| *** mohan43u_ has quit IRC | 03:17 | |
| *** mohan43u has joined #buildstream | 03:17 | |
| *** mohan43u has quit IRC | 07:16 | |
| *** benschubert has joined #buildstream | 07:26 | |
| gitlab-br-bot | BenjaminSchubert approved MR !1848 (juerg/buildbox-run-error->master: _sandboxbuildboxrun.py: Check for buildbox-run initialization errors) on buildstream https://gitlab.com/BuildStream/buildstream/-/merge_requests/1848 | 07:27 |
|---|---|---|
| *** tristan has quit IRC | 07:28 | |
| *** mohan43u has joined #buildstream | 07:42 | |
| gitlab-br-bot | juergbi merged MR !1848 (juerg/buildbox-run-error->master: _sandboxbuildboxrun.py: Check for buildbox-run initialization errors) on buildstream https://gitlab.com/BuildStream/buildstream/-/merge_requests/1848 | 07:44 |
| *** tpollard has joined #buildstream | 08:12 | |
| *** rdale has joined #buildstream | 08:31 | |
| *** tristan has joined #buildstream | 08:34 | |
| cphang | tpollard, I've come across this before, this is the way that urllib is setup. Some servers block the default user agent https://gitlab.com/libreml/libreml/-/blob/master/elements/matplotlib.bst | 08:40 |
| cphang | A workaround by the matplotlib developers is to pass an empty user agent, but I'm not sure that's something that should be done by default: https://github.com/matplotlib/matplotlib/commit/21150c002bcc3817917b010c9636b807634b2cbd#diff-2eeaed663bd0d25b7e608891384b7298R130 | 08:41 |
| tpollard | that is..... interesting | 08:41 |
| tpollard | this seemingly tracks elsewhere, I wonder if it's hitting a mirror | 08:45 |
| juergbi | could it make sense to set user-agent to BuildStream? | 08:45 |
| juergbi | or do some servers whitelist certain UAs? | 08:46 |
| WSalmon | might be worth asking the gnome people why it works in there ci but not locally | 08:46 |
| WSalmon | are they intersepting something.. | 08:46 |
| WSalmon | tpollard, have you checked that you can build the elements before you started changing things? | 08:47 |
| tpollard | I can track it in gbm | 08:48 |
| *** santi has joined #buildstream | 08:48 | |
| tpollard | if I track it in gbm, and then transpose that new url over to the fork, it works | 08:51 |
| * tpollard boggles | 08:51 | |
| *** lachlan has joined #buildstream | 09:07 | |
| *** phoenix has joined #buildstream | 09:35 | |
| *** narispo has quit IRC | 09:40 | |
| *** narispo has joined #buildstream | 09:41 | |
| gitlab-br-bot | willsalmon opened issue #1280 (Junction in junction requires top level junction) on buildstream https://gitlab.com/BuildStream/buildstream/-/issues/1280 | 09:47 |
| *** phoenix_ has joined #buildstream | 09:50 | |
| *** phoenix has quit IRC | 09:50 | |
| *** chipb has quit IRC | 09:50 | |
| *** jward has quit IRC | 09:50 | |
| *** douglaswinship has quit IRC | 09:50 | |
| *** lchlan has quit IRC | 09:50 | |
| *** chipb has joined #buildstream | 09:51 | |
| *** jward has joined #buildstream | 09:51 | |
| *** douglaswinship has joined #buildstream | 09:51 | |
| *** lchlan has joined #buildstream | 09:51 | |
| *** phoenix_ is now known as phoenix | 09:51 | |
| *** lachlan has quit IRC | 09:58 | |
| *** narispo has quit IRC | 10:05 | |
| *** narispo has joined #buildstream | 10:05 | |
| valentind | On Debian sid now, I cannot call tox with the current requirements. I had to run "make" in requirements/ | 10:07 |
| valentind | I think one of the issue is the version "packaging". | 10:08 |
| valentind | Which is too old from my installed pip. And tox seems to use the installed pip. | 10:09 |
| valentind | Are there any plan to use a systemd user socket to launch buildbox-casd? Would be better to make sure it is linked to the user session and that systemd kills if buildstream crashed. | 10:18 |
| valentind | It is also much better to have the socket in /run than /tmp. | 10:19 |
| juergbi | valentind: unfortunately, systemd socket activation is blocked by https://github.com/grpc/grpc/issues/19133 | 10:21 |
| juergbi | well, it might be possible to work around it by writing our own socket handling code bypassing the one from gRPC, but not sure how much effort that would be | 10:22 |
| valentind | juergbi, I would expect that systemd has already work arounds for that. | 10:22 |
| juergbi | I don't think this can really be worked around on the systemd side (in a reasonable way, without ptrace) | 10:24 |
| valentind | I would expect some kind of piping. | 10:24 |
| valentind | But OK, maybe it is more complex that I imagine. | 10:25 |
| juergbi | it might not be that difficult to work around it inside buildbox-casd | 10:25 |
| juergbi | a python workaround is linked in the bug I mentioned | 10:26 |
| juergbi | something similar may be possible in C++ | 10:26 |
| juergbi | I was originally hoping it would be properly fixed in gRPC by now | 10:26 |
| *** lachlan has joined #buildstream | 10:33 | |
| valentind | juergbi, I will create an issue on buildbox-casd. | 10:33 |
| *** lachlan has quit IRC | 10:58 | |
| *** lachlan has joined #buildstream | 11:02 | |
| *** lachlan has quit IRC | 11:29 | |
| abderrahim[m] | Hey guys, I just noticed my messages didn't get through as I didn't identify with nickserv | 11:31 |
| abderrahim[m] | was wondering why I was invisible | 11:31 |
| abderrahim[m] | am I visible now? | 11:31 |
| tpollard | yep | 11:31 |
| abderrahim[m] | so tpollard about your track problem: are you sure you're not defining a mirror in your fork? | 11:32 |
| abderrahim[m] | and tracking shouldn't change the url, btw | 11:32 |
| abderrahim[m] | did you mean the ref? | 11:33 |
| tpollard | the project.conf was un-altered in the fork | 11:33 |
| tpollard | it was subbing in the alias from the project.conf into the url field after tracking in the working tree though | 11:34 |
| tpollard | anyhow, back up & running now, not really sure what went on | 11:34 |
| gitlab-br-bot | tristanvb merged MR !1846 (tristan/bst-1/testsuite-py38->bst-1: Pass CI with python3.8) on buildstream https://gitlab.com/BuildStream/buildstream/-/merge_requests/1846 | 11:54 |
| abderrahim[m] | tristan: Hi :) | 11:55 |
| *** ChanServ sets mode: +o tristan | 11:55 | |
| * tristan will also set this one line https://gitlab.com/BuildStream/buildstream/-/merge_requests/1847 to merge when completed... | 11:55 | |
| tristan | *when pipeline completes | 11:56 |
| tristan | Hi abderrahim[m] :) | 11:56 |
| abderrahim[m] | for 1.4.2, I'd like to have !1842 merged too | 11:56 |
| gitlab-br-bot | MR !1842: mount: don't pass nonempty option https://gitlab.com/BuildStream/buildstream/-/merge_requests/1842 | 11:56 |
| tristan | aha | 11:58 |
| tristan | abderrahim[m], do you know off hand what that option does/means ? | 11:58 |
| abderrahim[m] | allow to mount on top of non empty directory | 11:59 |
| abderrahim[m] | we only mount on empty directories so it doesn't matter | 11:59 |
| tristan | I think the tests passing are a good indication of that :) | 12:00 |
| tristan | Ok lets do that, I suppose this is irrelevant to master ? | 12:00 |
| abderrahim[m] | already merged in master, this is a backport | 12:00 |
| tristan | Ah, more context :D | 12:01 |
| gitlab-br-bot | tristanvb approved MR !1842 (abderrahim/fuse3-1->bst-1: mount: don't pass nonempty option) on buildstream https://gitlab.com/BuildStream/buildstream/-/merge_requests/1842 | 12:01 |
| tristan | Alright it's on the way | 12:01 |
| *** phoenix has quit IRC | 12:36 | |
| *** hasebastian has joined #buildstream | 12:53 | |
| *** lachlan has joined #buildstream | 12:55 | |
| gitlab-br-bot | tristanvb merged MR !1842 (abderrahim/fuse3-1->bst-1: mount: don't pass nonempty option) on buildstream https://gitlab.com/BuildStream/buildstream/-/merge_requests/1842 | 12:55 |
| tristan | remote: GitLab: You are not allowed to push code to protected branches on this project. | 13:18 |
| tristan | That's new ^^^^ | 13:18 |
| gitlab-br-bot | tristanvb opened MR !1849 (tristan/bst-1/news-update-1.4.2->bst-1: NEWS: Updated for 1.4.2 release) on buildstream https://gitlab.com/BuildStream/buildstream/-/merge_requests/1849 | 13:19 |
| * tristan awaits another round of CI so he can create a new tag | 13:20 | |
| * tristan hopes he is allowed to push a tag ! | 13:20 | |
| *** phoenix has joined #buildstream | 13:34 | |
| tristan | valentind, seems we still have CI issues, e.g.: https://gitlab.com/BuildStream/buildstream/-/jobs/494132102 | 13:34 |
| tristan | happening now :-S | 13:34 |
| tristan | but please don't start killing jobs, I think we can get https://gitlab.com/BuildStream/buildstream/pipelines/131824526 through so I can push a release tag | 13:35 |
| coldtom | https://docs.gitlab.com/ee/ci/yaml/#skipping-jobs might be of use for doc/news changes like that, tristan | 13:36 |
| tristan | ohhhh that's good to know | 13:39 |
| tristan | coldtom, do you know why I don't have an enabled "merge" button here: https://gitlab.com/BuildStream/buildstream/-/merge_requests/1849 ? | 13:42 |
| coldtom | looks like you need approval if ci is skipped | 13:43 |
| coldtom | which i guess makes sense | 13:43 |
| *** phoenix has quit IRC | 13:43 | |
| tristan | Still says no approval required | 13:43 |
| valentind | tristan, looking at the logs | 13:43 |
| tristan | Can someone approve !1849 ? | 13:44 |
| tristan | See if that helps ? | 13:44 |
| tristan | I think that it might be a gitlab bug, as I may have had that on "Merge when pipeline succeeds", and then force pushed a version with [skip ci] | 13:44 |
| valentind | Same thing as yesterday I think. gitlab-runner seems to be stuck in some loop trying to remove machines. | 13:45 |
| valentind | 422 creating this/these droplet(s) will exceed your droplet limit | 13:48 |
| valentind | tristan, jjardon, have we set a droplet limit? | 13:48 |
| gitlab-br-bot | coldtom approved MR !1849 (tristan/bst-1/news-update-1.4.2->bst-1: NEWS: Updated for 1.4.2 release) on buildstream https://gitlab.com/BuildStream/buildstream/-/merge_requests/1849 | 13:50 |
| jjardon | valentind: IT always has been 50; it was increased to 100 like 2 months ago | 13:50 |
| tristan | coldtom, ummm, did the button become enabled ? I just pushed one more revision of that branch, can you reapprove please ? | 13:51 |
| gitlab-br-bot | coldtom approved MR !1849 (tristan/bst-1/news-update-1.4.2->bst-1: NEWS: Updated for 1.4.2 release) on buildstream https://gitlab.com/BuildStream/buildstream/-/merge_requests/1849 | 13:51 |
| jjardon | valentind: something is wrong if we need more than 100, I think | 13:52 |
| valentind | I can increase the frequency of running cleanup-docker-machine.timer | 13:52 |
| tristan | coldtom, thanks... unfortunately still cannot merge it :-S | 13:52 |
| tpollard | on the DO dashboard there's many droplets that are over 6 hours old | 13:52 |
| tpollard | I presume they're all stuck again | 13:52 |
| valentind | Wait that was not that script | 13:54 |
| tristan | Ok well, I'll remove the [skip ci], repush, and wait for the pipeline to chug along I guess | 13:54 |
| tristan | Should I wait for the runner situation to get fixed ? | 13:54 |
| valentind | There was one script that was removing old machines. | 13:54 |
| * tristan queues a CI of the NEWS entries | 13:56 | |
| tpollard | valentind: https://gitlab.com/BuildStream/infrastructure/infrastructure | 13:56 |
| tpollard | that should be cronning hourly | 13:56 |
| *** hasebastian has quit IRC | 13:58 | |
| valentind | tpollard, Does this delete the droplets? | 13:59 |
| jjardon | tristan: Id skip CI for README/ NEWS changes | 14:00 |
| valentind | Ah right. droplet.destroy() | 14:00 |
| jjardon | tristan: -o ci.skip when you push | 14:00 |
| valentind | Well it is not what is on the machine. | 14:00 |
| valentind | So it is broken. | 14:00 |
| valentind | Someone modified the script to remove docker machine files. | 14:01 |
| valentind | But this is done by another script. | 14:01 |
| coldtom | benbrown was looking around in there recently i believe | 14:01 |
| valentind | /root/destroy_droplets/destroy_old_droplets.py | 14:02 |
| valentind | This was modified to do something that /usr/local/bin/cleanup-docker-machine already does. | 14:03 |
| valentind | But it stopped removing droplet. | 14:03 |
| valentind | I put back the old file then. | 14:03 |
| valentind | So it deleted few. But many are not that old. I will automatically delete the ones that are older than 12 hours. | 14:07 |
| tpollard | Ok | 14:09 |
| tpollard | I think 12 is lose enough threshold | 14:09 |
| valentind | OK, that should be a bit bettr. | 14:09 |
| tpollard | *loose | 14:10 |
| valentind | "Invalid user appserver from 45.55.238.20 port 55422" | 14:12 |
| tristan | jjardon, that doesn't work, though | 14:12 |
| tristan | jjardon, we just tried it; it results in a merge request being confused and not being able to merge | 14:13 |
| valentind | tristan, there are still some issues | 14:13 |
| valentind | But now it is creating new builders. Let's see if those work. | 14:13 |
| valentind | "429 Too many requests" | 14:14 |
| jjardon | tristan: mmm, maybe you have some special config in buildstream ci, but it should work: https://docs.gitlab.com/ee/user/project/push_options.html#push-options-for-gitlab-cicd | 14:14 |
| tristan | still some issues with CI ? or issues with !1849 ? (sorry ambiguity... the merge request was telling me I needed to resolve "all the issues" above before I can merge) | 14:15 |
| gitlab-br-bot | MR !1849: NEWS: Updated for 1.4.2 release https://gitlab.com/BuildStream/buildstream/-/merge_requests/1849 | 14:15 |
| tristan | jjardon, we are not allowed to push to protected branches from the command line, that might be a special config | 14:15 |
| tristan | not sure | 14:15 |
| tristan | Anyway, it's not like these commits are super frequent | 14:16 |
| tristan | It's not particularly urgent either but I'd like to finish the release tonight heh | 14:16 |
| valentind | We are under attack, see: journalctl | grep "invalid user" | 14:20 |
| valentind | They are scanning lots of users. | 14:22 |
| tpollard | ouch, yep that's getting hit hard | 14:24 |
| tristan | heh | 14:25 |
| tpollard | they're scrambling the IP's too | 14:28 |
| valentind | Most are from digital ocean. | 14:29 |
| valentind | But I suppose it is compromised. | 14:29 |
| valentind | tristan, Can you tell me if the ci is working better for your now? | 14:33 |
| valentind | How do we contact the security team of digital ocean? | 14:34 |
| valentind | Probably abuse@ | 14:34 |
| tristan | valentind, it's been working alright except for that pipeline which had some retries going on earlier, at least for the couple of pipelines I've triggered | 14:36 |
| tristan | valentind, All I can really observe here is (A) Long time to assign a runner / not all jobs being processed in parallel... (B) Errors in trying to get a runner | 14:37 |
| tristan | Those two things don't seem to be happening right now | 14:37 |
| tristan | https://gitlab.com/BuildStream/buildstream/pipelines/131839312 seems almost done now | 14:38 |
| valentind | tristan, Can you tell me if you see it again happening? I fixed a few things. I think there will still be some issues with retries due to some throttling from digital ocean API. | 14:40 |
| valentind | But I hope that the complele failures that were due to a modified script and a limit in the number of droplet, should be fixed. | 14:41 |
| tristan | Ok nice :) | 14:43 |
| tristan | Yup I'll keep you posted | 14:43 |
| gitlab-br-bot | tristanvb merged MR !1849 (tristan/bst-1/news-update-1.4.2->bst-1: NEWS: Updated for 1.4.2 release) on buildstream https://gitlab.com/BuildStream/buildstream/-/merge_requests/1849 | 14:45 |
| *** tristan changes topic to "BuildStream 1.4.2 is out ! | https://gitlab.com/BuildStream/buildstream | Docs: https://docs.buildstream.build/ | IRC logs: https://irclogs.baserock.org/buildstream | Mailing List: https://mail.gnome.org/mailman/listinfo/buildstream-list | Roadmap: https://wiki.gnome.org/Projects/BuildStream/Roadmaps" | 14:49 | |
| *** tristan has quit IRC | 14:54 | |
| valentind | \o/ | 15:06 |
| tpollard | :) | 15:07 |
| valentind | I realize there is now a fix for fuse3 in master. Was this backported to 1.4? | 15:09 |
| coldtom | valentind, yep | 15:13 |
| valentind | Yeah! | 15:14 |
| valentind | I can now update! | 15:14 |
| *** lachlan has quit IRC | 15:20 | |
| juergbi | debian 9 docker image build fails because it can't install cmake | 15:38 |
| juergbi | from stretch-backports | 15:39 |
| juergbi | The following packages have unmet dependencies: | 15:39 |
| juergbi | cmake : Depends: libarchive13 (>= 3.3.3) but 3.2.2-2+deb9u2 is to be installed | 15:39 |
| juergbi | i.e., it seems cmake was updated in stretch-backports but the required dependency was not updated | 15:39 |
| juergbi | I can install more recent libarchive13 from stretch-backports-sloppy but I'm not sure whether that's the proper solution | 15:40 |
| juergbi | anyone more familiar with debian backports here? to me it seems like a distro bug | 15:41 |
| juergbi | it worked fine a few weeks back where stretch-backports still had an older cmake version but it seems the older package has been deleted from the archive :-/ | 15:42 |
| *** lachlan has joined #buildstream | 15:43 | |
| juergbi | maybe it's time to discuss dropping support for python 3.5 / debian 9 | 15:44 |
| benschubert | ^ I think that might be a good thing for bst 2.0 anyways. We could use async for some parts also now | 15:48 |
| *** tristan has joined #buildstream | 15:49 | |
| *** lachlan has quit IRC | 15:55 | |
| tpollard | There's a few other 3.5 niggles we can drop too | 15:56 |
| tpollard | black not supporting 3.5 is one that comes to mind | 15:58 |
| juergbi | regular EOL of Debian 9 is anyway expected in a few months, although LTS support will be available for another two years or so | 16:00 |
| juergbi | should probably propose this on the mailing list | 16:01 |
| tpollard | that reminds me, 20.04 Ubuntu will be LTS too won't it? | 16:01 |
| *** lachlan has joined #buildstream | 16:01 | |
| juergbi | yes | 16:04 |
| juergbi | 18.04 is already on Python 3.6 | 16:04 |
| * tpollard nods | 16:04 | |
| *** mohan43u has quit IRC | 16:59 | |
| *** tpollard has quit IRC | 17:01 | |
| *** mohan43u has joined #buildstream | 17:03 | |
| *** lachlan has quit IRC | 17:05 | |
| *** mohan43u has quit IRC | 17:10 | |
| *** mohan43u has joined #buildstream | 17:13 | |
| *** lachlan has joined #buildstream | 17:23 | |
| *** santi has quit IRC | 17:59 | |
| *** santi has joined #buildstream | 17:59 | |
| *** traveltissues has quit IRC | 18:06 | |
| *** traveltissues has joined #buildstream | 18:07 | |
| *** robjh has quit IRC | 18:07 | |
| *** jward has quit IRC | 18:07 | |
| *** valentind has quit IRC | 18:07 | |
| *** WSalmon has quit IRC | 18:07 | |
| *** bethw has quit IRC | 18:07 | |
| *** bethw has joined #buildstream | 18:07 | |
| *** jward has joined #buildstream | 18:07 | |
| *** robjh has joined #buildstream | 18:08 | |
| *** valentind has joined #buildstream | 18:11 | |
| *** WSalmon has joined #buildstream | 18:13 | |
| *** lachlan has quit IRC | 18:14 | |
| *** santi has quit IRC | 18:22 | |
| *** ikerperez has quit IRC | 18:28 | |
| *** cphang has quit IRC | 18:28 | |
| gitlab-br-bot | dylanmccall opened issue #1281 (Downloadable file source is not setting a user agent; rejected by gitlab) on buildstream https://gitlab.com/BuildStream/buildstream/-/issues/1281 | 18:46 |
| *** cphang has joined #buildstream | 19:19 | |
| *** toscalix has joined #buildstream | 19:24 | |
| *** benschubert has quit IRC | 19:55 | |
| *** phoenix has joined #buildstream | 20:26 | |
| *** rdale has quit IRC | 21:39 | |
| *** toscalix has quit IRC | 22:42 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!