rogermason> since cmake was mentioned... I'm doing a manual build of Paraview, preparatory, to putting together a t2 patch. Paraview has to be built outside its source tree. can t2 handle that automatically or must I do something special in the conf file?
17:55 <@rxr> some cmake packages have this alread
17:55 <@rxr> y
17:56 <@rxr> hook_add preconf 5 'mkdir objdir; cd objdir; cmake -DCMAKE_INSTALL_PREFIX=/$prefix -DCMAKE_LIBRARY_PATH=$libdir ..'
17:56 <@rxr> something like that
17:58 < rogermason> OK, thanks. We'll see how it goes. My manual build is still chugging along.
18:19 < rconstruct> Baldzius: so, autoconf 2.65 ?
18:19 < Baldzius> no, maybe leave the original
18:19 < Baldzius> we have in trunk
18:19 < rconstruct> 2.67?
18:19 < Baldzius> yes
18:20 < Baldzius> I started my own ref build from scratch
18:20 < rconstruct> quick question, do i need libiconv? or minimal+xorg template already adds it?
18:21 < rconstruct> brb
18:21 < Baldzius> don't add libiconv
18:22 < Baldzius> it's not needed
18:22 < Baldzius> I think it comes from glibc or something
18:22 < rconstruct> i haven't, just wondering if minimal+xorg added it
18:23 < rconstruct> i'll b back in 30, food time
18:23 < Baldzius> I don't remember what's the story but I fecked my system completely once I started playing with libiconv
18:23 < Baldzius> don't recall details now
18:53 -!- mjungwirth2 [~mjungwirt@213.174.252.112] has joined #t2
18:58 <@rxr> IIRC one needs it for non glibc builds, such as with uclibc
20:02 -!- Baldzius [~aldas@87-198-133-94.static.ptr.magnet.ie] has quit [Quit: Leaving]
20:10 -!- mtr [~michael@v29762.1blu.de] has quit [Ping timeout: 260 seconds]
20:11 -!- mtr [~michael@v29762.1blu.de] has joined #t2
20:26 < rconstruct> well, started from scratch
20:26 < rconstruct> but it's in stage 5 gcc already
20:26 < rconstruct> no errors yet
20:26 < rconstruct> i only start getting errors with some non essential pkgs atm, well, relatively non essential... poppler
20:26 < rconstruct> but let's see how things work out
20:28 < rconstruct> btw, gcc 4.5.1 is being built with mpc, mpfr, gmp, cloog-ppl, ppl
20:45 < rconstruct> brb
20:55 -!- mjungwirth2 [~mjungwirt@213.174.252.112] has quit [Quit: Nettalk6 - www.ntalk.de]
22:28 < rconstruct> well
22:28 < rconstruct> stage 5, gobject-introspection
22:28 < rconstruct> cannot find cairo
22:28 < rconstruct> cairo is installed
22:29 < rconstruct> gobject-introspection.conf has
22:29 < rconstruct> pkginstalled cairo || var_append extraconfopt ' ' '--disable-tests'
22:29 < rconstruct> cairo is installed, if i remove the pkginstalled cairo || tests
22:30 < rconstruct> and just add var_append extraconfopt ' ' '--disable-tests'
22:30 < rconstruct> detects cairo and proceeds normally
22:30 < rconstruct> gobject-introspection builds
22:32 < rconstruct> something fishy here
22:32 < rconstruct> cairo is not installed, build priority is 126.500 for gobject-introspection
22:33 < rconstruct> and 127.030 for cairo
22:33 < rconstruct> and yet, the .cache file has [DEP] cairo
22:34 < rconstruct> how can this be, if cairo according to the .desc priorities is built after gobject-introspection?
22:35 < rconstruct> going to change cairo from 127.030 to 126.030, gobject-introspection is 126.500, either that, or i disable tests and gobject-introspection is built without cairo, which raises the question of how is it in the .cache file
22:36 < rconstruct> in either caise, i can only change the .conf and force --disable-tests, and who knows where this will lead, or change the priority
23:14 < rconstruct> rxr: $stagelevel -ge = greater or equal, -gt = greater than?
23:15 < rconstruct> what are the flags here?
23:33 < rconstruct> ok fixed it
--- Log closed Tue Aug 10 00:00:49 2010