| 00:58.25 | *** join/#edev prpplague (~prpplague@ppp-70-249-149-70.dsl.rcsntx.swbell.net) |
| 02:08.56 | *** join/#edev GeorgeH (~George@c-69-141-105-145.hsd1.nj.comcast.net) |
| 03:27.44 | *** join/#edev bgamari (~ben@pool-96-233-179-97.spfdma.east.verizon.net) |
| 04:04.56 | *** join/#edev file (~file@asterisk/developer-and-muffin-lover/file) |
| 04:05.51 | *** join/#edev JoeLlama (~snork@unaffiliated/joellama) |
| 04:20.04 | *** join/#edev mbuf (user@nat/redhat/x-hfyxasiymwonhhuy) |
| 04:36.53 | *** join/#edev bgamari (~ben@pool-96-233-190-78.spfdma.east.verizon.net) |
| 06:14.34 | *** join/#edev clio (~andrej@85.159.109.222) |
| 09:08.55 | *** join/#edev florian (~fuchs@port-217-146-132-69.static.qsc.de) |
| 09:08.55 | *** join/#edev florian (~fuchs@Maemo/community/contributor/florian) |
| 09:48.34 | *** join/#edev rsFF (~rsFF@2.80.54.167) |
| 10:14.17 | *** join/#edev icanicant (~klawson@213.218.221.154) |
| 10:26.01 | *** join/#edev JoeLlama (~snork@cpe-75-85-45-219.socal.res.rr.com) |
| 10:26.03 | *** join/#edev JoeLlama (~snork@unaffiliated/joellama) |
| 10:30.52 | *** join/#edev madsy (~madsy@188.113.83.44) |
| 10:30.52 | *** join/#edev madsy (~madsy@fu/coder/madsy) |
| 11:01.39 | *** join/#edev bsdfox\ (~h36sa@c-76-20-20-210.hsd1.ca.comcast.net) |
| 11:10.50 | *** join/#edev g1powermac (~g1powerma@unaffiliated/g1powermac) |
| 12:01.02 | slashbeast | landley: Hi, any luck with debuging the mdadm? |
| 14:27.56 | *** join/#edev GPSFan (~kenm@64.92.145.112) |
| 14:27.56 | *** mode/#edev [+o GPSFan] by ChanServ |
| 14:48.40 | *** join/#edev risca (~risca@wi-secure-8080.cc.umanitoba.ca) |
| 15:04.57 | *** join/#edev gandhijee_ (~akp@50.12.169.99) |
| 15:38.52 | landley | slashbeast: alas, no. Been too busy. |
| 15:39.25 | landley | This morning dealing with paperwork, last night reviewing the mode parser contributed to toybox (getting it about halfway cleaned up), plus I have to install Fedora Core 16 for Denys Vlasenko to track down his bug. |
| 15:39.33 | landley | And the 3.3 kernel dropped, so I've had a test build of that running. |
| 15:39.59 | landley | I still have the window open on my desktop, but haven't downloaded your scripts to start building mdadm from source yet, which is the next debugging step. |
| 15:41.27 | landley | Oh, and the more/record-commands.sh stuff (which I'm using again for testing toybox) turns out to always have been subtly screwed up trying to record the host-tools.sh build (which itself edits the $PATH, sort of the point of that stage), and when a toybox change broke host-tools I went down the rathole of cleaning _that_ up... |
| 15:42.16 | landley | (The problem there is really that $OLDPATH is used to mean two different things, which used to coincide but it was a coincidence...) |
| 15:42.43 | landley | And now I'm back to daylight jobbery. |
| 15:43.06 | landley | I haven't forgotten! You're about #4 on my todo list right now, and I consider it a release blocker for aboriginal 1.1.2. |
| 15:43.37 | landley | (Haven't even _started_ testing the new uCLibc release yet...) |
| 15:57.09 | *** join/#edev prpplague (~danders@192.94.92.11) |
| 15:58.30 | *** join/#edev sndcrb (~sndcrb@gprs-prointernet-ff8e6a00-23.dhcp.inet.fi) |
| 16:47.57 | *** join/#edev __av500__ (~av500@pd95bf55a.dip0.t-ipconnect.de) |
| 17:42.39 | *** join/#edev NightMonkey (~NightMonk@pdpc/supporter/professional/nightmonkey) |
| 17:49.21 | *** join/#edev risca (~risca@wi-secure-8080.cc.umanitoba.ca) |
| 18:03.46 | *** join/#edev mrcan (~mrcan@unaffiliated/mrcan) |
| 18:09.21 | *** join/#edev mrcan (~mrcan@unaffiliated/mrcan) |
| 18:27.41 | *** join/#edev bgamari (~ben@72.19.78.86) |
| 18:37.07 | *** join/#edev sndcrb (~sndcrb@gprs-prointernet-ff8e6a00-23.dhcp.inet.fi) |
| 18:47.11 | *** join/#edev JoeLlama (~snork@unaffiliated/joellama) |
| 19:01.49 | *** join/#edev CosmicPenguin (~nobody@soa.codeaurora.org) |
| 19:01.49 | *** mode/#edev [+o CosmicPenguin] by ChanServ |
| 19:12.00 | MonMotha | I am amused at how broken jabber.org has been recently |
| 19:52.20 | landley | is amused that the SUSv4 definition of sys/stat.h does _not_ define actual values for S_IROTH and friends, but the SUSv4 page for the command line utility chmod _does_. :) |
| 19:52.42 | landley | (becuase "chmod 7777" has a defined meaning from the command line, but using 0x7777 from C code could mean _anything_ Yeah, right...) |
| 19:57.59 | alex4nder | I've often wondered how that doesn't break more often. |
| 20:21.37 | landley | alex4nder: it's posix. http://pubs.opengroup.org/onlinepubs/009604499/utilities/chmod.html |
| 20:22.10 | alex4nder | landley: no shit? |
| 20:22.26 | landley | And that means that Linux uses the same constants on all architectures, because doing translation from the userspace 7777 to something else is pointless. |
| 20:23.14 | landley | To clarify: posix specifies that userspace chmod must support 7 (suid, sgid, sticky), 777 (rwx+ugo). |
| 20:23.27 | landley | And as long as the userspace tool is using that constants, the kernel might as well. (Because it historically always has.) |
| 20:23.27 | alex4nder | my point is.. 0x7777 != 07777 |
| 20:23.38 | landley | Octal. 07777 |
| 20:23.51 | landley | Remember: Unix started out on a PDP-7, which was a 6 bit architecture. |
| 20:23.54 | alex4nder | you wrote ... 0x7777 |
| 20:24.05 | landley | Sorry, force of habit. :) |
| 20:24.06 | alex4nder | jesus, do you walk around assuming everyone is a total dumbass? |
| 20:24.15 | alex4nder | you're sitting in #edev |
| 20:24.38 | alex4nder | how many people in here are professional kernel programmers? |
| 20:24.39 | landley | shrugs. |
| 20:25.15 | landley | I actually know why posix _doesn't_ specify the kernel-level constants, I was assuming you were surprised that the constants were posix-specified in another context. |
| 20:25.27 | landley | But I'll stop bothering you. |
| 20:25.36 | landley | <PROTECTED> |
| 20:25.46 | landley | How do I get xchat to do that... |
| 20:25.51 | alex4nder | if you have to say you're doing it, you probably aren't going to. |
| 20:25.57 | alex4nder | that defeats the purpose. |
| 20:26.02 | ohsix | i don't think xchat even does that |
| 20:26.27 | landley | ohsix: I just installed it a few days ago. Pidgin before that... |
| 20:26.30 | ohsix | alex4nder: he's got history on his miiiind |
| 20:26.30 | alex4nder | landley: you're not bothering me, I just don't understand your perspective a lot of the time, when you start talking. |
| 20:27.45 | ohsix | fwiw, the impression that i get from him matches what he just said |
| 20:28.36 | landley | Ah, it's /ignore nick!blah ALL QUIET |
| 20:31.59 | *** join/#edev dijenerate (~dijenerat@173.225.251.175) |
| 20:33.16 | *** join/#edev Guest58937 (~simoirc@109.73.162.119) |
| 20:35.24 | ohsix | landley: hey you know a bit about qemu right, do you think i could use the one that translates syscalls to run a chroot from a much newer (2.6) kernel on a 2.2 host? |
| 20:35.54 | ohsix | host & chroot = 1386 |
| 20:43.58 | landley | Hmmm... |
| 20:44.11 | landley | Maybe? |
| 20:44.21 | landley | The main question is whether or not you can build qemu at all on a 2.2 host. |
| 20:44.24 | landley | (Probably yes.) |
| 20:44.42 | landley | And then the translation thing... |
| 20:44.55 | landley | It's essentially cross compiling along a different axis. |
| 20:46.03 | landley | qemu is built to distinguish host system calls from target system calls and translate between then, but what it needs in this case is two syscall definitions for the same architecture, but different kernel headers. |
| 20:46.56 | landley | I know you can use _system_ emulation to boot a kernel implementing the newer system calls on a 2.2 host. |
| 20:47.03 | landley | (Assuming you can build qemu on the 2.2 host at all.) |
| 20:47.17 | ohsix | the real issue is the glibc in the chroot does not like the old kernel, figured it'd might be a cute shortcut until i get around replacing it entirely |
| 20:47.37 | landley | You could even get it to export the host directory through virtfs and use a -nographic serial console so stdin/stdout went to your shell so it would act mostly likely a chroot. |
| 20:48.20 | landley | It sounds worth a try. I'd say you have maybe a 20-30% chance of it working out of the box? |
| 20:48.48 | landley | (That's between "can I compile qemu under 2.2" and "will that support what this glibc needs".) |
| 20:49.18 | landley | Lemme put it this way: qemu could easily be extended to do this by its' developers. But whether or not it _currently_ supports it... I'm not sure it's occurred to anybody to try. |
| 20:49.54 | landley | Hmmm... do you have any idea what glibc is barfing on? |
| 20:50.05 | landley | Most likely lack of tls and sysenter. |
| 20:50.40 | landley | What I'm worried about it stuff like stat growing new fields at the end: I have no idea what qemu would _do_ with them, translation-wise. |
| 20:51.24 | ohsix | it just says the host is too old |
| 20:51.39 | landley | Just tried it? |
| 20:51.47 | ohsix | the drive just remounted read only when i was doing something soi've got to deal with that now |
| 20:52.29 | landley | You might want to try user mode linux. |
| 20:53.00 | landley | THis is hideously out of date: http://landley.net/writing/docs/UML.html |
| 20:53.31 | landley | That just depends on ptrace, although... 2.2. It was a while ago. |
| 20:55.43 | MonMotha | is 2.2 really even relevant anymore? |
| 20:56.05 | landley | MonMotha: no, but it sounds like he's got an old system lying around. |
| 20:56.08 | MonMotha | ah |
| 20:56.16 | landley | One of those microvaxen that got drywalled in or some such. |
| 20:56.22 | MonMotha | heh, oh that story |
| 20:56.29 | MonMotha | though I've always heard it with novell |
| 20:56.32 | landley | The sad part is that story was _true_... |
| 20:56.38 | landley | And it happened more than once. |
| 20:56.45 | MonMotha | landley: oh, I'm sure |
| 20:56.57 | landley | There was a slashdot thread on it. |
| 20:57.10 | landley | The one that impressed me was a PDP-10 that had an air conditioning system built around it. |
| 20:57.19 | MonMotha | haha |
| 20:57.36 | calculus | yeah, it was too much of a space heater |
| 20:57.45 | landley | They finally tracked down where it _was_, then had to go dig into it, and when they took the 8 inch floppies out they crumbled to dust... |
| 20:58.01 | landley | Well, mylar fragments anyway. |
| 20:58.04 | MonMotha | last I checked, most of my 5.25" still worked |
| 20:58.14 | MonMotha | I should re-archive those |
| 20:58.25 | MonMotha | I have used 8", but not really personally |
| 20:58.26 | MonMotha | a bit before my time |
| 20:58.27 | landley | I think being near a climate control system had something to do with it. :) |
| 20:58.33 | MonMotha | yeah, that's probably true |
| 20:58.43 | ohsix | i've even got a machine to replace it at the moment, was just looking at the path of least resistance to running some up to date software |
| 20:58.43 | landley | 8" was before my time too. I have one in a box somewhere, though. |
| 20:58.51 | landley | And it's meeting time... |
| 20:59.03 | landley | ohsix: tar up the old root filesystem and use _that_ as a chroot under a new kernel. |
| 20:59.08 | landley | THat's the path of least resistance. |
| 20:59.22 | landley | Backwards compatability is a big thing with linux. Forwards compatability doesn't happen ever. |
| 20:59.46 | landley | SO new kernel with old root filesystem: sure. New root filesystem with old kernel: the pain is unending. |
| 20:59.47 | landley | Meeting time... |
| 21:00.09 | *** join/#edev bgamari (~ben@physicsnat56.physics.umass.edu) |
| 21:11.39 | *** join/#edev bgamari (~ben@physicsnat56.physics.umass.edu) |
| 21:43.51 | *** join/#edev florian (~fuchs@Maemo/community/contributor/florian) |
| 22:01.02 | *** join/#edev mrcan (~mrcan@unaffiliated/mrcan) |
| 22:08.00 | *** join/#edev Guest953 (~simoirc@109.73.162.119) |
| 22:14.23 | *** join/#edev bgamari (~ben@204.213.244.97) |
| 22:23.25 | *** join/#edev mrcan_ (~mrcan@unaffiliated/mrcan) |
| 22:26.35 | *** join/#edev mrcan (~mrcan@unaffiliated/mrcan) |
| 23:40.16 | *** join/#edev bgamari (~ben@pool-96-233-190-78.spfdma.east.verizon.net) |