dholbachgood morning07:00
ajmitchmorning dholbach07:03
dholbachhi ajmitch07:03
ronin___hi dholbach07:30
dholbachhi ronin___07:30
ronin___morning iulian07:31
hakermaniatumbleweed, thanks for the review. About the too many system calls, could we avoid it? I think that we only call it so as to set the desktop background, call the music players and to delete things08:09
hakermaniaMaybe the last one could be "C++azed"08:09
hakermaniabut the others seem fine, to me :/08:09
jtayloryou should be able to control everything over dbus08:20
hakermaniajtalor, I think im  good sending dbus messages, xD. Any other info about this? Should I send messages to the system so as to tell it what to run?08:21
hakermaniajtaylor, .08:21
tumbleweedhakermania: I was just suprised to find a C++  program that was half-shell08:36
tumbleweedhakermania: they could all handle errors a bit better (use && instead of ; or use set -e)08:37
tumbleweedhakermania: btw, the sounsd aren't included in the binary package08:37
hakermaniatumbleweed, thanks. I was surprised to find out that the build didn't work for you. Also I thought that the watch file was looking in a specific folder of sourceforge where i've placed the source. But it seems that it looks for the file that sourceforge considers as last08:45
hakermaniaAnd I don't know how to change it to the real last source (V2.0 GNOME 3)08:46
tumbleweedhakermania: the problem with the watch file is that you're matching .* as a version. maybe you should use [0-9.]+08:59
hakermaniatumbleweed, thanks. About the sound files, why aren't they included into the binary package? I mean they are into the DEB file and when zi did a complete unistallation of the program and re-installed through the DEB the sound files where there09:01
tumbleweedhakermania: they aren't in the deb I built09:01
tumbleweed"debc" will show you what's in the deb that just built09:01
hakermaniatumbleweed, I thought you said "Package doesn’t build", how did you manage to build the DEB? Also in what environment are you building? This source doesn't build in Natty but in Oneiric it does09:02
tumbleweedhakermania: by overriding dh_auto_install, as I mentioned in the comment09:02
tumbleweedoneiric, of course, hang on, I'll give you a build log09:02
jtaylor:( pypy does not work in oneiric anymore :( needs libffi.so.509:07
jtaylorand my machine is not powerful enough to rebuild it09:07
tumbleweedhakermania: your 'make install' calls 'make clean' That could be problematic :P09:07
* tumbleweed pokes a pypi developer09:08
tumbleweederr pypy09:08
hakermaniatumbleweed, I am a bit confsed with the clean step. Should it done at the end or at the begginning? In my case it is done in the beginning...09:09
hakermaniaAlso, you said I should run 'make clean' before uploading or something, while the source seems clean (no .o files, no binary etc)09:09
hakermaniatumbleweed: Fetched 74.8 MB in 3s (21.3 MB/s) o.O?09:09
tumbleweedlocal mirror09:10
jtaylorapt-cacher-ng ftw :)09:10
tumbleweedhakermania: the source package includes some foo~ files09:10
hakermaniatumbleweed, I hate gedit -_-09:11
tumbleweedclean is what you do when you're done and want to build everything from scratch again. you probably don't want to do it inbetween building and installing09:11
hakermaniatumbleweed, Ok, I'll take care of it. But now im more confused :/ you said that dist-clean should remove debian/wallch* and debian/files etc plus wallch (the exectuable), so I assume you need a total clean, leaving the source as it was initially09:13
hakermaniaSo, why not having the 'make clean'?09:13
tumbleweedhakermania: no I didn't say that09:13
tumbleweedI said you don't need to remove those manually in debian/rules, distclean will take care of the upstream parts, and debhelper will take care of the debhelper parts09:14
tumbleweedyou don't need your clean rule in debian/rules at all09:14
hakermaniaYes, but apparently (at least in my case), distclean leaves the wallch untouchable and also the debian/wallch* and debian/files files are not removed. So, somebody's not doing its job well (i assume me xD)?09:14
tumbleweedhakermania: no I don't mean *you* must run distclean09:15
tumbleweedI mean running debian/rules clean (without your clean rule in it) should do the trick, because debhelper will clean up after itself and will run make distclean09:15
=== almaisan-away is now known as al-maisan
hakermaniatumbleweed, yes but it doesn't09:16
tumbleweedit does for me09:17
tumbleweedif you remove the clean rule from debian/rules09:17
hakermaniatumbleweed, oh you mean that it finds that a clean rule exists and so it doesn't run its clean rules?09:17
tumbleweedyes, you are stopping "dh clean" from running09:18
hakermaniaOk thanks :)09:18
tumbleweedjtaylor: actaully, debmirror. I can live with being a bit out of sync, although some people call that herecy :P09:19
hakermaniatumbleweed, I found out why the source didn't build for you. You need to run 'sudo debuild -S', because during installation I call update-desktop-database which needs root09:22
hakermaniaI do that to associate the new mime type with the desktop file09:23
tumbleweedhakermania: that would be wrong09:23
hakermaniatumbleweed, it's the only way I think09:23
tumbleweedwe don't need the buildd to have an up to date desktop file database09:23
hakermaniatumbleweed, without this rule the DEB file doesn't associate the desktop file with the mime type09:24
tumbleweedhakermania: you're missing my point09:24
hakermaniaYes probabl09:24
tumbleweedhakermania: update-desktop-database updates the database of the machine you run it on, not the machine you install the deb on09:24
hakermaniatumbleweed, can't we assume that it would be the same machine? If not, how would I make it work :( ?09:25
tumbleweedit's pretty safe to assume it isn't the same machine. I don't think our build daemons have desktop backgrounds that need changing09:25
tumbleweedhakermania: but guess what. dh_installmime detects that you are installing a mimetype, and adds "update-mime-database /usr/share/mime" to the postinst script that gets run when your deb installs09:27
hakermaniatumbleweed, i think that's not enough09:30
hakermaniathat updates the system's mime lists but does not associate it with the desktop file09:31
tumbleweedhakermania: then add *that* to the postinst09:32
hakermaniatumbleweed, should I edit a debian/^ file to do so?09:33
hakermaniatumbleweed, in the current status, 'uscan --report' shouldn't output anything, right?09:46
tumbleweedyes. With a --verbose it should tell you you have the latest version09:47
hakermania-_- http://qa.debian.org/watch/sf.php/wall-changer/ failed: 500 Can't connect to qa.debian.org:80 (Bad hostname 'qa.debian.org')09:49
philipballewQUESTION: there is software in the current ubuntu repos that has updated versions but the version in the ubuntu repos is from 08. can this be fixed?09:49
* philipballew is willing to fix it to09:52
tumbleweedhakermania: that sounds like a problem on your end :)09:52
hakermaniatumbleweed, you mean its my ISP's fault? Also, when do I call libglib2.0-bin through system()?09:53
tumbleweedphilipballew: of course it can. The question is how important is it to update these packages (we are in feature freeze) and who maintains them (why hasn't it been updated yet)09:53
jtaylorphilipballew: which software?09:53
tumbleweedhakermania: gsettings?09:54
hakermaniatumbleweed, I see. What about the 1st question? My browser connects to qa.debian.org normally09:54
tumbleweedmaybe it was a spurious error then09:55
tumbleweed(your ISPs DNS resolvers just suck)09:55
hakermaniaok thanks :P09:55
philipballewjtaylor, kismet09:56
jtaylorphilipballew: the package seems orphaned in debian: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=53488409:58
ubottuDebian bug 534884 in kismet "new version and core of kismet" [Normal,Open]09:58
philipballewjtaylor, i see. so basically noone is packaging it anymore?09:59
jtaylorlast upload of that maintainer was in 200809:59
jtaylorit probably needs a new maintainer09:59
philipballewi see. well it appears jtaylor that the kismet team is making their own deb's for people and their own repo for people to add for their sources list10:00
philipballewif a maintainer is needed I can be that, but i am really bad at packaging. could learn, but whatever10:01
hakermaniatumbleweed, as for the copyright, should I just change the 'Format' field to the latest? (http://anonscm.debian.org/viewvc/dep/web/deps/dep5.mdwn?revision=174)10:01
hakermaniaBecause as I see there aren't any major format changes...10:02
tumbleweedhakermania: fine, I was just pointing it out10:03
jtaylorphilipballew: if you want to do that you should ask in #debian-mentors on irc.debian.org for advice, maybe also ask fourmond@debian.org who did two non-maintainer uploads recently if he has plans with this package or contact to the original maintainer10:04
hakermaniatumbleweed, I just think that the url I provided above isn't the correct one. I mean it does point out to the latest format but the page doesn't look like this: http://svn.debian.org/wsvn/dep/web/deps/dep5.mdwn?op=file&rev=15210:04
hakermaniait is the previous format I used. And replacing 152 with 174 returns error10:05
tumbleweedhakermania: we know about that, don't worry it was the correct one10:05
hakermaniaOk thanks10:05
hakermaniatumbleweed, what's the problem in installing to /usr/share?10:10
tumbleweedhakermania: please elaborate?10:12
hakermaniatumbleweed: "The upstream makefile probably should be installing to /usr/local not /usr/share"10:15
jtaylor /usr/share is package manager land, trespassers will be shot10:16
tumbleweedthat's not an issue for this package, it's an issue for people who install it from source10:17
hakermaniatumbleweed, so, source should install to /usr/local/ and package to /usr/share/? How can I achieve this? For somebody who builds from source it's like doing the same thing with building the deb but actually installing it immediately to his system (copying files etc immediately)10:21
jtayloryour makefile should support DESTDIR, default to /usr/local and the package overwrittes it to /usr10:22
tumbleweedyou can achive this by making your upstream makefile use /usr/local as PREFIX. dh_auto_configure will pass PREFIX=/usr to qmake10:23
tumbleweedjtaylor: this is qmake, so it uses INSTALL_ROOT rather than DESTDIR. Also, aren't you mixing up DESTDIR and PREFIX?10:23
hakermaniaWait, and then through the code how would I call a sound file if I don't know where it is located?10:24
jtaylordon't know qmake, but in normal Makefiles DESTDIR is the same as PREFIX10:24
tumbleweedhakermania: you know at build time, because you can use PREFIX10:25
tumbleweedjtaylor: PREFIX is where it will be installed, DESTDIR diverts the installation (e.g. into debian/tmp)10:25
jtaylorok then I mixed them  up10:25
hakermaniatumbleweed, I think i'm missing something. Let's say that somewhere in the code it says 'system("canberra-gtk-play /usr/share/wallch/a.ogg");10:26
hakermaniaHow am I supposed to change it if user builds from source?10:27
hakermania(and the it will be /usr/local not usr/share10:27
tumbleweedyou'd play PREFIX + "/share/wallch/a.ogg"10:27
hakermaniaWhere PREFIX would be /usr or /usr/local, got it10:28
hakermaniaDoes qmake knows which prefix to use each time :/ ?10:35
* hakermania feels a bit noob10:36
tumbleweedexport DH_VERBOSE=1 and you'll notice that qmake is called with PREFIX=/usr by dh10:36
tumbleweedobviously users installing at home wouldn't do that, so it would use the default (which you should make /usr/local)10:37
hakermaniatumbleweed, so, to make it clear, the prefix that dh uses is '/usr' no matter the default (being '/usr/local'). Also, the default should be specified during the creation of makefile (qmake PREFIX=/usr) or inside the project file? I suspect the latter10:42
tumbleweedinside the project file10:43
hakermaniatuumbleweed, it's the first time I use the PREFIX variable. Do I call it straightly in my program as you told previously (PREFIX+"/share/etc)? Also guys on #qt doesn't seem to know anything about PREFIX :P11:04
tumbleweedhakermania: read the gnu coding standrds (that jtaylor linked to). You'll also need to modify your qmake project to make it pass -DPREFIX=$PREFIX to g++11:08
* hakermania is something like http://bit.ly/qYeDvs11:11
=== al-maisan is now known as almaisan-away
hakermaniatumbleweed, how does the dh overrides the default PREFIX?11:32
tumbleweedhakermania: 12:36 < tumbleweed> export DH_VERBOSE=1 and you'll notice that qmake is called with PREFIX=/usr by dh11:34
hakermaniatumbleweed, I think I've found it!!!! It should look like this inside the makefile: http://paste.ubuntu.com/669150/ doesn't it????11:47
Laneywho has a lucid i386 installation on bare metal and wants brownie points?11:48
tumbleweedhakermania: yes, you need to make qmake generate that11:49
tumbleweedLaney: only have a lucid i386 kvm vm11:49
Laneynah, not tricksy enough, cheers though11:50
tumbleweedLaney: actually, I do have one of those11:51
Laneyyeah? fancy test building mono from oneiric? :-)11:51
hakermaniatumbleweed, I did it!!!!11:52
tumbleweedshould it build with nothing special?11:52
Laneyan oneiric chroot should be fine11:52
Laneydon't know about the backport11:52
=== almaisan-away is now known as al-maisan
Laneynobody can reproduce this ftbfs, but i don't think any of us have a similar enough setup to the buildds11:53
Laneyvms are no good, oneiric kernel is no good11:54
Laneyi386-on-amd64 is no good11:54
jtaylorsqueeze kernel is no good11:55
ajmitchi386 kernel with i386 oneiric vm on amd64 host is no good11:55
ajmitchwe've been having a fun time building11:55
Laneywoe is the debian mono group11:55
* tumbleweed is debootstrapping...11:56
ajmitchit's a sad day when we're disappointed that the build works fine11:56
Laneyso if you can reproduce, it's bisectin' time11:56
* tumbleweed should probably ionice this a bit, it *is* a production fileserver, but I don't think anyone will notice11:57
ajmitchproduction? who cares about that, packages are on the line here!11:57
tumbleweedit's for an unmaintained lab with 5 year old hardware, if they can log in, that's better than average11:58
Laneythere won't *be* a production if we don't fix this!11:58
nigelbLaney: lucid i386?11:59
nigelbI do11:59
hakermaniaShould I have the Makefile in the orig.tar.gz?12:00
Laneynigelb: you can also try rebuilding mono from oneiric if you want :-)12:00
nigelblp-pull-source and put it into a pbuilder?12:00
Laneypull-lp, yeah12:01
nigelbdo I need to make any changes?12:01
nigelbtrying now12:01
nigelbneed to find if  still have a lucid pbuilder though12:01
Laneyif it's pbuilder-oneiric12:01
ajmitch2.10.4-2 is what failed on a couple of buildds12:01
Laneywe want oneiric guest, lucid host12:01
nigelbmy pbuilder is probably lucid pbuilder12:02
ajmitchupdate it!12:02
nigelbput it throgh an oneiric pbuilder?12:02
nigelbalso, source package name12:02
* ajmitch hopes it's not stupidly hardware dependent somehow12:03
nigelbgetting source12:03
nigelbanyone seen persia around lately?12:03
ajmitchsee you in a couple of hours12:03
tumbleweednigelb: not since debconf12:03
nigelb*facepalm* ssh'd into a machine and typed up arrow for byobu. The last command was reboot.12:04
ajmitchwell done :)12:04
tumbleweedbetter than halt :)12:05
Laneydidn't it need a password?12:05
nigelbno passwd12:05
nigelbor whatever that settign is in /etc/sudoers12:05
Laneyssh root@...12:05
tumbleweedLaney: how long until expected build failure?12:05
Laneydepends how fast your hardware is12:05
Laneyit was about 40 minutes on the buildd12:06
tumbleweeddual-xen 3Ghz, 2G RAM12:06
Laneybuilds a load of native stuff, then mcs [basic] stuff, then [net_2_0], then [net_4_0], then mdoc which is where it fails12:06
tumbleweedok, it's building, I'm going to go out for a bit12:07
nigelbdammit, I have so much source code lying around that I have no more space.12:09
ajmitchmono will take a fair bit12:10
hakermaniaOk, about the code now, you told me that I could solve the system calls through sending dbus messages. How-to?12:10
ajmitchabout 1.1GB used on my clean VM's ~ so far12:10
nigelbFirefox source + compiled code eats my space.12:12
sladennigelb: reboot & history -r12:21
nigelbsladen: heh, I just typed it out manually instead of arrows12:26
nigelblater, I'll find who did the exit instead of Ctrl + D :)12:26
nigelbeverytime, I see the gpg key password dialog box12:39
nigelbI remember UDS lighting talks and james_w :)12:39
hakermaniatumbleweed, OK, now that I made qmake take PREFIX, how do I use it inside my code? What is it? char?13:30
hakermaniaShould I define it?13:30
tumbleweedhakermania: you are defining it with -D13:32
tumbleweedso yes, it' sa string instrted by the preprocessor, you can do "foo " PREFIX "/share/bar"13:32
tumbleweedLaney: my mono has got as far as running make check13:37
Laneytumbleweed: alright, that's too far then13:37
Laneyyou can't reproduce it either, thanks for trying :(13:38
Laneyshall we ask that guy for a shell yet?13:41
hakermaniatumbleweed, what about the 'uninstall' scipt? Should it check if wallch is installed in /usr/share or /usr/local/ or automatically trying to remove from /usr/local/ ?13:41
tumbleweedhakermania: it'll only be run by people who installed from source, so it sholud only remove from /usr/local13:43
hakermaniaGOT IT13:45
hakermaniatumbleweed, as for the pictures, the monitor comes from a site with NO WARRANTY and we've send some emails to request for the other images (i assume we should have done so before using them :P)13:46
hakermaniatumbleweed, also: error: ‘PREFIX’ was not declared in this scope13:47
tumbleweedhakermania: then you aren't declaring it with -DPREFIX=whatever13:48
tumbleweedhakermania: "NO WARRANTY" does not equate to claming copyright for it and distributing to others under the GPL13:48
hakermaniatumbleweed, http://www.clker.com/disclaimer.html13:54
tumbleweedhakermania: that does say that everyone uploading to the site certifies their artwork is in the public domain. You should probably state where you got it from in your licence information13:59
hakermaniaNo I don't think so14:01
hakermaniatumbleweed, I made PREFIX to work but it seems like a straight /usr/local inside the program... For example "a" + PREFIX + "b" returns: usr/local was not declared in this scope14:03
tumbleweedhakermania: you don't think so?? you probably want to define it with quotes around the value.14:08
hakermaniatumbleweed, I already have (DEFINES +="PREFIX=\"/usr/local\"") and I get the same errors. And yes, I think that the site is quite clear that anybody can do anything with its pictures unless it has to do with porn or breaking a law14:15
tumbleweedhakermania: and the way you avoid archive-admins / ftp-masters asking the same questions I did is by including the necessary informationn in debian/copyright.14:18
hakermaniaOk, so should I make debian copyright say Files: Pictures/monitor.png and then License? What?14:21
tumbleweedDEP5 explains how you should handle public domain14:22
AnAntdholbach: Hello, thanks for the ACK.14:49
AnAntdholbach: you've been away for a while ?14:49
dholbachAnAnt, yes, I was on holidays for 2 weeks :)14:49
AnAntdholbach: note, I will just fix the title of #819692, it is -11 to be sync'ed not -1014:50
dholbachsounds good14:50
AnAntdholbach: do I really have to comment on 812120 ?14:57
AnAntdholbach: I mean , the changelog speaks for itself14:57
LaneyWhy should we break the freeze for this version — i.e. does it fix any bugs that are present in Ubuntu currently?14:58
AnAntLaney: not that I know of15:00
dholbachI talked to Colin about it and it was his suggestion - I can see why somebody would be a bit hesitant to update a package which virtually every other package build depends on :/15:00
AnAntwell, I filed it before freeze actually, but since we are after freeze, the changes in changelog might be scary15:01
=== al-maisan is now known as almaisan-away
=== bdrung_vacation is now known as bdrung
tumbleweedcyphermox: I just got one of your usb-modeswitch segfaults16:16
cyphermoxtumbleweed: yeah16:16
tumbleweedthe bug's still marked as incomplete, need anything tested?16:17
=== JanC_ is now known as JanC
cyphermoxtumbleweed: does the workaround at the end work for you? and can you reproduce and see where it crashed if you were to run this in ltrace or something?16:37
cyphermoxI'd really like to understand why it's crashing, which is what I don't have now, just a change in how the file handles are redirected, and I don't understand why it makes it work16:38
tumbleweedcyphermox: the workaround seems crazy to me, esp as that code path isn't involved16:40
tumbleweedcyphermox: it doesn't segfault under strace (loops through something big, forever), but it does under ltrace16:41
cyphermoxas run manually or though the udev script?16:42
tumbleweedrun manually16:42
tumbleweeddamn, it just worked16:42
cyphermoxsee, the problem is no matter how much I try it always works for me, except the first time I tried to reproduce it; at which point I obviously had nothing ready to catch where the bug was ;)16:43
cyphermoxif your system is looping through things for a long while, your 3g dongle likely needs scsi attributes; which is a little less well tested.16:44
cyphermoxbut if you can give me some piece of info more than just the segfault message from syslog (I can't parse that, unfortunately), then it's going to be a win already :)16:45
tumbleweedyeah, let me spenda little more time on it (and grab some debug symbols)16:45
cyphermoxeven just a the usb_modeswitch.log from /var/log if you enable logging in /etc/usb_modeswitch.conf first, is good16:46
tumbleweedcan't trigger it if I enable logging :/16:49
cyphermoxtumbleweed: maybe that's why I can't reproduce it either16:53
cyphermoxdo you have the ltrace output where it failed?16:54
tumbleweedcommented on the bug, but it's only ltrace, no output from usb_modeswitch_disptatcher16:55
tumbleweedyeah if I try to catch stdout and stderr, it doesn't segfault16:56
cyphermoxso it's got to be something about file handles17:09
tumbleweedcyphermox: got it in gdb, see bug17:10
cyphermoxamazing, you rock17:11
tumbleweedrigh,t now I can go to pub quiz and leav you to fix it :)17:13
cyphermoxfound two issues with the data you gave there17:15
=== medberry is now known as med_out
lfaraonemicahg: myself, I'd prefer to just  have the new upstream version packaged and SRUed, but I think that's about a snowball's chance in hell of happening.18:34
micahglfaraone: I think SRUing version 2 would be better, even though it has new features as the current version is totally broke, but IANA release team member18:34
micahglfaraone: no, if the current version worked it would be a problem, but it doesn't :)18:35
lfaraonemicahg: hm. I'll repeat my offer, want to maintain another package in Debian? :)18:36
micahglfaraone: I'm still working on the first :), I have no inherent interest in blogtk, there seems to be interest in UBuntu for it though as the original bug for deprecating python-gtkhtml keeps getting a ping on blogtk18:37
micahglfaraone: but if the package isn't too bad, maybe I can get the ball rolling on it18:38
lfaraonemicahg: I've seen more lintian warnings. https://launchpad.net/~jayreding/+archive/ppa/+files/blogtk_2.0%7Eppa6%7Elucid.dsc18:41
=== med_out is now known as med
=== med is now known as medberry
hakermaniatumbleweed, DEP5 says that "If a work has no copyright holder (i.e., it is in the public20:03
hakermania     domain), that information should be recorded here."20:03
hakermaniaAnd it refers to the Copyright, field20:04
hakermaniaBut how?20:04
hakermaniaShould we include the site's name?20:04
hakermaniaAlong with our names?20:04
hakermaniaAlso what about the straight PREFIX?20:12
=== yofel_ is now known as yofel
=== almaisan-away is now known as al-maisan
=== al-maisan is now known as almaisan-away
=== almaisan-away is now known as al-maisan
=== al-maisan is now known as almaisan-away
=== medberry is now known as med_out
=== Amaranthus is now known as Amaranth

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