[02:01] <jbailey> lamont: ping?
[02:02] <jbailey> lamont: Now that we're blocked on Carlos ack'ing the patch, tausq commiting gdb patch, benc/amit uploading a new kernel, the world answering how to deal with missing kernel headers, the result of which might be somewhat easily debugging of gij, do you have any other tasks that I should be hacking on? =)
[02:02] <jbailey> Otherwise I might spend the evening playing with 64 bit user space.
[03:01] <lamont> well, we could put together a slightly tweaked 8.16 tree with our patch in it, and maybe call it 8.16hppa1 :-)
[03:02] <lamont> that way we could have SMP support before monday
[03:04] <lamont> gmp is also holding up some stuff
[03:04] <lamont> and then there's perl
[03:43] <jbailey> lamont: That might do.
[03:43] <jbailey> Actually, do you mind playing with the kernel and seeing if a known configuration will boot?
[03:43] <jbailey> I don't know if the non-bootingness of it is code or config.
[03:44] <jbailey> I can take a look at perl this WE, though.
[04:04] <lamont> non-bootingness?
[04:05] <lamont> my evening involves finishing the reorg of my office, before mitzi gets back home and wants to reclaim her kitchen (currently my overflow area for organizing...)
[04:06] <lamont> and then there are other honey-do items
[04:32] <lamont> devel/ecj_3.3~M7-2: Dep-Wait-Removed by buildd-hppa+bld-4 [optional:uncompiled] 
[04:32] <lamont> why does it hate me?
[04:33] <lamont> amusingly, ecj is not in PaS.. how did it get removed, I wonder...
[04:34] <jbailey> lamont: You reproduced the build failure with perl, yes?
[04:35] <jbailey> using dpkg-buildpackage -B ?
[04:35] <lamont> well, it builds just fine on hppa
[04:35] <lamont> it's everywhere else that's broken
[04:35] <jbailey> 'k.
[04:35] <jbailey> It's working its way through the testsuite on my ppc box.
[04:35] <lamont> OTOH, it fails to build on a feisty kernel on my i386, with a completely diff failure than the buildds
[04:36] <lamont> thread tests fail on the feisty i386 kernel, at least for me.
[04:37] <lamont> I'm disinclined to annoy people by having the buildds retry on all 5 architectures... making perl uninstallable would tend to annoy people.  I know it annoys me.. :-(
[04:37] <jbailey> Hey, do you have a Debian box handy?
[04:37] <jbailey> Can you try building it there?
[04:37] <lamont> checking if krita should be compiled... no
[04:37] <lamont> ...
[04:37] <lamont> dh_install -pkrita
[04:37] <lamont> cp: cannot stat `./debian/tmp/usr/bin/krita': No such file or directory
[04:37] <lamont> go koffice!!!
[04:37] <jbailey> I'm curious if it will fail now that they're on a current glibc.
[04:37] <lamont> I have a debian chroot...
[04:38] <jbailey> Should be close enough.
[04:38] <jbailey> Debian boxes don't run Debian kernels anyway.
[04:38] <lamont> lol
[04:39] <lamont> for the moment, I want to pause and finish moving /home to its new, uh, home.
[04:39] <jbailey> Sure, no worries.
[04:39] <jbailey> I have a perl build in one window and  abinutils build in the other. =)
[04:39] <jbailey> Tomorrow I'll probably spend the day studying, and hack in the evening.
[04:40] <jbailey> But I'm hoping we can magic together enough answers for getting gdb to work.
[04:40] <lamont> I wonder if the DC x86 box would be a good one to test-build on
[04:40] <lamont> working GDB would rock
[04:40] <lamont> do you have a gdb with the patch, and does that work?
[04:40] <lamont> or is the jury still out?
[04:41] <jbailey> It works for tausq.
[04:41] <jbailey> It's unbuildable on Ubuntu because we don't export the kernel header needed for gdb.
[04:41] <lamont> oh.  right.  sigh
[04:41] <jbailey> Which is to say, upstream doesn't either.
[04:41] <jbailey> So I started a thread on the list for solving that.
[04:41] <lamont> "the list"?
[04:41] <jbailey> parisc-linux
[04:41] <lamont> ah, ok
[04:41] <jbailey> Oh, also, mjw and doko helped find a gij hacker who will help us.
[04:42] <jbailey> But he's going to need gdb anyway.
[04:42] <lamont> ROCK!
[04:42] <lamont> anyway, about a cubic meter of crap to sort through so I can move things out of the kitchen before my wife gets home and kills me.
[04:42] <lamont> laggy, but still here
[04:45] <jbailey> Death by mitzi sounds probably spectacular.
[04:46] <lamont> irish redhead.
[04:46] <lamont> very spectacular
[04:47] <lamont> oh.  uic3 glibc corruption-detected thing would be useful for you to figure out/fix too.
[04:49] <jbailey> I can reproduce the build failure for perl./
[04:50] <lamont> tulip_2.0.6-5_20070713-1305:*** glibc detected *** /usr/bin/uic3: free(): invalid pointer: 0x00124098 ***
[04:51] <jbailey> gdb it, see what is being freed.
[04:51] <jbailey> If it's not threaded, gdb ought to work.
[04:51] <lamont> ah, ok
[04:51] <lamont> I've been being distracted by AT
[04:52] <jbailey> Automated Tests?
[04:52] <jbailey> Or your modem singing away to you?
[04:53] <jbailey> Trying obvious solution to perl build failure.
[04:53] <jbailey> SHELL = /bin/bash =)
[05:51] <lamont> assigned taskes
[06:28] <lamont> wow.  floor
[02:16] <jbailey> lamont: I've confirmed that the perl problem is a dash/bash thing.
[02:16] <jbailey> I'm guessing your chroot still has /bin/sh -> /bin/bash
[03:00] <jbailey> shinmen: Heya
[03:12] <jbailey> lamont: Looks like bingo!  Trying.
[03:43] <lamont> jbailey: updating the hppa chroots to use dash instead of bash.  thanks
[03:44] <jbailey> Just checking to make sure these lintian warnings aren't my fault and then I'll upload perl to gutsy.
[03:48] <jbailey> It's not, but I built it -B, feh.  Rebuilding to make sure I can do an install test.
[04:06] <shinmen> jbailey: Hey.
[04:11] <jbailey> lamont: I keep having this vague thought that the compiler should be written in something interpreter but hotspot optimised.
[04:11] <jbailey> Java's the logical choice for existing technology, but LISP or something could do as well once Parrot is real.
[04:12] <jbailey> And then we could have a massive distributed.net farm for all of our compilation.
[04:24] <jbailey> lamont: putting perl 5.8.8-7ubuntu2 now.  Please build it when you see it, kthxbye
[04:24] <jbailey> I should probably study for the rest of the day.  I'll be back when my brain can't handle more Unified Process.
[06:52] <jbailey> lamont: Bah, your buildd is being lazy =)
[06:52] <jbailey> But it's built succesfully on i386, amd64 and ppc now. =)
[08:23] <lamont> jbailey: rock
[08:24] <lamont> fwiw, cron.hourly actually runs hourly now on hppa...
[08:24] <lamont> perl is building as of 30 min ago
[08:24] <jbailey> Lovely.
[08:25] <jbailey> perl-modules has hit the archive.
[08:25] <jbailey> After that are we debootstrappable?
[08:25] <lamont> and gutsy-stage0 archive, too.
[08:26] <lamont> I expect so
[08:26] <lamont> gonna have to reboot again...
[08:26] <lamont> randomly overwriting /etc/group with the "right" file is kinda fraught with pain
[08:32] <lamont> this is a fun experiment... apt-get remove --purge hal dbus-1; apt-get install ubuntu-desktop
[08:32] <lamont> hope it works.
[08:42] <lamont> how do I tell the system to rescan for disks?
[08:42] <lamont> "reboot" seems kinda invasive.
[08:43] <jbailey> It should do it automatically.
[08:43] <jbailey> Assuming it's a pluggable bus.
[08:43] <lamont> that was what I was afraid of
[08:43] <jbailey> It should have sent an event.
[08:43] <lamont> SATA
[08:43] <jbailey> Right.
[08:43] <jbailey> Sata should send an event.
[08:44] <lamont> mind you, I did screw up everything's permissions
[08:46] <lamont> brb (I hope...)
[10:08] <jbailey> lamont: This lag isn't promising...