IRC logs for #buildstream for Wednesday, 2018-05-09

« Tuesday, 2018-05-08 Index Thursday, 2018-05-10 »
albfan[m]tristan: Still trying to create that stack.bst, getting progress, no blocks05:43
*** tristan has joined #buildstream07:30
*** toscalix has joined #buildstream07:32
*** bochecha_ has joined #buildstream08:12
*** jennis has quit IRC08:12
*** jonathanmaw has joined #buildstream08:40
gitlab-br-botbuildstream: merge request (valentindavid/359_cross_junction_elements->master: WIP: Cross junction elements) #454 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/45408:41
*** aday has joined #buildstream08:44
gitlab-br-botbuildstream: merge request (jmac/virtual_directories->master: WIP: Abstract directory class and filesystem-backed implementation) #445 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/44509:04
gitlab-br-botbuildstream: merge request (valentindavid/359_cross_junction_elements->master: WIP: Cross junction elements) #454 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/45409:06
*** dominic has joined #buildstream09:06
*** tiago has joined #buildstream09:18
*** finn_ has joined #buildstream10:00
*** finn has quit IRC10:02
finn_Hi, so I tried downloading Alpine directly from the website and I came accross this:11:01
finn_FAILURE tar source at base-alpine.bst [line 7 column 2]: Error staging source: [Errno 1] Operation not permitted11:01
finn_Now if I download the file and try:11:01
finn_tar -xvf alpine.tar.gz11:01
finn_I get:11:01
finn_tar: ./dev/null: Cannot mknod: Operation not permitted11:01
finn_Which I'm guessing causes the original error. Unsure how to proceed, any suggestions?11:01
*** finn has joined #buildstream11:04
*** finn_ has quit IRC11:05
gitlab-br-botbuildstream: merge request (valentindavid/359_cross_junction_elements->master: WIP: Cross junction elements) #454 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/45411:11
*** cs_shadow has joined #buildstream11:23
gitlab-br-botbuildstream: merge request (jmac/virtual_directories->master: WIP: Abstract directory class and filesystem-backed implementation) #445 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/44511:32
tristanfinn, that is very weird, is it the same image we're using in tests ? or does it differ because you got it from the website ?11:33
tristanfinn, maybe ssssam[m] doctored the image to not contain static dev nodes ?11:33
tlatertristan: That's exactly what ssssam[m] did - or well, what his plugin does11:34
tristanthat makes sense yeah11:36
tristanyou cant have static dev nodes, maybe we can produce them one day in artifacts, but certainly not at source staging time11:36
tlaterfinn: You'll have to extract it without the device nodes and tar it back up - that's what I did, at least. Odd that alpine wouldn't consider this11:36
toscalixwhen did BuildStream officially start? when is the project birthday?11:36
tlaterI'll find my makefile for you...11:37
tristantoscalix, hmmm, hard to say, there is not an exact date11:37
toscalixfirst commit?11:38
tristanfirst commit is November 7th 201611:38
tristansorry november 6th11:38
toscalixI will take that one then11:38
tlaterSo BeaverCon is November 7th 201811:38
tristanbut we had meetings and design before writing code of course11:39
toscalixI like the idea of celebrating the first commit11:39
toscalixit is demonstrable11:39
tristanHeh11:39
*** Guest32547 is now known as ernestask[m{11:41
*** ernestask[m{ is now known as ernestask[m]11:41
*** tristan has quit IRC11:53
finnIs there a way to run a script on a downloaded file before it's staged?12:06
tlaterfinn: There's the patch plugin, not sure if it does exactly that12:07
finnAlthough, again for a simple example, is this necessary?12:08
tlaterProbably not :)12:08
finnWhich brings us back to - should we be hosting this .tar?12:08
tlaterRight, that's a point...12:09
tlaterSo, the problem is that alpine's tarball contains device nodes12:10
tlaterWhich we can't extract as a normal user12:10
tlaterSurely we wouldn't be the first to run into this?12:11
tlaterHave you googled around to see if this is a known issue for tarballs?12:11
finnhave done12:11
tlaterOr is everybody just extracting their sysroots as root?12:11
tlaterI think the "correct" solution here is to talk to upstream and figure out what's going on on their end12:12
tlaterI don't think a tarball should contain device nodes12:12
*** tristan has joined #buildstream12:13
tlaterIn the mean time, we will probably want to host this tarball... And then migrate to freedesktop-sdk as soon as that becomes available.12:14
tlatertristan: Any ideas about where we could host a slightly modified alpine tarball?12:15
tlaterfinn: Alternatively, I think the tar source plugin could specify a source regexp without /dev, but I'm not sure that's helpful for a starter project.12:16
tristantlater, dont we already host one ?12:16
tristantlater, if we're using it in integration tests, then we must be hosting it somewhere right ?12:16
finntristan, isn't that a docker file?12:17
tristanI dont know, is it ?12:17
tlaterHrm, it can't be - was originally, perhaps, but certainly isn't anymore12:17
tlaterBecause the docker plugin isn't in the main repo12:17
tlaterLemme check what it is12:17
tristanright12:17
tristanhttps://gnome7.codethink.co.uk/tarballs/12:18
tristanis where we host it currently12:18
tlaterOh, right, this was ssssam[m]'s hacky little script12:18
tlaterHaha, finn, ssssam[m] wrote a script that turns the alpine mini tarball into something usable12:18
tlaterAnd is deploying the result of that into what tristan linked12:18
tlaterfinn: The script is here: https://gitlab.com/BuildStream/buildstream/blob/master/tests/integration/base/generate-base.sh12:19
tlaterSo yeah, I think the solution here is to use the alpine tarball in the repository tristan linked12:20
finnworked12:25
finnta12:25
*** xjuan has joined #buildstream12:49
gitlab-br-botbuildstream: issue #394 ("Abstraction of storage used by sandbox") changed state ("opened") https://gitlab.com/BuildStream/buildstream/issues/39413:09
gitlab-br-botbuildstream: merge request (jmac/virtual_directories->master: WIP: Abstract directory class and filesystem-backed implementation) #445 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/44513:12
valentindStrange error: https://gitlab.com/BuildStream/buildstream/-/jobs/6729900213:49
valentindThis does not seem to be my fault. Should I open an issue for that?13:49
tristanvalentind, this is very strange, you know... I know which commit triggered that13:55
tristanit makes ABSOLUTELY no sense to me why13:56
tristanbut this causes it to happen: https://gitlab.com/BuildStream/buildstream/merge_requests/45213:56
tristanjuergbi, have any idea how MR 452 above could cause that failure ?13:57
tristanI honestly thought it was a very rare race condition, but now it's happened twice since merging that13:57
tristanit *is* a race, but now it seems to trigger more in CI, just because we're doing a version check instead of a feature check for --die-with-parent13:58
juergbihave you verified that the feature check works?13:59
tristanMaybe we are wrong about the version ? maybe it's available since earlier than 0.1.8 and we have < 0.1.8 in the image ?13:59
tristanjuergbi, you mean the version check ?13:59
tristani.e. now we dont do the feature check13:59
juergbieh, yes13:59
tristanI will look into it14:01
tristanbut not tonight, first thing tomorrow14:01
juergbii.e., I would verify that --die-with-parent is actually used in the CI now. if it is, I have no idea why this race condition would be triggered14:02
tristanindeed it seems that we must not be using --die-with-parent14:02
tristanyeah, I will print a bwap --version in CI to check what's in the image too14:02
tristanvalentind, you can open an issue for it if you like yes, and link to MR 452, it's certainly not your fault14:03
* tristan thinks we'll fix it quickly14:04
gitlab-br-botbuildstream: issue #395 ("Job Failed #67299002: FileExistsError") changed state ("opened") https://gitlab.com/BuildStream/buildstream/issues/39514:10
valentindHere it is.14:10
tristanthanks!14:31
toscalixbstbuild then?15:07
toscalixas twitter name?15:07
*** mohan43u has quit IRC15:07
toscalixhaving buildstream is important I believe15:07
toscalixbut I can live without it15:08
dominictoscalix, apparently if @buildstream hasn't had any activity in 6 months you can try and claim it (looks like nothing has happened on it since 2015)15:10
paulsherwooddid we get buildstream.io ?15:10
toscalixtom pollard told me how to claim it15:11
*** mohan43u has joined #buildstream15:12
jmacWhat does "bstbuild" mean? buildstreambuild?15:12
toscalixjmac: that is the point I made on the mailing list15:12
toscalixI would like to claim @buildstream but I need a domain, based on the advices provided by some people on internet related with impersonation claims at twitter15:14
paulsherwoodwhich is why i asked if we got buildstream.io15:14
toscalixbuildstream.org is taken so I assume buildstream.io has been discussed before15:14
paulsherwoodanswer is no :/15:14
paulsherwoodffs15:15
toscalixI will send a mail then with options15:15
* paulsherwood is having an irritable day15:15
toscalixabout domains15:15
*** bethw has joined #buildstream15:31
toscalixI sent to the list some suggestions about the domain, adter some search16:05
toscalixs/adter/after16:05
toscalixbuildstreambuild ja :-) I am getting nuts.16:18
*** toscalix has quit IRC16:31
*** dominic has quit IRC16:44
gitlab-br-botbuildstream: merge request (jmac/virtual_directories->master: WIP: Abstract directory class and filesystem-backed implementation) #445 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/44516:59
*** tiago has quit IRC17:01
*** jonathanmaw has quit IRC17:01
*** tiago has joined #buildstream17:02
*** tristan has quit IRC17:04
*** tiago has quit IRC17:16
gitlab-br-botbuildstream: merge request (valentindavid/359_cross_junction_elements->master: WIP: Cross junction elements) #454 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/45417:32
gitlab-br-botbuildstream: merge request (valentindavid/359_cross_junction_elements->master: Cross junction elements) #454 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/45417:32
*** aday has quit IRC18:46
*** aday has joined #buildstream18:47
*** aday has quit IRC18:51
*** aday has joined #buildstream18:52
*** tristan has joined #buildstream19:05
*** aday has quit IRC19:21
*** aday has joined #buildstream19:22
*** aday has quit IRC19:26
*** aday has joined #buildstream19:27
*** bethw has quit IRC19:29
*** aday has quit IRC19:33
*** bethw has joined #buildstream19:44
*** xjuan has quit IRC20:02
*** tristan has quit IRC20:50
*** bethw has quit IRC21:01
gitlab-br-botbuildstream: issue #396 ("BuildStream is not reporting element-path in the startup summary") changed state ("opened") https://gitlab.com/BuildStream/buildstream/issues/39623:23
« Tuesday, 2018-05-08 Index Thursday, 2018-05-10 »

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!