/srv/irclogs.ubuntu.com/2013/11/26/#ubuntu-devel.txt

maxbThere probably isn't a place in launchpad to report bugs with http://people.canonical.com/~ubuntu-archive/madison.cgi00:01
maxbAlternatively, rmadison could fix up the columns client side00:02
twbMm, what puzzle[sd] me is that it's only the ubuntu side00:07
tumbleweedtwb: ubuntu is running madison-lite (which is a package in the archive)00:09
twbOK.  I'll see if I can remember how to use launchpad00:10
twb...and report it against that.00:10
infinitytwb: Oh hey, I've never noticed the difference in prettiness.  We could certainly fix that.00:17
twbCool.00:18
twbpiping into column -ts'|' doesn't DTRT because of the spaces on the LHS of the release name00:18
twbI sent the bug to malone00:18
twbBuuuuut I fat-fingered the package name.  Oopsie.00:19
infinityHrm.  I was going to look at what raw 'dak ls' output looks like, but ries seems... Down?  Or something?00:19
infinityIrksome.00:19
twbinfinity: dunno.  alioth was sick last week00:19
twbMight still be a little sick00:19
infinityYeah, the two don't relate at all.00:19
=== freeflying_away is now known as freeflying
infinitytwb: FWIW, your bug is https://bugs.launchpad.net/ubuntu/+source/madison-lite/+bug/31583300:24
ubottuUbuntu bug 315833 in madison-lite (Ubuntu) "Ubuntu’s rmadison should have wider column defaults" [Low,Triaged]00:24
twbCool00:25
twbI stupidly did affects/rmadison instead of devscripts00:25
infinitytwb: Where's your bug, I'll dupe it to this one.00:26
twbinfinity: it didn't get created AFAIK.00:26
infinityOh.00:26
infinityKay.00:26
twbhttp://sprunge.us/KLMG is the malone error message00:27
twbI got confused by you saying "this is your bug" ;_)00:27
twbI was also about to add "how curious that the last correspondent also picked the same options as me" until I looked at the timestamp :P00:29
infinity*laugh*00:29
twbAnyway, it looks like this is all under control.  Thanks for your help.00:29
=== freeflying is now known as freeflying_away
=== freeflying_away is now known as freeflying
=== philipballew is now known as philip
pittiGood morning05:57
=== Logan_ is now known as Guest61371
=== Logan__ is now known as Logan_
pittiStevenK, wgrant: can we enable https://translations.launchpad.net/ubuntu/trusty/+language-packs early in the cycle?06:27
wgrantpitti: With the usual schedule (ie. each series back one level)?06:30
pittiwgrant: yes, that WFM; we don't need to produce saucy updates twice a week any more06:31
pittiwgrant: in fact, I think we could/should even switch quantal/raring to "on demand", it's very unlikely that we'll ever do an official update06:31
Mirvany core-dev up for acking a packaging change from cu2d? http://pastebin.ubuntu.com/6477621/ - because upstream touched debian/changelog manually, it now misses the mention of "Adds a test suite for the libupstart-app-launch code."06:46
Mirvotherwise looks ok, although no explanation why as part of re-enabling zeitgeist the -Wno-error=unused-function was added06:46
MirvI checked also that the new tests were really run as part of the build in daily-build PPA06:48
pittiMirv: LGTM06:53
Mirvthank you06:55
=== doko_ is now known as doko
BrotherBrickmorning07:56
=== iahmad is now known as iahmad|afk
dholbachgood morning08:42
DktrKranzinfinity: ries is no more, now you want to login to coccia.d.o09:11
=== iahmad|afk is now known as iahmad
mlankhorstlooking for sru team https://bugs.launchpad.net/ubuntu/precise/+source/libdrm/+bug/125304109:49
ubottuUbuntu bug 1253041 in libdrm (Ubuntu Precise) "lts-saucy enablement in precise" [High,In progress]09:49
mlankhorst:P09:49
itsme_Hello everyone,  When I clicked on my Ubuntu power button on the top-right corner, a user account by the name of "J Random User" appeared on the list of users . I restarted my computer and now the user account has disappeared. Should I be worried that my computer has been compromised by someone/something, or is there a logical explanation for this?09:55
itsme_I'm currently using Ubuntu 13.1009:56
pittiitsme_: hm, my initial thought was "this is a test string in indicator-session", but it isn't10:08
pittiitsme_: was it exactly "J Random User"? (capitalization, periods, etc.)10:08
itsme_It was exactly "J Random User"10:09
itsme_pitti: No periods or comma10:10
pittiitsme_: I cannot find this string anywhere on my system, so no immediate idea; you can check "last" for some implausible logins10:11
lotuspsychjepitti: i also sugested him a peak at auth.log, nothing there neither10:12
itsme_pitti: Shll I paste (pastebin) the output of the command "last"? Nothing unusaull in here10:13
pittiitsme_: won't help then10:13
=== freeflying is now known as freeflying_away
diwicpitti, hi, just a quick question - do you know if a dbus activated service (on the session bus) can talk to the GUI? The idea is to be able to pop up a dialog without having a daemon running all the time just for that10:17
pittidiwic: yes, it can; session bus services retain all environment variables like $XDG_*, $DISPLAY, etc.10:18
pittidiwic: (unlike the system bus, where services have zero env)10:18
diwicpitti, sounds good, thanks!10:20
lotuspsychjeitsme_: do you have any card reader on ubuntu?10:21
itsme_No I don't10:21
lotuspsychjehmm10:21
lotuspsychjei found a thread with that j random user, not sure it has something to do with this10:22
lotuspsychjehttp://menari.eu/post/42508697963/feitian-epass2003-with-opensc-on-archlinux10:22
lotuspsychjemaybe its used in same way of John doe10:23
lotuspsychjecant find any related stuff for ubuntu...10:23
itsme_lotuspsychje :  varunendra: "Secure Boot" or UEFI feature was widely implemented in 2012 and its an Intel technology. My motherboard (Gigabyte) and processor is AMD and I haven't updated my Bios since 200810:23
* cjwatson notices the discussion of bug 315833 overnight, and tidies up rmadison output10:30
ubottubug 315833 in madison-lite (Ubuntu) "Ubuntu’s rmadison should have wider column defaults" [Low,Fix committed] https://launchpad.net/bugs/31583310:30
mlankhorstwhy is https://launchpad.net/ubuntu/+source/xorg-server/2:1.14.3-5ubuntu1 still in proposed?10:45
Laneymlankhorst: http://people.canonical.com/~ubuntu-archive/proposed-migration/update_excuses.html#xorg-server10:45
mlankhorstoh fu firefox tests10:45
LaneyQuite10:48
LaneyIt's not exactly ideal having those fail all the time10:48
mlankhorstyou could say that :-)10:49
mlankhorstcan you override it?10:49
Laneyokay10:51
mlankhorstthanks10:51
Laneydone10:52
mlankhorstit has a fix for loading mesa 10 :)10:54
mlankhorst../../../../src/gallium/winsys/radeon/drm/.libs/libradeonwinsys.a(radeon_drm_bo.o): In function `radeon_bo_map':10:56
mlankhorst/build/buildd/mesa-10.0.0~rc2/build/dri/src/gallium/winsys/radeon/drm/../../../../../../../src/gallium/winsys/radeon/drm/radeon_drm_bo.c:520:(.text+0xf34): relocation truncated to fit: R_ARM_THM_JUMP11 against symbol `radeon_bo_do_map' defined in .text section in ../../../../src/gallium/winsys/radeon/drm/.libs/libradeonwinsys.a(radeon_drm_bo.o)10:57
mlankhorsthm what is that error in english?10:58
=== freeflying_away is now known as freeflying
pete-woodsdidrocks: hi! Is there somewhere I can read about how to make package releases now? Is every landing still supposed to go through the spreadsheet, or are some projects still automatically daily releasing?11:56
didrockspete-woods: it's still the spreadsheet for now11:57
didrockspete-woods: so nothing really knew11:57
didrocksnew*11:57
pete-woodsdidrocks: okay, and that applies to like every project? whether it's phone related or not?12:07
didrockspete-woods: only everything that is seeded in the phone image12:08
voodHello, I have a probleb with developer.ubuntu.com/publish - I added my application and uploaded source archive but app still appear as "Draft". How can I fix it or send to the review&?12:08
=== _salem is now known as salem_
pete-woodsdidrocks: so, specifically I have two libraries for testing dbus stuff (libqtdbustest/mock), but they aren't auto-releasing, and they aren't on the phone image, either12:09
didrockspete-woods: are they used by phone components?12:10
didrockspete-woods: like build-deps or in the tests?12:10
pete-woodsdidrocks: yeah, they're build deps, does that count me as part of the image then?12:10
didrockspete-woods: so yeah, they are in this rule :)12:10
pete-woodsdidrocks: okay, I will get them added to the spreadsheet then12:11
pete-woods:)12:11
didrocksthanks!12:11
=== MacSlow_ is now known as MacSlow|lunch
=== MacSlow is now known as MacSlow|lunch
pittifor the C++ types here: If I have "std::string dest_string;", and don't do anything else with that, woudl that potentially crash at the end of the function when it gets auto-freed?12:43
pitti(because it's not initialized or something?)12:43
cjwatsonnot that I'm willing to admit to being a C++ type, but I would expect that to use the default std::string constructor12:44
pittiyes, me too12:44
cjwatsonwhich constructs an empty string12:44
pittiI was looking into bug 1254996, and fixed it with https://code.launchpad.net/~pitti/autopilot-gtk/fix-1254996/+merge/19670312:44
ubottubug 1254996 in autopilot-gtk (Ubuntu) "ubiquity crashed with SIGSEGV in GtkNode::MatchStringProperty()" [Critical,Confirmed] https://launchpad.net/bugs/125499612:44
pittibut it's still rather inexplicable to me12:45
pittiaside from the ugliness of tossing some bits of C++ into an otherwise C function, I can't see anything wrong with the original code12:45
cjwatsonit might be something to do with the str != NULL case causing dest_string to refer to something which is now out of scope12:45
pittialso, it's excruciatingly painful to reproduce12:46
pitticjwatson: string's = operator makes a copy of teh assigned C string, though12:46
cjwatsonyeah12:46
pittithat was actually my first thought, but that seems fine to me12:46
pittiI ran the tests through valgrind, it didn't have anything to complain about; and we never saw it with anything else that uses ap-gtk, just with the recent ubiquity tests12:47
pittiand on top of that, it doesn't seem to happen on today's image even, just on yesterday's12:47
* cjwatson cops out with "how strange"12:48
pitticjwatson: yeah, I settle with that, too12:48
pitticjwatson: thanks12:48
cjwatsonafraid it's been over ten years since I wrote C++ professionally, and I'm quite happy with this state of affairs :-)12:48
pittiheh, likewise12:48
pitti(well, not even professionally, I just touched it a bit during uni)12:49
cjwatsonI did quite a bit in $job-2, but not with the STL12:50
pittipresumably it only crashes on Tuesdays anyway12:51
kfunkyofel: any idea about https://bugs.launchpad.net/ubuntu/+source/icecc/+bug/1182491 -- i see that you have an updated version of icecc in your PPA.12:55
ubottuUbuntu bug 1182491 in icecc (Ubuntu) "icecc always running with max_jobs=0" [Undecided,Confirmed]12:55
kfunkcan't we just ship a more recent version of icecc?12:55
kfunkwe're using icecream at the office, some people are using ubuntu 13.04. and right now we were wondering why the overall performance of our icecream network degraded :)12:55
mlankhorstyeah13:02
mlankhorsttoolchain bug \o/13:02
mlankhorstmesa FTBFS on arm if I don't add -fno-optimize-sibling-calls to cflags13:03
=== MacSlow|lunch is now known as MacSlow
=== ara is now known as Guest1754
rbasakpitti: it looks to me as if you've got something else stomping on the stack and writing over dest_string's space, which causes its destruction to blow up.13:23
rbasakpitti: I suspect convert_value.13:23
rbasakpitti: I'm not sure valgrind would catch this necessarily.13:23
rbasakpitti: if you can reproduce this in gdb, you could set up a watch over dest_string's stack space. Then you might get a break where dest_string isn't supposed to be being accessed.13:24
=== ara_ is now known as ara
=== ara is now known as Guest55288
=== Guest55288 is now known as ara_
infinityDktrKranz: Oh, that would explain it.  machines.cgi doesn't seem to reflect that.15:05
=== ara_ is now known as ara
barryxnox, doko this autopkgtest failure in trustyp py33 is very strange.  it shouldn't be happening afaict :/15:48
zulbarry:  ping how do you use tox with dh_python?16:37
barryzul: pybuild has a --test-tox option16:38
barry          --test-tox16:38
barry                 use  tox command in test step, remember to add python-tox to16:38
barry                 Build-Depends. Requires tox.ini file16:38
barry 16:38
zulbarry:  do you have an example package that uses pybuild?16:39
barryzul: i do, but not yet one that also uses tox16:39
zulbarry:  the pybuild example is good enough for me right now :)16:40
barryzul: :)  here's two:16:40
barryhttp://anonscm.debian.org/viewvc/python-modules/packages/enum34/trunk/debian/16:40
barryhttp://anonscm.debian.org/viewvc/python-modules/packages/flufl.i18n/trunk/debian/16:41
zulbarry:  cool thanks16:41
alexblighI'm trying to package a library (distributed as binary+headers - yuck) into a libxyz and libxyz-dev package. I'm trying to do it from scratch with dh_make. This calls fakeroot debian/rules build (which runs Makefile) and fakeroot debian/rules binary. I /thought/ the Makefile was meant to put its output into debian/libxyz (etc.), but fakeroot debian/rules binary then runs dh_prep which clears this. How is this me16:45
alexblighant to work?16:45
xnoxalexbligh: remove all makefiles (since you can't re-compile/install binary) & simply list files where they need to go in debian/libxyz.install & debian/libxyz-dev.install16:50
alexblighxnox, I need the makefiles to expand the stupid tar format thing, rename stuff etc. etc.16:51
xnoxalexbligh: build, typically compiles. And binary stage usually does $ rm -rf debian/$pkg && make install DESTDIR=debian/$pkg16:51
xnox(roughly, it's actually more than that)16:51
alexblighxnox, right, so my 'Compile' step is actually a copy, rename, some perl etc., but the question is 'where should its output go'?16:52
xnoxalexbligh: what are the binaries?16:52
alexblighxnox, parallels-virtualisation-sdk - the binaries are some .so files16:52
xnoxalexbligh: the output should go anywhere, e.g. a dir called "build/" that you create and clean up.16:52
xnoxalexbligh: anywhere, as long as ./debian/rules clean, cleans all of the temp outputs.16:52
alexblighOIC. So why does dh_make tell me to put it in debian/libxyz? (which is what dh_prep unhelpfully deletes)?16:53
alexblighI'll just change the makefile back then!16:53
alexblighSo I should just just put them in (e.g.) build, and then make the .install file install them from build/ to /usr/include etc.16:53
xnoxi don't think that's what dh_make meant =)16:54
xnoxalexbligh: correct.16:54
alexbligh"Done. Please edit the files in the debian/ subdirectory now. You should also check that the libparallelssdk Makefiles install into $DESTDIR and not in / ."16:55
psusixnox: fyi, I looked at switching dmraid to the new driver.. it's a bit more complicated than we thought.  It seems that the new driver lacks the offset parameters ( to specify where in the disk the array begins )16:55
alexblighxnox, thanks16:55
psusiso it will need to create a linear dm target to create the offset and insert it between the raid target and the disk16:55
psusithat exceeded my pain threshold so I gave up16:55
xnoxalexbligh: "install into DESTDIR" means that "make install" will be called with a variable DESTDIR=$something. Which could be debian/$pkg for a single package, or debian/tmp for a multi-binary package. So "upstream Makefile" may not assume anything about $DESTDIR variable.16:57
xnoxalexbligh: but e.g. install headers into $DESTDIR/usr/include/16:58
xnoxalexbligh: later dh_install can be used to split what goes into what package.16:58
alexblighxnox, the trouble is if you do that, it runs dh_prep afterwards which deletes $DESTDIR/usr/include. But anyway, I will just do it manually!16:58
xnoxalexbligh: make build -> may not create anything in debian/[$pkg|tmp], make install -> may not be called as part of build target, ./debian/rules binary -> calls dh_prep (to clean everything) and then calls `make install DESTDIR=debian/pkg` to do a clean fresh installation of build-targets to $DESTDIR17:16
alexblighah! so the real fix is that I should be doing the install under the install target. Doh.17:16
xnoxalexbligh: try experimenting with bear minimal Makefile. E.g.: build: "echo hello > myfile" install: mv myfile $DESTDIR/myinstalledfile17:17
xnoxalexbligh: that's what dh_make meant by install =) the "install:" target =) and make sure it's _not_ the first target in your Makefile. So your first target must be e.g. "build:" and the second one "install:"17:17
alexblighxnox, doh. Thanks17:18
xnoxalexbligh: setting export DH_VERBOSE=1 also helps, cause that dh would actually print what it is invoking, and it makes troubleshooting easier.17:18
catbus1srinira: ping17:20
sriniracatbus1 ping17:27
catbus1srinira: pong17:27
happyaronwanna have some talk about input methods? and damned I got some "jet lag" now...17:45
=== greyback is now known as greyback|away
alexblighIf I am building a library package and a -dev package, what is it that is meant to create the relevant symlinks? Is there some handy howto guide on this?17:47
xnoxalexbligh: usually upstream build-system does (e.g. libtool etc). in your case, you can create them by-hand.17:57
alexblighxnox, ah :-/17:57
xnoxalexbligh: e.g. just drop debian/$pkg.links17:57
xnoxalexbligh: $ man dh_link17:57
alexblighxnox, yeah there's a pile of them. I was going to read the SONAME out the elf table and work out what they should be dynamically so when $upstream upgrades their stuff it works it out17:58
alexblighxnox, do I need to do the ldconfig myself on .postinst etc.?17:58
xnoxcheck your packages with lintian, it will tell you if there are too many or missing ldconfig calls.18:01
alexblighxnox, thanks18:02
=== Logan_ is now known as Guest1674
=== Logan__ is now known as Logan_
seb128ev, is e.u.c using trust-proposed for dbgsym? I'm wondering why the most reported trusty issue doesn't have a valid retracting (it's happening with a proposed version)18:11
evseb128: nope, trusty-proposed is commented out: http://bazaar.launchpad.net/~daisy-pluckers/daisy/trunk/files/head:/retracer/config/Ubuntu%2014.04/18:56
seb128ev, can we get it added? (should I open a bug/where?)18:57
evseb128: I *think* that breaks things. pitti do you remember more about this? I vaguely recall us saying that we get broken retraces on crashes from package versions that predate the version in proposed because apport complains that there's a newer version18:58
evthis is why we've been only flipping on proposed once the release is out the door18:58
evfor the daisy retracers, that is18:58
seb128hum, k18:59
seb128I though apport was smart enough to find the right version when it's available in the index18:59
seb128or one of the index/sources18:59
seb128let's wait for pitti to reply, I guess that's going to be for tomorrow19:00
seb128ev, thanks19:00
evsure thing - keen on getting this resolved in a more long term way :)19:00
seb128;-)19:03
=== thomi_ is now known as thomi
=== hholtmann_ is now known as hholtmann
=== salem_ is now known as _salem
=== tkamppeter__ is now known as tkamppeter
=== freeflying is now known as freeflying_away
=== freeflying_away is now known as freeflying

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