T2 IRC Log: 2010-08-23

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.

« prev | next »

--- Log opened Mon Aug 23 00:00:49 2010
04:29 -!- old_gaffer [~quassel@211.238.240.87] has joined #t2
06:09 -!- old_gaffer [~quassel@211.238.240.87] has quit [Read error: Connection reset by peer]
06:15 -!- old_gaffer [~quassel@211.238.240.87] has joined #t2
06:25 -!- old_gaffer [~quassel@211.238.240.87] has quit [Read error: Connection reset by peer]
06:27 -!- old_gaffer [~quassel@211.238.240.87] has joined #t2
08:00 -!- snowwoolf [~fedorowp@CPE0013f7acc3c0-CM0013f7acc3bc.cpe.net.cable.rogers.com] has joined #t2
08:02 < snowwoolf> Hi all. Are there any directions for getting up and running with an 8.0 install?
08:03 < snowwoolf> I did an install using t2-8.0-minimal-i486.iso and I'd like to know how to install packages.
08:05 <@rxr> usually from source, by checking out the t2 sources into /usr/src/t2-trunk (or so) and using scripts/Emerge-Pkg ...
08:07 < snowwoolf> There is no scripts directory in the installation ;-)
08:09 <@rxr> you can either unpack the release tarball or use svn to checkout the source
08:09 <@rxr> http://dl.t2-project.org/source/
08:10 <@rxr> http://t2-project.org/download/
08:10 <@rxr> gotta go into the office - cu then - by
08:10 < snowwoolf> Okay, thanks! Btw, it would be nice if stone had an option to do that.
09:01 <@rxr> ok, noted
09:03 < CIA-32> rene * r37398 /trunk/package/base/busybox/busybox.desc: * updated busybox (1.17.1 -> 1.17.2)
09:05 < CIA-32> rene * r37399 /trunk/package/xorg/pixman/pixman.desc: * updated pixman (0.18.2 -> 0.18.4)
09:06 < CIA-32> rene * r37400 /trunk/package/xorg/xf86-input-evdev/xf86-input-evdev.desc: * updated xf86-input-evdev (2.4.0 -> 2.5.0)
09:19 < CIA-32> rene * r37401 /trunk/package/network/lighttpd/ (lighttpd.conf lighttpd.desc): * updated lighttpd (1.4.27 -> 1.4.28)
09:19 <@rxr> http://t2-project.org/packages/vx32.html
09:36 < CIA-32> rene * r37402 /trunk/package/archiver/xz/ (. xz-noclobber-lzma.patch xz.cache xz.desc):
09:36 < CIA-32> Roger Mason :
09:36 < CIA-32> * added xz (4.999.9beta) - XZ Utils
09:37 <@rxr> oh, now that makes the fans spinning: load average: 7.31, 5.70, 3.24
09:37 <@rxr> temp1: +51.2 C (high = +70.0 C)
09:37 <@rxr> /dev/sda: WDC WD20EARS-00S8B1 : 38 C or F
10:33 < CIA-32> rene * r37403 /trunk/scripts/ (Build-Target config.in): * added support for xz compressed binary archives
10:41 <@rxr> http://rene.rebe.de/2010-08-23/so-how-good-is-the-new-xz-compressor/
10:48 < koan> rxr: did you also measure the decompression times? these are also relevant
10:49 < koan> xz generally decompresses faster than bzip2
10:49 <@rxr> ah, ok
10:49 <@rxr> you can leave a comment :-)
10:50 < koan> that's why Fedora, Arch Linux and Slackware are using xz as default for their packages now: they are compressed just once, which isn't relevant for the user, but decompression is faster and the packages are generally smaller, so downloading is faster
10:51 * koan is entering these comments on rxr's blog at the moment
10:56 <@rxr> just leave a tiny quick one, I got decompression numbers
10:56 < koan> submitted
10:57 < koan> I also remember xz doesn't compress plain text well
10:57 < koan> so it really depends on your use case if it's better
10:57 <@rxr> plain text was the only thing where it was smaller than lzma on that blog post :-)
10:58 <@rxr> but you are right, for decompression it looks better
11:02 <@rxr> http://rene.rebe.de/2010-08-23/so-how-good-is-the-new-xz-compressor/#comment-71312
11:07 <@rxr> strange, xz --long-help tells me:
11:07 <@rxr> On this system and configuration, this program will use at maximum of roughly
11:07 <@rxr> 1584 MiB RAM and 7 threads.
11:07 <@rxr> hwoever, it did not appear to put more than one core under load, strange thing
11:08 <@rxr> especially as time accounted user time was not higher than teal wall clock time elapsed, which it should if it used more than one core, ...
11:09 <@rxr> whatever, got more important stuff, e.g. real work to do ...
11:10 <@rxr> 32040 root 20 0 678m 224m 652 R 100 5.7 0:14.17 xz
11:11 <@rxr> PID SYSCPU USRCPU VGROW RGROW USERNAME THR ST EXC S CPU CMD 1/2
11:11 <@rxr> 32040 0.06s 9.94s 0K 5996K root 1 -- - R 100% xz
11:11 <@rxr> ah!
11:11 <@rxr> Multithreaded compression and decompression are not implemented yet, so this option has no effect
11:11 <@rxr> for now.
11:11 < koan> :-)
11:12 <@rxr> cause we got parallel bzip2 and gzip versions that significantly speed up things, ...
11:12 <@rxr> Multithreaded compression and decompression are not implemented yet, so this option has no effect
11:12 <@rxr> for now.
11:12 <@rxr> oops
11:12 <@rxr> http://www.t2-project.org/packages/lbzip2.html
11:13 <@rxr> http://www.t2-project.org/packages/pigz.html
11:13 <@rxr> was thinking about making those multi-threading frontends the default if enable, to significantly speed up my reference test builds ;-)
11:13 <@rxr> serial unpack, pack time is the biggest serialiization bottleneck on my many-core VM T2 builds, ...
11:24 <@rxr> little warm already, ...
11:27 -!- snowwoolf [~fedorowp@CPE0013f7acc3c0-CM0013f7acc3bc.cpe.net.cable.rogers.com] has quit [Quit: Leaving.]
11:41 < rogermason> After the tinkering I did to make xz compile and not overwrite lzma it is a miracle that a) It works and b) your computer does not glow in the dark.
11:42 < koan> lol
11:42 <@rxr> heh
11:50 < rogermason> By the way, is there a short option for tar to use xz, analogous to -j for bzip2? I don't see it in the tar man page.
11:52 <@rxr> looks like latest tar version has it - I saw on google --xz may work
11:53 <@rxr> but looks like ours does not have it yet
12:19 -!- Baldzius [~aldas@87-198-133-94.static.ptr.magnet.ie] has joined #t2
12:19 < Baldzius> moin
12:19 <@rxr> moin moin Baldzius
12:19 < Baldzius> moin rxr
12:20 < Baldzius> could you restart my vm :)
12:20 < Baldzius> can't ssh anymore, made something stupid
12:28 -!- mtr` is now known as mtr
12:29 <@rxr> hm, looks up
12:29 <@rxr> Connection refused
12:29 <@rxr> hm
12:30 <@rxr> I sent it a Ctrl-Alt-Del
12:30 < Baldzius> super
12:30 < Baldzius> should come up fine
12:31 <@rxr> its currently doing an fsck, 198 days without being checked, check forced.
12:31 <@rxr> stuipd ext4 or what you got :-(
12:31 <@rxr> guess it's time to work on the web view of the vm infrastructure, so you could see the framebuffer yourself, ...
12:32 <@rxr> now it's up, with ssh
12:32 <@rxr> did you kill sshd or so =?
12:33 <@rxr> maybe we/I should start experiementing with VMs based on btrfs
12:33 <@rxr> in the hope it becomes the better fs, ...
12:33 < Baldzius> thanks, works
12:34 < Baldzius> I am planning start playing with lxc
12:34 < Baldzius> should be fun
12:34 <@rxr> which is like vserver, openvz?
12:34 < Baldzius> yes
12:34 < Baldzius> looks pretty good
12:34 <@rxr> I have my problems with such stuff
12:35 <@rxr> a) can you trust they really got every syscal patched?
12:35 <@rxr> b) you can only run the same linux kernel (e.g. not even a newer one to fix bugs, nor BSD, or Windows, etc.
12:35 < Baldzius> yeah, security is an issue here , unfortunately :/
12:36 <@rxr> c) incomplete, vserver did not even support mknod, needed for t2 builds, and openvz patchery was so messy, that you know we never reached an uptime of over 26days or so
12:36 < Baldzius> not sure if you need same kernel as for Linux, but no other flavour , yes
12:36 <@rxr> with "containers" it's all running in one big kernel, ...
12:37 <@rxr> soo all containers get the same linux kernel version, the exactly same, all running in one address space
12:37 <@rxr> openvz also hat this performance starving problems, when one container produced lots of memory or i/o pressure usually all other containers suffered, badly
12:39 <@rxr> but I do not want to stop you playing with it :_)
12:40 < Baldzius> :)
12:40 < Baldzius> btw we really have some issues with xorg fonts going into different location than before
12:40 < Baldzius> not sure why yet
12:40 < Baldzius> 8.0 is affected also
12:51 <@rxr> hm
12:51 < CIA-32> aldas * r37404 /trunk/package/xfce4/libxfce4menu/libxfce4menu.desc: * removed extra space from Download tag for libxfce4menu
12:54 < Baldzius> rconstruct mentioned same week or so ago
12:57 < rogermason> Hello all. I'm running into this when trying to build poppler: error: possibly undefined macro: AS_MESSAGE_LOG_FDdnl. This after doing svn up. Isn't this the same error that Baldzius' hotfix patch cured?
12:57 < Baldzius> you are on trunk, right?
12:58 < rogermason> Right.
12:58 < Baldzius> that's messy at the moment
12:58 < Baldzius> 8.0 is fine so trunk should be fine too
12:59 < Baldzius> I still need to start fresh build for trunk to see if this happens also
12:59 < Baldzius> on trunk
13:01 < rogermason> autoconf is 2.66.
13:04 < Baldzius> open /usr/share/aclocal/pkg.m4
13:05 < Baldzius> Consider adjusting the PKG_CONFIG_PATH environment variable if you
13:05 < Baldzius> installed software in a non-standard prefix.
13:05 < Baldzius> _PKG_TEXT
13:05 < Baldzius> ])],
13:05 < Baldzius> [AC_MSG_RESULT([no])
13:05 < Baldzius> $4])
13:05 < Baldzius> elif test $pkg_failed = untried; then
13:05 < Baldzius> this this from mine 8.0 build
13:05 < Baldzius> yours I believe should be slightly different
13:11 < rogermason> This is what I have:
13:11 < rogermason> _PKG_TEXT])dnl
13:11 < rogermason> ])
13:11 < rogermason> elif test $pkg_failed = untried; then
13:11 < rogermason> AC_MSG_RESULT([no])
13:11 < rogermason> m4_default([$4], [AC_MSG_FAILURE(
13:12 < rogermason> [The pkg-config script could not be found or is too old. Make sure it
13:12 < rogermason> is in your PATH or set the PKG_CONFIG environment variable to the full
13:12 < rogermason> path to pkg-config.
13:12 < rogermason> _PKG_TEXT
13:12 < Baldzius> yep
13:13 < Baldzius> change _PKG_TEXT])dnl -> _PKG_TEXT])[]dnl
13:13 < Baldzius> pkgconfig has been fixed upstream, I guess we can expect new release soon
13:19 < rogermason> This is what I have now: _PKG_TEXT])[]dnl
13:19 < rogermason> ])
13:19 < rogermason> elif test $pkg_failed = untried; then
13:19 < rogermason> AC_MSG_RESULT([no])
13:19 < rogermason> m4_default([$4], [AC_MSG_FAILURE(
13:19 < rogermason> [The pkg-config script could not be found or is too old. Make sure it
13:19 < rogermason> is in your PATH or set the PKG_CONFIG environment variable to the full
13:19 < rogermason> path to pkg-config.
13:19 < rogermason> _PKG_TEXT
13:20 < rogermason> but I still get the same error :-(. Time for breakfast.
13:24 < Baldzius> running build or compiling on host?
13:24 < Baldzius> make sure you change correct file
13:39 < koan> hm rxr have you by any chance taken a look at the docbook-tools build error I got yesterday?
14:15 < rogermason> Baldzius: my fault. I did a dumb thing: made a backup buthen edited and overwrote the backup. poppler is compiling now. Thanks.
16:33 -!- ccole [~cole@coledd.com] has joined #t2
17:53 <@rxr> re
17:53 <@rxr> maybe just another docbookx package missing?
17:53 <@rxr> sorry for the abense just got Susan from the airport :-)
17:54 < CIA-32> aldas * r37405 /trunk/package/multimedia/gnash/gnash.desc: * updated gnash (0.8.7 -> 0.8.8)
18:02 < koan> docbookx is installed
18:05 <@rxr> this was during emerge?
18:06 < koan> yes
18:06 < koan> http://pastebin.com/vFbuwzBq
18:06 <@rxr> yes, still had it in scrollback
18:19 <@rxr> in my clean ongoing trunk build the build, ... :-(
18:21 < Baldzius> haven't had this error, ever
18:21 <@rxr> Baldzius: do you know why we got this if x86-64 then autoregen junk in some gnome package?
18:21 <@rxr> I currently try a build with cleanups in this regard
18:22 <@rxr> can it be that we mostly only have this damn autojunk problems because some freaking "missing" scripts decide it is time to auto-regenate?
18:22 <@rxr> most of the "if x86-64 then regen" stuff build with this and another automatig quirk removed, ...
18:23 < Baldzius> hm, not sure
18:23 < Baldzius> I am doing fresh ref build
18:23 < Baldzius> we'll see how it goes
18:24 < Baldzius> the problem is that autogen stuff can't find needed files as we get gnome into /opt
18:24 <@rxr> yes, but this autogeren is unintended
18:24 < Baldzius> that's not an issue for 32 bits
18:24 <@rxr> this are just broken missing scripts going havoc
18:24 <@rxr> I do not see that we want them to regen, ...
18:24 < Baldzius> not sure why they want to regenerate files for 64 build
18:25 <@rxr> beside that autojunk is fragile, ever was, it wasts lots of CPU cycles during builds, ...
18:25 < Baldzius> and that's when they fail
18:25 <@rxr> this missing scripts are stupid they check some files and versions and some br0ken stupid stuff then decide "hey, let's regen with new versions"
18:25 <@rxr> I already run a build since yesterday that globally, automatically replaces the missing files with "#!/bin/true' all is fine go away :-)
18:26 -!- xbj9000 [~xbj@74.43.146.108] has joined #t2
18:26 < Baldzius> then it's super duper
18:26 <@rxr> I only once maintained a project with autojunk, it was GSMP, back some 10 years, and it was a pure waste of time
18:27 <@rxr> for all new stuff, from exact-image, to the now legacy hotplug++ and some internal things I use my own set of pure Makefiles that just work(tm)
18:28 <@rxr> I know from Lars that they even use it in some uni projects adapted them for their intern code, ...
18:41 <@rxr> it's really still warm after the rain, :-(
18:42 <@rxr> 1~some humid 25°C...
18:43 < Baldzius> in 4 years I live here we had the coldest summer this year :/
18:43 <@rxr> yeah, global warming sucks, looks like it's overall the hottest year world-wide
18:44 < Baldzius> not here apparently
18:45 <@rxr> wow, even usatoday writes that !!!
18:45 <@rxr> http://www.usatoday.com/weather/climate/2010-07-15-heat-record_N.htm
18:45 <@rxr> Baldzius: yeah, global warming always meant that some regions get cooler, ...
18:46 < Baldzius> so why they call it global :)
18:46 <@rxr> when gulf stream and friends change
18:46 <@rxr> well, overall warmer, and yet due mother earth subtle energy transfer mechanisms change, or break down completely, some regions will get colder ..
18:47 < Baldzius> anyway , we had coldest winter and coldest summer :/
18:47 <@rxr> this summer was really hot in Germany this year, very very hot
18:47 <@rxr> and the winter was indeed also very gold
18:47 <@rxr> for us in europe this can very well mean that it gets more "extreme", not as mild anymore
18:48 <@rxr> http://en.wikipedia.org/wiki/Gulf_Stream
18:49 <@rxr> (beside rising sea levels, ...)
18:50 -!- snowwoolf [~fedorowp@CPE0013f7acc3c0-CM0013f7acc3bc.cpe.net.cable.rogers.com] has joined #t2
18:53 <@rxr> ah, a little fresh air coming thru the 2 fully open windows, ...
18:53 <@rxr> maybe I should leave the servers alone and go home :-(
18:55 -!- snowwoolf [~fedorowp@CPE0013f7acc3c0-CM0013f7acc3bc.cpe.net.cable.rogers.com] has quit [Quit: Leaving.]
19:31 <@rxr> cu later then
19:33 < Baldzius> bye
20:26 -!- Baldzius [~aldas@87-198-133-94.static.ptr.magnet.ie] has quit [Quit: Leaving]
21:41 -!- jkoerner [~jkoerner@p5B06F689.dip.t-dialin.net] has joined #t2
21:52 -!- jkoerner [~jkoerner@p5B06F689.dip.t-dialin.net] has quit [Quit: jkoerner]
--- Log closed Tue Aug 24 00:00:49 2010