/srv/irclogs.ubuntu.com/2014/02/05/#ubuntu-ci-eng.txt

fgintherveebers, yes, is this for an existing project (i.e. can you just submit an MP)02:15
veebersHi fginther, yeah this is for libautopilot-qt02:16
fgintherveebers, it appears to be already setup. If you submit an MP to test, the armhf ci job should save the deb files02:18
veebersfginther: ah right good idea :-) Thanks02:18
=== Ursinha is now known as Ursinha-afk
=== tvoss is now known as tvoss|test
=== tvoss|test is now known as tvoss
asacxnox: maybe your uploads have made our image go bloody? :)07:45
asachmmm07:46
asacor not07:46
asac/usr/bin/dialer-app: symbol lookup error: /usr/lib/arm-linux-gnueabihf/libmirprotobuf.so.0: undefined symbol:07:46
asac_ZN6google8protobuf18GoogleOnceInitImplEPiPNS0_7ClosureE07:46
asachttp://ci.ubuntu.com/smokeng/trusty/touch/mako/164:20140205:20140115.1/6457/dialer-app-autopilot/735638/07:46
asaclibprotobuf8:armhf from 2.5.0-5ubuntu2 to 2.5.0-7ubuntu107:46
asacslangasek: ^07:47
asacnot good to have a InitImpl function in an inline thingy07:50
asaci am sure that defeats the purpose of InitImpl07:50
asacfeels like one of those SONAME bumps due to header bustage07:51
asachttps://launchpadlibrarian.net/164839352/protobuf_2.5.0-5ubuntu2_2.5.0-7ubuntu1.diff.gz -> see: once.h diff07:52
=== vila_ changed the topic of #ubuntu-ci-eng to: Ubuntu CI Engineering Team | Vanguard: vila | Landing instructions: http://paste.ubuntu.com/6292280/ | Known issues: -
xnoxasac: all my changes are to do with shutdown on the desktop, plymouth is disabled on touch and i tested all my uploads on grouper before uploading.09:00
asacxnox: aye... already found protobuf as you can see :)09:06
sil2100ogra_: MEETING!09:31
sil2100Ouch, caps-lock09:31
tvossdidrocks, I'm here, too09:32
=== tvoss is now known as tvoss|test
=== tvoss|test is now known as tvoss
=== lool- is now known as lool
asacdidrocks: so think to be safe we might want to have someone look at the ones that dont show the symbol crash ... like weather or terminal09:37
asaci guess its still the same09:37
asacbut... you never know :)09:37
cjwatsonI suspect libmirprotobuf being broken will make it pretty painful to investigate anything :)09:43
tvosscjwatson, so downgrading libprotobuf8 helps here09:45
cjwatsonthat's no surprise at all09:46
cjwatsonI don't need help diagnosing this, I'm busy testing the fix :)09:46
cjwatsonbut yes, I suppose you could check that e.g. terminal works after downgrading libprotobuf809:46
cjwatsonmight be worth doing09:47
ogra_terminal input is broken ... (return and backspace dont work)09:49
ogra_so probably pick something else09:49
cjwatsonor whatever, something that isn't showing the protobuf symbol crash but where autopilot just gives up in a huff09:50
ogra_system-settings doesnt start in the latest image for me ... should be good for a test09:51
asacterminal worked yesterday09:55
asacthe AP09:55
asacso should be fine to check if the AP regressions go away09:56
asactvoss: ^^ can you run terminal AP as well? or weather?09:56
asactvoss: try clock09:56
asachttp://ci.ubuntu.com/smokeng/trusty/touch/mako/164:20140205:20140115.1/6457/ubuntu-clock-app-autopilot/736007/09:56
asacthat also failed and worked yesterday and we dont see the symbol crash09:56
asacso if that goes green with downgrade i feel pretty safe09:57
asactvoss: trying :)?10:02
asac(sorry, but your device just seems ready to running that clock AP :))10:02
tvossasac, trying now10:12
tvossasac, just completed the terminal tests successfully10:17
=== tvoss is now known as tvoss|test
=== tvoss|test is now known as tvoss
tvossasac, clock app tests are running, too10:22
asacnice one10:26
asacseems we just need to get those two things in, kick image and then can continue at reasonable pace :)10:27
cjwatsonso, I can't get my emulator to work properly right at the moment.  when the slowest build ever finishes, can I hand somebody .debs to try out?10:27
cjwatsonbuilding on porter-armhf.c.c at the moment10:27
asaccjwatson: give them to tvoss10:27
asacsorry tvoss :)10:27
tvosscjwatson, yup, just shoot them over10:27
davmor2didrocks: mako is screwed no apps opening on whatever image is currently out10:40
davmor2back in 2010:40
didrocksdavmor2: yeah, would be nice if you can join our morning call :)10:40
didrocksdavmor2: all that is dealt, for now, just wait for next image (we hope to kick it ASAP)10:41
asaccjwatson: tvoss: already shot something over to tvoss?10:46
cjwatsonthe hamsters are grinding as fast as they can10:47
cjwatsonwhich is not very fast10:47
cjwatsonI'll shoot over .debs as soon as I have them10:47
cjwatsonyou know, sod it, I'm going to dump this into trusty-proposed in parallel and block it there until it's tested10:52
cjwatsonotherwise it's just another mumble-dozen minutes of build time10:53
cjwatsonmumble and ccsm work fine with a local build of the new source here10:54
* didrocks does the same with location-service meanwhile10:55
didrocksget that to a silo10:56
asacthanks :)10:57
asachamsters :P10:57
asacstaging in proposed sounds appropriate10:57
davmor2didrocks: what time is the morning meeting?11:01
didrocksdavmor2: 9:30 UTC11:04
davmor2didrocks: hangouts?11:05
didrocksyep, it's a hangout11:06
cjwatsonlet's see if https://launchpad.net/ubuntu/+source/protobuf/2.5.0-8ubuntu1 beats my build on the porter box11:07
ogra_the last upload took 22mins on armhf11:08
ogra_wow11:09
ogra_and 19 on arm64 ... pretty slow for twice the bits :)11:09
cjwatsonas I said on #ubuntu-devel, backlogged + slow database master11:09
cjwatsonpackage size won't have been particularly relevant11:10
ogra_no, but i would have expected a 64bit CPU to build faster11:10
cjwatsonoh you mean the protobuf build11:11
ogra_(though looking at ppc vs ppc64 the pcc variant is faster than the ppc64 one)11:11
ogra_yeah11:11
cjwatsonour arm64 hardware is not exactly production-quality11:11
ogra_yeah, i guess11:11
ogra_well, even amd64 builds it slower than i38611:12
cjwatsonalso, arm64 and ppc64el lose out because those builders are hosted in 1SS, and there's some network slowness between that and the main Launchpad DC that nobody's ever been able to track down well enough to fix11:12
cjwatsonthat will utterly dominate any CPU performance in this case11:12
ogra_ah, so it isnt exactly the build time11:12
ogra_yeah, that makes sense11:12
cjwatsonthough it's true that the build itself was 14 seconds slower on ppc64el here11:13
cjwatsonit also shows 300M more disk space used though11:13
ogra_wow11:13
ogra_thats a lot11:14
davmor2didrocks: I'll see what I can do then11:14
cjwatsonthe amd64 build itself was one second faster than its i386 counterpart, but about 340M more disk used11:14
ogra_seems the 64bit arches use more deps11:15
cjwatsonyou can't really do this kind of fine-grained comparison on single packages very sensibly anyway :)11:15
ogra_yeah, i just found it curious to see the numbers11:15
* ogra_ grins about xnox' "say-no-to-x11" branches ... 11:17
ogra_its a quest !11:17
cjwatsontvoss: deb http://people.canonical.com/~cjwatson/tmp/protobuf/ ./11:20
cjwatsontvoss: though https://launchpad.net/ubuntu/+source/protobuf/2.5.0-8ubuntu1/+build/5555680 is done so you should just be able to pull it from trusty-proposed shortly11:20
cjwatson(or from the librarian)11:20
cjwatsonin fact better to test the librarian builds really11:21
tvosscjohnston, ack and thax11:21
tvossthx11:21
sil2100s/cjohnston/cjwatson11:22
cjwatsontvoss: I need to go to the bike shop; if you test this and it looks good before I get back, feel free to close bug 1276531 and that should cause it to slide into trusty11:28
ubot5bug 1276531 in protobuf (Ubuntu) "ABI regression" [Critical,In progress] https://launchpad.net/bugs/127653111:28
tvosscjwatson, ack11:28
cjwatsonthanks11:28
tvosslooks good so far, clock tests are running after installing the deb11:29
cjwatsonok, good, hopefully there'll be no surprises11:29
ogra_as long as apps start again ...11:29
ogra_:)11:29
tvossogra_, could you please install https://launchpad.net/ubuntu/+source/protobuf/2.5.0-8ubuntu1/+build/5555680/+files/libprotobuf8_2.5.0-8ubuntu1_armhf.deb and see if the apps start working again?11:40
tvossogra_, works for me11:40
ogra_sure11:40
ogra_tvoss, seems fine11:44
ogra_cjwatson, ^^^11:44
* ogra_ has system-settings running again 11:44
tvossogra_, ack and thx11:44
tvossogra_, asac closing the bug that cjwatson pinged me11:44
asacnice one11:45
asacso we have ti testd?11:45
asacif so colin can release the proposed block11:46
asacdidrocks: ?11:46
didrocksasac: doing CITrain training, but sil2100 is looking/helping11:47
didrockshe's working on dbus-cpp as well11:47
tvossasac, closing the bug now, fine with you?11:48
tvossasac, that unblocks the package in proposed and let's it migrate as I understand it11:48
asacis the stuff in archive?11:48
asacclosing bug unblocks the package in proposed?11:48
asaci dont think so :)11:48
asacdidrocks: fine with releating proposed block?11:49
sil2100asac: cjwatson said that closing the bug will cause the package to migrate11:49
asacfor protobuf?11:49
asacah cool11:49
didrocksasac: yep :)11:49
sil2100asac: ^11:49
asacif didrocks is happy lets try that11:49
asacinteresting feature11:49
tvosssil2100, so I'm setting it to fix committed or fix released?11:51
tvossdidrocks, ^?11:52
sil2100tvoss: not sure, cjwatson would be the one to know ;) We're betting on Fix Committed11:52
tvosssil2100, done :)11:53
sil2100\o/11:54
=== vila_ changed the topic of #ubuntu-ci-eng to: Ubuntu CI Engineering Team | Vanguard: cihelp | Landing instructions: http://paste.ubuntu.com/6292280/ | Known issues: -
=== cprov changed the topic of #ubuntu-ci-eng to: Ubuntu CI Engineering Team | Vanguard: cprov | Landing instructions: http://paste.ubuntu.com/6292280/ | Known issues: -
=== MacSlow is now known as MacSlow|lunch
cjwatsontvoss,sil2100: fix committed won't do anything, any actual closed state (e.g. fix released) is fine12:14
tvosscjwatson, thx12:15
cjwatsonI've fix-released it now12:15
tvosscjwatson, thank you :)12:15
cjwatsonasac: https://lists.ubuntu.com/archives/ubuntu-devel-announce/2013-October/001068.html FYI12:15
sil2100cjwatson: hah, so me and didrocks guessed wrong ;)12:15
ogra_cjwatson, wow, sweet !12:16
cjwatsonthanks for the tests12:16
ogra_(why did i never notice that mail)12:17
davmor2ogra_: because you get more than 1 mail a day?12:20
asaccjwatson: wow... so thats avail for everyone? cool stuff12:21
ogra_asac, note that date12:21
asacyes12:21
asachence12:21
asac:)12:21
ogra_we have that since like ... forever12:21
* ogra_ didnt know about it either ... i somehow missed that mail 12:21
asacwell... i guess it this mechanic detail12:21
cjwatsonstart of this cycle isn't forever :)12:21
asacwas not in my head12:22
ogra_:)12:22
asaci thought touch had a way to block through bzr12:22
cjwatsonyou do as well, yes12:22
ogra_we do12:22
asacyeah, but i was stuck at that :)12:22
* ogra_ has only used it at release time though 12:22
ogra_i dont think we had to touch it since12:22
cjwatsonwhat that mechanism gives you that the bug mechanism doesn't is that you can *unblock* during freezes12:22
ogra_(dider probably did)12:22
rsalvetimorning12:25
sil2100davmor2: hi! I fixed the SMS OSK problem, it might be released sooner-or-later with some other changes12:38
davmor2sil2100: nice :)12:39
rsalvetiargh, launchpad is just giving me timeout errors today12:41
sil2100Same here12:41
ogra_rsalveti, yeah12:42
davmor2rsalveti: it knows you are going to try and make it work12:42
ogra_some upgrade ... something needs regeneration ... everything gets slow for a while12:42
rsalvetiright12:42
rsalvetididrocks: mlankhorst is finishing the xorg migration and then we should be good to publish mir12:43
sil2100rsalveti: \o/12:44
rsalvetihopefully nothing will explode :-)12:44
cjwatsonogra_: no, it's not about anything needing regeneration, please don't make things up :)12:46
cjwatsonthe master database system was being upgraded to precise (that's just finished), and so it was taken out of rotation and a slower slave was temporarily promoted to master12:47
ogra_i thought it needs to re-fill a db12:47
ogra_sorry12:47
cjwatsonit's older hardware and can't quite do the job as quickly12:47
cjwatsonbut it's a lot better than taking all of LP down for an hour or two12:47
asacindeed12:47
cjwatsonthe slaves are replicated, it had everything, just not necessarily well-cached etc.12:47
cjwatsontrying a few times should be sufficient; and I think wildcherry is being swapped back in RSN12:48
didrocksrsalveti: probably not, but I just want to kick an image with location-service and protobuf fix first12:48
didrocksrsalveti: then, we just publish xorg, mir… whatever12:48
rsalvetididrocks: great12:49
rsalvetididrocks: https://code.launchpad.net/~sergiusens/qtubuntu/empty/+merge/204183 can be merged already btw (from landing-1)12:50
didrocksrsalveti: you can click on "merge and clean"12:51
didrocksrsalveti: that would do it for you12:51
didrocks(in landing-001)12:51
rsalvetiDONE12:51
didrocks\o/12:52
didrocksthanks rsalveti :)12:52
rsalveti:-)12:54
sergiusensrsalveti, yeah, don't confuse the landing box ;-)12:59
rsalvetinice, it keeps spinning to see if the packages are really gone13:05
rsalvetisergiusens: hey, welcome back13:05
didrocksrsalveti: yeah, it can take times to have them really cleaned13:05
didrocksrsalveti: see http://ppa.launchpad.net/ci-train-ppa-service/landing-001/ubuntu/dists/trusty/main/binary-armhf/Packages13:05
didrocksfor instance13:05
rsalvetiyeah13:05
=== Ursinha-afk is now known as Ursinha
didrocksbut then, you should be fine, once everything is really cleaned, the line will change to "landed" and landing-001 will be available13:06
sil2100tvoss: how's it proceeding? Will we have to land dbus-cpp only or process-cpp as well?13:07
=== alan_g is now known as alan_g|lunch
tvosssil2100, dbus-cpp has to land13:12
sil2100tvoss: ACK13:12
cjwatsonfixed protobuf is in trusty now, btw13:21
ogra_yay13:21
=== MacSlow|lunch is now known as MacSlow
sil2100\o/13:27
sil2100So now just the fix from tvoss is needed ;13:28
sil2100tvoss: you have many eyeballs eyeballing you!13:30
tvosssil2100, sure13:31
tvosssil2100, that's not that new :)13:31
tvossdidrocks, I could use some help here: https://code.launchpad.net/~thomas-voss/dbus-cpp/force_gcc_4.7/+merge/20490913:37
tvossdidrocks, when trying to build with dpkg-buildpackage, gcc 4.8.x is chosen13:37
didrockshum13:38
didrocks-Package: libdbus-cpp1.113:38
didrocks18+Package: libdbus-cpp113:38
didrockswhy was it changed in trunk and now it's back to compatbility?13:38
tvossdidrocks, shouldn't have been switched on trunk13:39
didrocks-- The C compiler identification is GNU 4.7.313:41
didrocks-- The CXX compiler identification is GNU 4.7.313:41
didrocks-- Check for working C compiler: /usr/lib/ccache/gcc-4.713:41
didrocks-- Check for working C compiler: /usr/lib/ccache/gcc-4.7 -- works13:41
didrocks/usr/lib/ccache/g++-4.7 here (but I'm using ccache)13:41
tvossdidrocks, not working here, but let's just wait for the silo13:45
didrockstvoss: hum, we aren't building dbus-cpp in a silo, right?13:46
didrocksas we can't ship the whole trunk13:46
didrockswhich is != archive13:46
didrockswe only want your fix, right?13:46
=== alan_g|lunch is now known as alan_g
sil2100didrocks: yes13:49
sil2100didrocks: that's why I'm dputting dbus-cpp to the PPA13:49
didrockssil2100: still not published for me13:54
* didrocks hesitates to bother LP guys as I guess it's a followup on the upgrade (and maybe catching up)13:54
sil2100Still not popping up for me as well ;/13:55
cjwatsonthe cron jobs are disabled right now for the second of the two db switches today13:55
cjwatsoni.e. switching back to the real master13:55
cjwatsonthey should return in a few minutes13:55
didrocksok, thanks cjwatson :)13:55
=== cprov changed the topic of #ubuntu-ci-eng to: Ubuntu CI Engineering Team | Vanguard: cihelp | Landing instructions: http://paste.ubuntu.com/6292280/ | Known issues: -
sil2100tvoss, didrocks: https://launchpad.net/~ci-train-ppa-service/+archive/landing-005/+build/5557271 <- it's using 4.7 here14:30
sil2100\o/14:30
tvosssil2100, thanks14:30
didrocks/usr/bin/g++-4.714:30
didrocksphew14:30
didrocksand the cherry-pick works14:31
didrockscontrary to trunk with half renaming :)14:31
didrockslet's see as soon as it's published14:31
didrocksgot everything tested14:31
* ogra_ likes cherries 14:31
ogra_with cream !14:31
didrocksand picks? :p14:31
ogra_indeed !14:32
jibelfginther, hey, there is something I don't understand with http://s-jenkins.ubuntu-ci:8080/job/address-book-service-trusty-amd64-ci/46/ maybe you can help14:42
jibelfginther, a header has been moved from src/addressbook.h to lib/addressbook.h and modified to add new methods14:42
jibelfginther, but the build seems to be grabbing the old version of the source package14:42
jibeland it fails with xxx as no member start()14:42
jibelfginther, any idea what could happen here?14:42
fgintherjibel, looking14:46
sil2100tvoss: ok, so dbus-cpp failed because of symbols...14:46
sil2100tvoss: https://launchpad.net/~ci-train-ppa-service/+archive/landing-005/+packages14:47
fgintherjibel, I know what this is and have an MP in the works to fix. The job configuration needs to be modified to bump the package version so that bzr bd will treat it as a 'new' upstream14:47
sil2100tvoss: to make sure I cherry-picked your fix correctly, here's the diff https://launchpadlibrarian.net/165051845/dbus-cpp_1.0.0%2B14.04.20140123-0ubuntu1_1.0.0%2B14.04.20140123-0ubuntu2.diff.gz14:47
fgintherjibel, should be ready shortly14:47
jibelfginther, good, how shortly is it?14:48
tvosssil2100, the cherry picking looks sane14:49
sil2100I'm not familiar with inter-toolchain problems, but can that be caused by the different toolchain?14:50
fgintherjibel, I'll have address-book-service ready in a few minutes14:51
jibelfginther, perfect, thank you!14:51
fgintherjibel, I've triggered a rebuild on that MP14:56
rsalvetididrocks: xorg-server is in, so once the build with location-service is done, we can try to land mir again14:57
didrocksrsalveti: right, I'll keep an eye and keep you posted14:57
rsalvetithanks14:57
didrocksrsalveti: btw, I forgot, but there is another override (which was in the message, but it's not that clear)14:58
didrocksrsalveti: there is an "ignore dest version check" option in publish14:58
rsalvetididrocks: tried that yesterday, but didn't make any difference14:58
didrocksIGNORE_VERSIONDESTINATION14:59
didrockshum14:59
didrocksah, that is a bug, the override should work14:59
didrocksdo you remember which run?14:59
didrocks(I'll have a look)14:59
rsalvetilet me check14:59
rsalvetididrocks: http://162.213.34.102/job/landing-002-2-publish/13/14:59
didrocksrsalveti: thanks, that should have worked, I'll give it a look15:04
* didrocks first look at why google apps script is "thinking"15:04
davmor2sil2100: I got bored and now I am a time.sleep lord http://bazaar.launchpad.net/~davmor2/+junk/is_there_an_update/view/head:/isthereanupdate.py15:20
balloonsping sergiusens15:32
=== alan_g is now known as alan_g|tea
sil2100tvoss: once you're back from lunch, give us a status report ;)15:38
=== gatox is now known as gatox_lunch
=== alan_g|tea is now known as alan_g
sil2100davmor2: sadly you'll have to wait some more even, we're still waiting for the dbus/location-service fix15:48
davmor2read the else it check once an hour after the first hour is gone :)15:48
sergiusensballoons, is it wrt to the click buddy mr bundle? :-)16:04
tvossxnox, can gou give me a hand in tricking a dpkg-buildpackage to use gcc 4.7?16:12
tvossxnox, my debian/rules looks like this: http://pastebin.ubuntu.com/6879879/16:13
xnoxtvoss: drop line 7, not needed, as DEB_HOST_MULTIARCH is defined inside the include from line 916:17
balloonssergiusens: sorry wifi here is horrid :-) Yes, let's talk click-buddy please16:17
plarsdidrocks: ogra_: any idea when to expect the next image?16:17
xnoxtvoss: and you need a build-dependency on g++-4.716:17
xnoxtvoss:  and that should work, can I fetch the whole thing to test here?16:17
didrocksplars: no idea at all, tvoss and sil2100 are working on dbus-cpp fix16:18
tvossplars, on it, need to figure out symbol file changes16:19
sergiusensballoons, I've have 3 hours in meetings now; but I can still pay attention to irc; so feel free to shoot16:19
tvossxnox, sure @build dep, that's there, here is what I get: -- The C compiler identification is GNU 4.8.216:21
tvoss-- The CXX compiler identification is GNU 4.8.216:21
tvoss-- Check for working C compiler: /usr/bin/x86_64-linux-gnu-gcc16:21
tvoss-- Check for working C compiler: /usr/bin/x86_64-linux-gnu-gcc -- works16:21
tvoss-- Detecting C compiler ABI info16:21
tvoss-- Detecting C compiler ABI info - done16:21
tvoss-- Check for working CXX compiler: /usr/bin/x86_64-linux-gnu-g++16:21
tvoss-- Check for working CXX compiler: /usr/bin/x86_64-linux-gnu-g++ -- works16:21
davmor2ogra_: new n7 is razor right?16:21
ogra_davmor2, flo16:21
davmor2ogra_: right thanks16:22
ogra_razor is the old grouper with 3G iirc16:22
xnoxtvoss: and what is the cmake package version number?16:22
davmor2thanks16:22
tvossxnox, 2.8.12.1-1ubuntu516:22
xnoxtvoss: that should just work.16:23
xnoxtvoss: can I please fetch the full source package / branch please?16:23
tvossxnox, well, that would be my wish, too :)16:23
xnoxtvoss: does adding -DCMAKE_CXX_COMPILER=$(CXX) and -DCMAKE_C_COMPILER=$(CC) to the end of dh_auto_configure make it work?16:24
tvossxnox, let me check16:24
xnox(ugly & explicit and shouldn't be necessary...)16:24
xnoxif that also doesn't help, it could be that there are custom toolchains or BuildTypes overriden in the CMakeLists in those sources.16:25
tvossxnox, lp:~thomas-voss/dbus-cpp/force_gcc_4.7/16:28
tvossxnox, that actually works16:30
xnoxtvoss: =) and it does the right thing in the clean chroot / sbuild as well. \o/16:34
tvossxnox, nope, the dirty line works16:36
tvoss:)16:37
xnoxtvoss: i've cloned your branch and without modifying it, built it, and it used gcc-4.7.16:37
tvossxnox, might well be that my system is just screwed up16:38
rsalvetididrocks: any news regarding the location service & image?16:38
rsalvetiI know you'll be off soon :-)16:38
xnoxtvoss: well, i built in a clean trusty sbuild, as ~= done by launchpad. So it will work there. And extra -DCMAKE_*_COMPILER will not hurt.16:38
=== gatox_lunch is now known as gatox
tvossxnox, great, as long as I can get dpkg-buidpackage to throw .symbol file issues at me16:40
xnoxtvoss: ah, that =) note it will typically throw mangled symbols at you.16:42
sil2100rsalveti: tvoss is working on that ^16:42
rsalvetialright :-)16:42
sil2100rsalveti: the problem got found, fixed but now there's a lot of symbols mis-matches16:42
ogra_freking symbols ... rip them out !16:43
xnoxtvoss: https://wiki.ubuntu.com/DailyRelease/FAQ#I.27m_exposing_a_new_C.2BAC8-C.2B-.2B-_symbols_in_my_library.2C_it_seems_that_some_packaging_changes_are_needed.2BICY- and read the portion about C++ mangled symbols and how to demangle them.16:43
xnoxstarting from "NB! For C++ ..."16:43
xnoxtvoss: if something is not clear feel free to poke me, i've documented that bit of symbols management.16:43
sil2100xnox: the funny thing here is that we have some arch-dependent symbols as well16:47
sil2100xnox: like, some symbols differ for 32 bit archs and 64 bit archs16:47
sil2100xnox: I hacked around the symbols file before, but it's so terrible that my eyes burn16:48
xnoxsil2100: you simply mark them optional, or restrict them by arch.16:52
xnoxsil2100: if you filter and demangle C++ symbols they bleed your eyes less, and there is less per-arch variance.16:52
xnoxsil2100: e.g. _ZN26AccelerometerSensorReading11qt_metacallEN11QMetaObject4CallEiPPv@Base 0.5.1+13.10.20130412ubuntu.unity.next-0ubuntu116:53
xnoxsil2100: becomes this:16:53
xnox(c++)"AccelerometerSensorReading::qt_metacall(QMetaObject::Call, int, void**)@Base" 0.5.1+13.10.20130412ubuntu.unity.next-0ubuntu116:53
xnoxsil2100: and it no longer varies per-arch (since it's encoded as "int", instead of a "N amount of bits int")16:54
sil2100xnox: yes, that's what I did previously16:54
sil2100xnox: but it took a long time to do, with a lot of vimdiff and copy+paste madness16:55
xnoxsil2100: there isn't anything better than that =/ and i do agree that C++ generally makes ones eyes bleed =) hence the http://tgceec.tumblr.com/post/74534916370/results-of-the-grand-c-error-explosion-competition16:55
sil2100xnox: I know all this as the current symbols file uses a lot of (c++|optional) and (c++|arch=amd64 arm64 ppc64el) and such16:55
xnox(if one starts to poke symbols, or debug bugs in nested template initialisation)16:55
sil2100But yeah...16:55
sil2100xnox: :(16:55
xnoxsil2100: yeah, "optional" is bad, cause it wouldn't catch regression on that arch. and it's sad that one has to expand 64bit arches like that =(16:56
ogra_well, we dont use that package anywhere except on armhf currently16:57
* ogra_ would go with an interim fix and make it "Architecture: armhf"16:57
ogra_to finally get us unblocked16:57
didrockstvoss: seems that you are pushing still some commits to the branch17:02
didrocksare you done or not really?17:02
sil2100tvoss: I have some symbols files ready right now, let's see how it goes17:08
tvosssil2100, thx17:24
sil2100tvoss: actually, work on the symbols files on your branch normally17:24
sil2100tvoss: we'll temporarily drop the symbols file for this one upload17:24
tvosssil2100, yup, that's what I'm trying to understand17:24
tvosssil2100, build passes just fine here17:24
sil2100tvoss: to unblock stuff we'll be doing shlibs -V17:24
tvosssil2100, ack17:41
tvosssil2100, we can clean up the dbus-cpp branch tomorrow then17:41
tvossdidrocks, mind clarifying what you mean with abi not being in good shape?17:42
didrockstvoss: I'm working on fixing things now, can explain afterwards, but weird that symbols keep changing for something which is at a 1.0.017:42
tvossdidrocks, I really don't think that our assumption should be that interfaces don't change anymore17:43
tvossdidrocks, but we have had that conversation before17:43
didrocksin that case, it's not a library, but let's move on for now17:44
thostr_didrocks: sil2100: can anybody reconfigure silo8? (we found an issue while testing and have a new MP added)17:44
ogra_tvoss, shouldnt that become a 2.0.0 then ?17:44
tvossogra_, sure, happy to bump the major version number17:44
ogra_well first have some symbols files ;)17:44
tvossdidrocks, what is our easy to use way of sharing code between projects then?17:44
tvossogra_, I have, believe me, and I have been throught the process of updating them more often then I actually wanted to17:45
tvossdidrocks, and updating the symbols file really isn't easy but a real pain, just saying17:45
didrocksthen, start acting as a library17:45
didrocksanyway again, I'm fixing things17:45
didrocksno time for chatting17:45
sil2100thostr_: in a moment, we're in firefighting mode right now ;)17:49
thostr_sil2100: have you reconfigured? it seems it got autoupdated, at least the silo now shows all MPs, so is this just a spread sheet feature or is jenkins job properly set up?18:03
sil2100thostr_: I did not do anything yet, still busy, but we noticed problems with google spreadsheets today already18:04
didrocksthostr_: I'm doing it now18:04
thostr_didrocks: thanks!!!18:04
didrocksthe spreadsheet is updated automatically18:04
didrocks(it's a formula)18:04
didrocksthis doesn't come from the backend18:04
thostr_didrocks: ok, so just let me know when it's reconfigured18:04
didrocksthostr_: done18:06
thostr_didrocks: thanks18:07
didrocksyw18:07
didrocksbasic testing done18:08
didrocksjasoncwarner: asac: we didn't get any location as we have to wait for 10+ minutes, but the process doesn't go crazy ^18:08
didrocksso, we'll publish this18:08
didrockstvoss: ogra_: entering proposed18:10
tvossdidrocks, ack18:11
asacricmm: can you come back?18:11
ogra_didrocks, yay, finally18:12
=== alan_g is now known as alan_g|EOD
asacricmm: your P&P tab came up :)18:13
asacas scaring people18:13
didrockskgunn: ok, so to keep you posted (even if the infos are already on the touch ML)…18:17
didrockskgunn: we just published the fixes that will enable to get a better image18:17
didrocksthis is migrating to distro18:17
ogra_didrocks, rsalveti is informed ... i'll do the image and he can do the rest then18:18
didrocksogra_: thanks!18:18
ogra_both packages are not blocked in proposed or anything, right ?18:18
kgunndidrocks: ack18:18
ogra_(i.e. just a matter of waiting)18:19
didrockskgunn: so, once the image is kicked (and I hope someone will have time to do very very basic dogfooding to ensure at least the image isn't is a worse shape…), rsalveti will publish Mir18:19
didrocksogra_: shouldn't, but I'm staying, don't worry18:19
ogra_ok18:19
didrocksuntil they are in the release pocket at least18:19
kgunnogra_: rsalveti just hit me up...i don't mind doing some quick testing18:19
ogra_the office should get better sofas for you :)18:19
didrocksogra_: clearly… tell that to my manager :p18:19
ogra_jasoncwarner, ^^18:19
didrocks"special didrocks' order"18:19
ogra_:)18:20
sil2100;)18:20
rsalvetikgunn: alright18:26
rsalvetididrocks: yeah, will quickly test the image, test the mir stuff to make sure nothing explodes and we'll try to land it again18:27
didrocksrsalveti: excellent, thanks a lot :)18:27
rsalvetiogra_: will brb (~1h), hopefully in time to test the new image :-)18:31
didrocksrsalveti: get a bigger dinner18:32
rsalvetiyeah18:32
didrocks:)18:32
ogra_rsalveti, lol18:33
ogra_wishful thinking18:33
didrockslet's get them published in proposed first :p18:34
ogra_details18:34
didrocksahah18:34
ogra_seems location-service is in dep-wait on ppc ppc64 and arm6418:35
ogra_and dbus-cpp on ppc6418:36
ricmmasac: sorry was cooking while on another call, no irc then18:44
ogra_in madrid you should get an IRC capable stove ;)18:45
didrocksogra_: but that's expected18:48
didrocksit's not new18:48
* didrocks recheck18:49
ogra_whee, location-service moved on18:49
didrocksdbus-cpp18:50
didrockshum18:50
ogra_well, its just done on arm6418:50
ogra_give it a bit18:50
didrocksMissing build dependencies: g++-4.718:51
didrocksargh18:51
ogra_sigh18:53
didrocksok18:53
ogra_but thats ppc6418:54
didrocksso, let's remove them, checking rdepends first18:54
* didrocks uses even more drasticness18:56
* ogra_ looks in the other direction18:57
* didrocks flushes19:06
* ogra_ hears the gurgling 19:06
jibelcihelp can you tell me what I miss to have this https://code.launchpad.net/~jibel/address-book-app/abook_delete_contact_pickmode/+merge/203790 merged?20:14
=== fginther changed the topic of #ubuntu-ci-eng to: Ubuntu CI Engineering Team | Vanguard: fginther | Landing instructions: http://paste.ubuntu.com/6292280/ | Known issues: -
fgintherjibel, looking20:14
fgintherjibel, this project is now under the ci-train process, are you familiar with that?20:17
jibelfginther, not at all20:18
fgintherjibel, I was hoping you were and could tell me :-(20:18
fgintherjibel, let's see what I can find20:18
jibelfginther, okay, I'll see with renato if he knows20:18
balloonsfginther: LOLOLOLOLOL20:19
fgintherjibel, I found some email sent to ubuntu-phone list on Jan 17 "[Ubuntu-phone] Introducing the new upstream release process: CI Train"20:23
ogra_=== Image 165 Building ===20:26
ogra_rsalveti, ^^^20:26
* rsalveti back20:46
rsalvetiogra_: great, almost in time20:46
ogra_yeah, dbus-cpp got stuck so it took a bit longer20:46
ogra_(no gcc-4.7 pn ppc64)20:46
ogra_*on20:47
rsalvetiright20:47
ogra_didier did some evil behind the scenes shuffling (not sure what) and it ended up in the archive now20:47
rsalvetigreat20:49
Saviqfginther, was calxeda-pbuilder offline for some time? the armhf builder got pretty queued up: http://s-jenkins.ubuntu-ci:8080/job/generic-mediumtests-builder-trusty-armhf/20:51
balloonsrsalveti, fginther, or someone else on the phablet-tool team, would you be able to review https://code.launchpad.net/~jibel/phablet-tools/run_tests_from_custom_location/+merge/205033 ? I'd like to try and build upon this this week @ the sprint, so it would be helpful to get this merged in phablet-tools and released20:58
rsalvetiballoons: sergio should be able to get that reviewed quickly20:59
rsalvetihe's just not on-line now it seems20:59
rsalvetibut I'll ping him once he's back20:59
balloonsrsalveti: I was hoping to avoid bogging down sergio :-)20:59
balloonsI've asked him a few other favors21:00
rsalvetidon't worry, it's better for him to review this, since he's the maintainer anyway21:00
ogra_yeah, and just means more free drinks for him the next time you meet him21:00
rsalvetiballoons: + wdir=/home/phablet21:00
rsalvetiballoons: mind use wdir=/home/$USER ?21:01
rsalveti*using21:01
balloonsogra_: hehe.. so let's say I'm just trying to keep my drink bill lower eh? :p21:01
jibelrsalveti, I can change that, previously it was /home/phablet/autopilot/ which didn't exist anyway21:01
ogra_heh21:01
rsalvetijibel: yeah, as you're changing it anyway, let's get that fixed :-)21:02
jibelagreed :)21:02
jibeldone21:09
fgintherSaviq, in this case there is nothing wrong, we're just down a few machines due to an upgrade that needs to be done on the machines that are offline21:09
fgintherSaviq, I also noticed the backlog and will try to get that upgrade finished21:09
Saviqfginther, yeah ok, saw that it's progressing now, just wanted to let you know in case something's gone awryt21:15
Saviq-t21:15
fgintherSaviq, feel free to ping if you think something might be wrong, there's a chance that the monitoring is missing something21:17
rsalvetiogra_: did the cdimage part of the build finished already?21:21
rsalvetiseems so, like 10 mins ago21:21
ogra_yeah21:22
ogra_but we should test before pushing the button on mir21:22
ogra_system-image will still take a bit21:22
rsalvetiright, just wonder if the build was already done21:22
rsalveti*wondered21:23
rsalvetiseems to be fine after I updated it by hand, will reflash from scratch using the latest image and enable the ppa to do some more testing with it21:23
rsalvetisergiusens: for you: https://code.launchpad.net/~jibel/phablet-tools/run_tests_from_custom_location/+merge/20503321:24
sergiusensrsalveti, hmmm, I already commented... that breaks ci21:39
ogra_=== Image 165 DONE ===21:39
rsalveti\o/21:40
* ogra_ OTAs on maguro 21:40
ogra_i really like the new upgrader21:40
* tvoss grabs phone21:40
jibelsergiusens, how does it break CI?21:45
sergiusensjibel, it's in my comment; the change of the /home/phablet/autopilot path21:46
ogra_rsalveti, kgunn, i'd say go wild with Mir21:47
ogra_image looks ok on maguro21:47
rsalvetiogra_: looks fine indeed21:47
rsalvetitesting on mako21:47
rsalvetiwill do a quick testing with latest mir again and will try to land it for real21:47
ogra_calls/sms work, apps start etc21:47
jibelsergiusens, but this path doesn't exist currently and there is always an error during tests21:47
ogra_(and top output looks sane too)21:47
jibelsergiusens, that's a bug that has nothing to do with this MP imo21:47
jibelsergiusens, for example https://jenkins.qa.ubuntu.com/job/generic-mediumtests-runner-mako/5089/console21:48
jibelbash: cd: /home/phablet/autopilot: No such file or directory21:48
kgunnack flashing now21:48
sergiusensjibel, all the tests in ci run phablet-click-test-setup to get the click tests21:48
sergiusensjibel, I was acually hoping to have a generic test getter and a test runner; or do you think it better to merge all under the same script?21:49
sergiusensjibel, that last thing is actually a question :-)21:50
jibelsergiusens, from a test writer perspective it is better to have a single tool that makes everything. from a CI perspective separate tool is more flexible21:51
jibelsergiusens, now we can have a wrapper for dev/test writers that glues together separate tools21:51
sergiusensjibel, yeah, for click, I had click-buddy for that; in the end would build, install, provision and run21:52
sergiusenshaven't really thought about the debs yet though21:53
sergiusensfor click as well, I was removing the need to know the python module name and just knowing the click package name21:53
asacogra_: we really got 165?21:54
asac:021:54
asacthought that would never happen :)21:54
* asac crosses finger21:54
asacs21:54
asac:)21:54
ogra_asac, yeah, and it is fine22:01
ogra_(hopefully the dashboard will agree ;) )22:02
asacogra_: what is the best case? 1 test failure?22:02
asacor all green?22:02
ogra_i guess 1 failure on mako 15-17 on maguro22:03
ogra_i havent run the AP tests ... didrocks just wanted a short smoketest to make sure it doesnt go wild ... before we let Mir in22:03
* ogra_ would expect us to be back at the old 1 and 16 failure scheme 22:04
tvossogra_, asac just started ap tests on my phone22:06
ogra_great22:06
ogra_http://people.canonical.com/~ogra/touch-image-stats/20140205.1.changes ...22:09
ogra_the changeset looks fine and small as well22:09
* ogra_ calls it a day 22:13
ogra_rsalveti, there is no need that you kick an image after Mir landed, at 3:00 UTC the cronjob will kick off anyway22:13
rsalvetiogra_: alright22:21
rsalvetiogra_: enjoy22:21
ogra_:)22:21
rsalvetikgunn: alright, landing22:25
rsalvetitrying at least22:25
kgunnrsalveti: yep...all looked good here22:25
kgunni just test u1 acct and loading a click app and all ok22:26
kgunnseems people got the AP stuff ok...and getting same results22:26
rsalvetikgunn: Finished: SUCCESS22:27
rsalvetiit's happening22:27
tvossogra_, asac ap looks good here22:29
asaccool22:30
asactvoss: all ?22:30
tvossasac, nope, not finished yet22:30
asacbut cool anyway22:31
rsalvetikgunn: should, should be all in proposed already22:35
rsalvetikgunn: hopefully landing in release in ~1 hour22:35
rsalvetiwill be afk for a bit, but will check in one hour22:35
=== fginther changed the topic of #ubuntu-ci-eng to: Ubuntu CI Engineering Team | Vanguard: cihelp | Landing instructions: http://paste.ubuntu.com/6292280/ | Known issues: -
kgunnrsalveti: thanks!22:38
didrocksrsalveti: hey, just checking back, now that the image is built (not sure about your dogfooding result), were you able to publish Mir? (seems google spreadsheet isn't really responsive here)22:45
didrocksoh, -changes tell me yes ;)22:47
didrocksI assume everything's fine then, great!22:47
kgunndidrocks: you still on? now can i hit merge/clean ?22:47
rsalvetididrocks: yeah22:47
rsalvetikgunn: not yet22:47
didrockskgunn: you need to wait for Mir to be in the release pocket22:48
rsalvetikgunn: let's wait it to be released22:48
kgunnrsalveti: ah yeah...i see, gotta wait for that to say taht22:48
kgunnthat22:48
kgunnok22:48
didrockskgunn: no harm done before, it will just tell you "not ready yet" :)22:48
rsalvetididrocks: all good for now22:48
didrocksrsalveti: excellent!22:48
didrockskgunn: the status won't update on the spreadsheet, but I guess rsalveti will help you to know when it's ready. I'll try to work on updating the status automatically on the status when I get some time for that22:49
kgunnack...thanks all for the help22:49
kgunntime to do it again :)22:50
didrocksyw! I hope everything's will be fine from now on :)22:50
didrockskgunn: no client breakage/xorg involved this time? :)22:50
kgunndidrocks: sure...22:50
didrocksif so, should be easier22:50
kgunndidrocks: the real answer is to bring xorg guys into the fold and make them use landing22:50
didrockskgunn: more than agreed. I think you saw my answer…22:51
* didrocks sad is to have warned but seems the advice was dismissed22:51
didrockss/sad is/is sad/22:51
didrocksbut anyway, done now ;)22:52
rsalvetiyeah22:52
didrockssee you tomorrow guys!22:53

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