/srv/irclogs.ubuntu.com/2011/05/21/#ubuntu-devel.txt

bluefoxicy 1531 root      20   0  350m 166m  11m R   85  4.5   0:27.97 update-apt-xapi00:45
bluefoxicywhat is this process00:45
cjwatsonupdate-apt-xapian-index00:49
bluefoxicycjwatson:  it eats a LOT of CPU, for reasons I'm not sure of.00:52
=== asac_ is now known as asac
=== jtechidna is now known as JontheEchidna
=== MTecknology is now known as EvilMTeck
=== _LibertyZero is now known as LibertyZero
lucidfoxbluefoxicy, I also ran into that problem06:31
lucidfoxwith update-apt-xapian-index hogging CPU06:31
ohsixslow drive / lots of files? it's ran periodically and niced; it will use some cpu to do its work06:33
lucidfox"Some" CPU? More like 99%07:07
lucidfoxnot on this computer, on my work one07:07
lucidfoxAlso, what's with the build failures?07:08
lucidfoxactually, wait07:09
lucidfoxhrm07:10
lucidfoxhttps://launchpadlibrarian.net/72114285/buildlog_ubuntu-oneiric-i386.steadyflow_0.1.6-0ubuntu1_FAILEDTOBUILD.txt.gz07:10
lucidfoxany idea why this is failing?07:10
lucidfoxThe following packages have unmet dependencies:07:10
lucidfox libnotify4-dev : Depends: libnotify4 (= 0.7.2-0ubuntu2) but it is not going to be installed07:10
lucidfoxoh right, libnotify4 was removed from oneiric while the package was in NEW? o_O07:15
ohsixlucidfox: run it manually and see what it's doing07:29
ohsixif it isn't i/o bound i don't see why it wouldn't use the cpu available07:30
lucidfoxWell, I fixed it already. It failed because libnotify4-dev was not completely removed from oneiric yet, and the buildd never even considered libnotify-dev07:30
lucidfoxI've removed libnotify4-dev and left only libnotify-dev07:30
ohsixinteresting07:32
lucidfox...oh what now... dput is not accepting my signature08:09
lucidfoxHuh, but it did upload08:12
lucidfoxdput complained on .changes about "verification failed 3 times (no public key)", but then I received a mail saying the package did get accepted08:13
=== Quintasan_ is now known as Quintasan
Laneylaunchpad bug08:37
pittihallyn: libcap2> getting it through autosyncs is deal IMHO09:56
pittibroder: ah, thanks for confirming; that was my impression, too09:56
pittiSpamapS: ok, will do a round of SRU09:57
cjwatsondoko_: may I merge eina?09:59
doko_cjwatson: sure, merge anything you can't resist =)10:00
cjwatsonheh, just trying to reduce the FTBFS list a bit10:00
apwanyone else seeing issues with uploads to ubuntu ?11:14
cjwatsonapw: if it's just verification-failed messages, ignore them11:15
cjwatsonit's a known LP bug11:16
cjwatsonyour upload was successful11:16
cjwatson(I checked)11:16
apwcjwatson, ahh thanks11:17
cjwatsonoh, for goodness' sake - eina is failing a test of some silly joke function11:38
cjwatsonand doesn't fail on the porter box11:38
=== TheDaniel0108 is now known as Daniel0108
bluefoxicydoes Cool'n'Quiet work in Ubuntu?16:00
bluefoxicyI heard you need powernow-k8 loaded in the kernel16:00
bluefoxicywhich is compiled in16:01
hvis appmenu not working properly at the moment (oneiric)?16:11
=== JasonO is now known as Rapture
bluefoxicyyay16:56
bluefoxicymy 31 line bash script to monitor my CPU temperature works!16:56
bluefoxicynow if my computer is at risk of overheating, it'll send SIGSTOP to a bunch of stuff and then wait for it to cool down.16:56
ionHow about fixing the cooling instead? :-)16:59
bluefoxicyion:  I tried that, too :)16:59
bluefoxicyi suggested also that maybe a small daemon that throttles back the system in interesting ways when the threat of overheating is present would be useful, and people were vehemently against the idea of software fixing a hardware problem17:00
ionI should do that for my laptop. The thermal grease probably needs reapplying.17:00
bluefoxicyThe discussion kind of broke down after I recommended to the entire mailing list to remove everything tied to "fsck" since the real solution is to fix your disk/power source17:00
bluefoxicythere's also that silly checksum thing in the kernel, that checks if dirty blocks are corrupted (due to bad RAM) before writing them back to disk; again, the real solution is to replace the faulty RAM ...17:01
bcurtiswxi have a 32 bit build machine, is there any way to get it to build a 64 bit debian pacakge ?23:30
penguin42I don't see why you can't cross build it23:31
bcurtiswxso how do I do that exactly?23:33
slangasekit's quite non-trivial today23:33
bcurtiswxslangasek, hmm, is there a wiki page I should be referencing?23:33
penguin42bcurtiswx: http://wiki.debian.org/BuildingCrossCompilers is one23:34
penguin42bcurtiswx: Then there are various bits of magic to actually build the packages - none of them are particularly pretty23:34
penguin42bcurtiswx: xdeb/xapt is the one I've used23:34
penguin42bcurtiswx: And it doesn't work for all packages23:34
bcurtiswxpenguin42, OK thanks for the links/help :)23:34
penguin42bcurtiswx: Good luck23:35
slangasekin theory, an x86_64-linux-gnu-gcc wrapper that simply calls 'gcc -m64 "$@"' would be sufficient, provided that gcc-multilib is installed23:35
cjwatsonFWIW xdeb and xapt are entirely separate projects23:35
slangasekbut actually getting the 64-bit dependencies where you need them requires something like xdeb at the moment, yes23:36
penguin42yeh? Oh i'd assumed xapt used xdeb23:36
directhexdo they work?23:36
penguin42directhex: Depends on their mood23:36
bcurtiswxi thought maybe pbuilder had a special way to create a 64bit build env23:37
cjwatsonno, both xdeb and xapt use dpkg-cross, but they have different philosophies23:37
cjwatsonultimately, we hope both will be replaced by stage<whatever> multiarch23:37
slangasekbcurtiswx: the only 64bit build env you can create with pbuilder per se is a *native* 64-bit env; if your machine (or your kernel) is 32-bit, that won't be usable23:38
directhexyou can do 64-bit kernels on 32-bit debian. but afaik not on ubuntu23:38
bcurtiswxslangasek, yeah that makes sense.  Was worth a thought at least23:38
cjwatsonthough, I mean, it's basically the same idea, grab all the dependencies, cross them, install them, build23:38
penguin42cjwatson: I don't think I'd succeeded in getting xdeb to fetch dependencies for me23:39
cjwatsonxdeb tries to do more work for sequences of builds and is more willing to apply heuristics to get there23:39
slangasekdirecthex: well, you can in that you can cross-install an amd64 kernel on your system using multiarch ;P23:39
cjwatsonit worked for me when I wrote it, but it's been a while since I did any work on it :-)23:39
cjwatsonit's wookey's problem nowadays23:39
bcurtiswxinformation overload here :) hehe, i appreciate the help tho :)23:40
penguin42bcurtiswx: What was the package out of interest?23:41
bcurtiswxpenguin42, i've rebased seahorse 3.0.0 with debian, it built and I wanted to test it on my 64 bit machine.. i was hoping there was a "fairly easy" way to cross build, but my laptops quite powerful and i'll just build it on here :)23:44
bcurtiswxi don't have upload, so i wanted to really verify this package was built well before i went requesting a merge23:44
penguin42oh I wouldn't give that a good chance of crossbuilding23:44
penguin42but I'm a pessimist23:44
bcurtiswxhaha, hence my building on my 64 bit laptop23:45
cjwatsonbcurtiswx: you can use a PPA too, if you like23:45
bcurtiswxcjwatson, yup, i have empathy 3.1 on that for interested parties to test, and if i didn't have the immediate availability to this laptop i would23:46
lfaraonecjwatson: so in https://launchpad.net/ubuntu/+spec/foundations-o-apt-mirror-method, we'd just use mirror:// rather than having a mirror picker?23:47
cjwatsonlfaraone: yes23:47
* bcurtiswx waves to lfaraone23:47
cjwatsonor at least rather than requiring a mirror picker - we're not out to destroy people's choice, just to make things work better without everyone having to choose23:47
lfaraonecjwatson: but would software-properties still have a static list, or would it source its list of mirrors from the mirror method?23:48
cjwatsonI would hope that it wouldn't matter any more23:48
cjwatsonbut you'll have to ask mvo, I don't know the details of software-properties23:48
cjwatsonI'm mostly involved in that spec to ensure that the installer is synced up23:48
* penguin42 hopes it requires a signed reply to the list of mirrors23:48
lfaraonepenguin42: why would that matter?23:49
lfaraonepenguin42: if I can MITM the mirror list, I can MITM whichever mirror you select anyway.23:50
penguin42yeh I guess so23:50
lfaraonethat's why we have http://wiki.debian.org/SecureApt23:50
cjwatsonI suspect, without having asked, that mvo is no great fan of having to keep a static list of mirrors; but we don't have a specific work item for replacing it23:51
highvoltagejbicha: congrats on the membership! I thought you were one already :)23:54
lfaraonehey bcurtiswx23:55
lfaraonecjwatson: hmm. well, I could do that, I guess. If the mirror:// list was defined somewhere, I could just add a hook to software-properties to poll it whenever you go to change the mirror.23:55
=== dendro-afk is now known as dendrobates
jbichahighvoltage: thank you, I should have applied sooner23:59

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!