| *** Prince781 has quit IRC | 00:18 | |
| *** inigomartinez has joined #buildstream | 00:18 | |
| *** Prince781 has joined #buildstream | 00:21 | |
| *** Prince781 has quit IRC | 00:24 | |
| *** Prince781 has joined #buildstream | 00:26 | |
| *** Prince781 has quit IRC | 00:27 | |
| *** Prince781 has joined #buildstream | 00:28 | |
| *** xjuan has quit IRC | 02:29 | |
| *** mohan43u has quit IRC | 03:23 | |
| *** mohan43u has joined #buildstream | 03:23 | |
| *** catonano has joined #buildstream | 06:16 | |
| *** finn has joined #buildstream | 07:08 | |
| *** finn has quit IRC | 07:09 | |
| Kinnison | tlater[m]: Hmm, the pipeline failed at setup because versioneer wasn't available | 07:56 |
|---|---|---|
| *** Prince781 has quit IRC | 08:05 | |
| *** alatiera_ has joined #buildstream | 08:24 | |
| *** Prince781 has joined #buildstream | 08:25 | |
| *** catonano has quit IRC | 08:28 | |
| *** alatiera has quit IRC | 08:28 | |
| *** ssssam[m] has quit IRC | 08:28 | |
| *** WSalmon has quit IRC | 08:28 | |
| *** benschubert has quit IRC | 08:28 | |
| *** paulsherwood has quit IRC | 08:28 | |
| *** laurence has quit IRC | 08:28 | |
| *** mablanch has quit IRC | 08:28 | |
| *** milloni has quit IRC | 08:28 | |
| *** gitlab-br-bot has quit IRC | 08:28 | |
| *** Demos[m] has quit IRC | 08:28 | |
| *** abderrahim[m] has quit IRC | 08:28 | |
| *** juergbi has quit IRC | 08:28 | |
| *** rafaelff[m] has quit IRC | 08:28 | |
| *** connorshea[m] has quit IRC | 08:28 | |
| *** doras[m] has quit IRC | 08:28 | |
| *** krichter[m] has quit IRC | 08:28 | |
| *** pro[m] has quit IRC | 08:28 | |
| *** kailueke[m] has quit IRC | 08:28 | |
| *** asingh_[m] has quit IRC | 08:28 | |
| *** cgmcintyre[m] has quit IRC | 08:28 | |
| *** theawless[m] has quit IRC | 08:28 | |
| *** awacheux[m] has quit IRC | 08:28 | |
| *** jjardon has quit IRC | 08:28 | |
| *** mattiasb has quit IRC | 08:28 | |
| *** jjardon[m] has quit IRC | 08:28 | |
| *** m_22[m] has quit IRC | 08:28 | |
| *** tlater[m] has quit IRC | 08:28 | |
| *** oknf[m] has quit IRC | 08:28 | |
| *** hergertme has quit IRC | 08:28 | |
| *** lchlan has quit IRC | 08:28 | |
| *** Kinnison has quit IRC | 08:28 | |
| *** ironfoot has quit IRC | 08:28 | |
| *** tintou has quit IRC | 08:28 | |
| *** inigomartinez has quit IRC | 08:28 | |
| *** dineshdb[m] has quit IRC | 08:28 | |
| *** abderrahim3 has quit IRC | 08:28 | |
| *** anahuelamo has quit IRC | 08:28 | |
| *** slaf has quit IRC | 08:28 | |
| *** flatmush has quit IRC | 08:28 | |
| *** phildawson has quit IRC | 08:28 | |
| *** albfan[m] has quit IRC | 08:28 | |
| *** segfault3[m] has quit IRC | 08:28 | |
| *** waltervargas[m] has quit IRC | 08:28 | |
| *** tpollard has quit IRC | 08:28 | |
| *** persia has quit IRC | 08:28 | |
| *** cs-shadow has quit IRC | 08:28 | |
| *** valentind has quit IRC | 08:28 | |
| *** aiden has quit IRC | 08:28 | |
| *** thinkl33t has quit IRC | 08:28 | |
| *** Prince781 has quit IRC | 08:28 | |
| *** catonano has joined #buildstream | 08:28 | |
| *** inigomartinez has joined #buildstream | 08:28 | |
| *** dineshdb[m] has joined #buildstream | 08:28 | |
| *** alatiera has joined #buildstream | 08:28 | |
| *** ssssam[m] has joined #buildstream | 08:28 | |
| *** abderrahim3 has joined #buildstream | 08:28 | |
| *** WSalmon has joined #buildstream | 08:28 | |
| *** benschubert has joined #buildstream | 08:28 | |
| *** anahuelamo has joined #buildstream | 08:28 | |
| *** paulsherwood has joined #buildstream | 08:28 | |
| *** laurence has joined #buildstream | 08:28 | |
| *** valentind has joined #buildstream | 08:28 | |
| *** mablanch has joined #buildstream | 08:28 | |
| *** milloni has joined #buildstream | 08:28 | |
| *** aiden has joined #buildstream | 08:28 | |
| *** slaf has joined #buildstream | 08:28 | |
| *** tintou has joined #buildstream | 08:28 | |
| *** ironfoot has joined #buildstream | 08:28 | |
| *** Kinnison has joined #buildstream | 08:28 | |
| *** lchlan has joined #buildstream | 08:28 | |
| *** hergertme has joined #buildstream | 08:28 | |
| *** oknf[m] has joined #buildstream | 08:28 | |
| *** tlater[m] has joined #buildstream | 08:28 | |
| *** m_22[m] has joined #buildstream | 08:28 | |
| *** jjardon[m] has joined #buildstream | 08:28 | |
| *** mattiasb has joined #buildstream | 08:28 | |
| *** jjardon has joined #buildstream | 08:28 | |
| *** awacheux[m] has joined #buildstream | 08:28 | |
| *** theawless[m] has joined #buildstream | 08:28 | |
| *** cgmcintyre[m] has joined #buildstream | 08:28 | |
| *** asingh_[m] has joined #buildstream | 08:28 | |
| *** kailueke[m] has joined #buildstream | 08:28 | |
| *** pro[m] has joined #buildstream | 08:28 | |
| *** krichter[m] has joined #buildstream | 08:28 | |
| *** doras[m] has joined #buildstream | 08:28 | |
| *** connorshea[m] has joined #buildstream | 08:28 | |
| *** rafaelff[m] has joined #buildstream | 08:28 | |
| *** juergbi has joined #buildstream | 08:28 | |
| *** abderrahim[m] has joined #buildstream | 08:28 | |
| *** Demos[m] has joined #buildstream | 08:28 | |
| *** gitlab-br-bot has joined #buildstream | 08:28 | |
| *** flatmush has joined #buildstream | 08:28 | |
| *** thinkl33t has joined #buildstream | 08:28 | |
| *** phildawson has joined #buildstream | 08:28 | |
| *** albfan[m] has joined #buildstream | 08:28 | |
| *** segfault3[m] has joined #buildstream | 08:28 | |
| *** irc.eagle.y.se sets mode: +oo ironfoot jjardon | 08:28 | |
| *** waltervargas[m] has joined #buildstream | 08:28 | |
| *** tpollard has joined #buildstream | 08:28 | |
| *** cs-shadow has joined #buildstream | 08:28 | |
| *** persia has joined #buildstream | 08:28 | |
| *** finn has joined #buildstream | 08:28 | |
| *** jonathanmaw has joined #buildstream | 08:34 | |
| *** Prince781 has joined #buildstream | 08:34 | |
| adds68 | What do i need to pass to the Element class in order to receive and Element object back in Python? | 08:37 |
| adds68 | i just want to look at what is inside an Element object, but it seems passing a yaml file to Element.configure() does not work | 08:38 |
| valentind | adds68, You cannot construct an element. They are constructed by buildstream. | 08:38 |
| skullman | adds68: mostly you load them via the pipeline, they're normally instantiated from MetaElements, but you'll want to avoid having to do that yourself | 08:39 |
| skullman | you need to run something like: app.stream.load_selection(elements, selection=PipelineSelection.NONE) | 08:39 |
| *** bochecha has joined #buildstream | 08:43 | |
| adds68 | skullman oh ok thanks, so i need to import app from BuildStream? | 08:43 |
| skullman | you're writing an external tool? | 08:44 |
| adds68 | skullman a bst plugin, but i want to mess around with an element object to understand the data structure a bit more | 08:45 |
| skullman | hm, I've not done a plugin, can't say I know where the entry point is or what gets passed along | 08:46 |
| adds68 | skullman, i shall keep looking around, but thanks anyway :) | 08:47 |
| skullman | IIRC it's a layering violation for plugins to have access to the app, and maybe even to the stream, so I'm not sure there is an answer | 08:47 |
| adds68 | skullman, yea i'd assume this isn't a normal thing to do, but i just wondering if i could do that in iPython for example | 08:48 |
| adds68 | i was* | 08:48 |
| laurence | following the talk at the gathering last week, is there actually a WIP doc of buildstream's architecture available to look over ? | 08:50 |
| *** raoul has joined #buildstream | 08:51 | |
| *** catonano has quit IRC | 08:53 | |
| phildawson | laurence, it's not an mr yet, but theres a branch along the lines of tristan/architecture-docs | 08:54 |
| phildawson | or at least it wasn't when I looked a couple of days ago | 08:54 |
| laurence | phildawson, cheers | 08:55 |
| valentind | adds68, the element is different depending on the plugin used. They all inherit on Element which inherit from Plugin. So Element and Plugin class are the two places to look at for API. | 08:55 |
| *** catonano has joined #buildstream | 08:56 | |
| valentind | tiagogomes, jjardon, What was the reason for having separate builder for overnight builds? More memory? Bigger CPUs? Pricing of DO? | 08:57 |
| tiagogomes | valentind, main reason was, we had those two virtual machines provided by Codethink, so why not use them for something | 08:58 |
| valentind | tiagogomes, can you forward me the last email from/to operations about those virtual machines. | 09:00 |
| adds68 | valentind, ok i shall take a look at Plugin, seems annoying though that you can't just get an object to play around/debug with | 09:00 |
| tiagogomes | valentind sure | 09:01 |
| tiagogomes | done | 09:01 |
| valentind | adds68, in a plugin you are already in an element, so you can look at self. | 09:02 |
| valentind | tiagogomes, thank you. | 09:02 |
| adds68 | valentind, ahhh ok, i was confused as for the Manifest plugin, Element is passed, so i assumed you needed an Element object | 09:03 |
| valentind | adds68, if you want to look at a specific Element, you can create a plugin that gets Element object through dependencies with self.dependencies(), and then in your .bst of your element just add a dependency to the element you want to look at. | 09:03 |
| tiagogomes | Is there any way of removing some files from an artifact other than using a script element. The files that I need to remove are not in any domain | 09:04 |
| valentind | adds68, I do not see where an element is passed. | 09:04 |
| valentind | adds68, CollectManifestElement is a class not a function. | 09:05 |
| adds68 | valentind, class CollectManifestElement(Element), i would have expected to see "self" | 09:05 |
| valentind | adds68, this is class inheritance. | 09:05 |
| adds68 | valentind, i should have read the Plugin guidelines before though! :P | 09:05 |
| valentind | It means CollectManifestElement is a subclass of Element. | 09:06 |
| adds68 | valentind, ahhh yes! 0_0 | 09:06 |
| valentind | adds68, I recommend reading the Python documentation it is nicely written: https://docs.python.org/3/tutorial/classes.html | 09:07 |
| benschubert | tiagogomes: there is the "compose" element that has an exclude parameter https://docs.buildstream.build/elements/compose.html | 09:13 |
| tiagogomes | yup, but you exclude domains there. I want to filter at file level, not domain level | 09:14 |
| benschubert | oups sorry didn't read you well enough, I think there isn't anything else sadly, since both filter and compose are doing the same :/ | 09:16 |
| adds68 | valentind thanks :) I've spent to long looking at YAML :P | 09:17 |
| benschubert | https://gitlab.com/BuildStream/buildstream/merge_requests/886 ready for review: removing pytest-runner to simplify the lifes of people behind vpns/using other pypi servers :) | 09:20 |
| gitlab-br-bot | willsalmon opened MR !897 (willsalmon/defaultWorkspaces->master: WIP: Updated Workspace CLI) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/897 | 09:28 |
| *** lachlan has joined #buildstream | 09:30 | |
| gitlab-br-bot | BenjaminSchubert approved MR !891 (aevri/include-error->master: More user-friendly reporting on include errors) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/891 | 09:38 |
| gitlab-br-bot | BenjaminSchubert approved MR !889 (valentindavid/ruamel_warnings->master: Fix some ruamel warnings during tests) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/889 | 09:41 |
| * Kinnison wonders if anyone here is able to think about the docker CI systems? https://gitlab.com/BuildStream/buildstream/merge_requests/892 is currently failing to merge because tlater[m]'s updated docker images seem to lack Versioneer, but I have no idea what to do to diagnose/resolve this | 09:48 | |
| tpollard | hmm | 09:51 |
| cs-shadow | Kinnison: versioneer is not part of the Docker image, it is included in the buildstream repo | 09:52 |
| Kinnison | cs-shadow: then might you be able to shed any light on https://gitlab.com/BuildStream/buildstream/-/jobs/112255926 and why it might have failed? | 09:52 |
| * Kinnison is a smidge out of his depth with this stuff | 09:52 | |
| cs-shadow | What I suspect has happened is that we have got newer versions of setuptools and we don't explicitly include versioneer (and some other necessary things) in our MANIFEST.in and we should do that | 09:52 |
| jmac | benschubert: I'm taking a look at !886 now. I think jonathanmaw will need to check his previous questions are addressed. | 09:53 |
| cs-shadow | I think we need add "include versioneer.py" to the end of | 09:54 |
| cs-shadow | https://gitlab.com/BuildStream/buildstream/blob/master/MANIFEST.in#L26 | 09:54 |
| Kinnison | cs-shadow: I'll give that a go, ta | 09:54 |
| benschubert | jmac: thanks! | 09:55 |
| adds68 | valentind, is there anyway to get a "node" without calling buildstream? | 09:56 |
| Kinnison | cs-shadow: that got it through source_dist. Thanks! | 09:56 |
| valentind | adds68, You mean to parse a file? | 09:57 |
| cs-shadow | Kinnison: glad to hear that | 09:57 |
| Kinnison | cs-shadow: I'd have been wandering in the reeds assuming it was a docker issue for hours otherwise | 09:57 |
| adds68 | valentind, from the docs it says "Parameters:node (dict) – The loaded configuration dictionary" | 09:57 |
| cs-shadow | I am secretly rather happy that we hit this issue in the CI as I always had this issue with sdist locally and was procrastinating to create an issue :) | 09:58 |
| adds68 | valentind, which the CollectManifestElement expects to receive? | 09:58 |
| Kinnison | cs-shadow: hah | 10:00 |
| Kinnison | cs-shadow: I tagged you in the MR, as thanks :-D | 10:00 |
| valentind | adds68, yes, configure expects a dict as parameter. Are you trying to test the plugin without running buildstream? | 10:01 |
| adds68 | valentind, yea correct | 10:02 |
| phildawson | Can anyone tell me what the fetch_subproject parameter in Stream._load does? I've followed it about 10 methods down so far and haven't got to the bottom yet. My assumption is that if true, the buildstream project files for any junctioned projects will be fetched, but that is just an assumption. | 10:02 |
| *** catonano has quit IRC | 10:02 | |
| valentind | phildawson, I think you are correct. | 10:03 |
| Kinnison | cs-shadow: I fear even more fallout - I'm now seeing issues with "--integration" being an unknown argument. URGH | 10:04 |
| valentind | phildawson, I think it fetches junctions during loading. | 10:04 |
| * Kinnison thinks tlater[m] needs to weigh in on this | 10:04 | |
| valentind | So that means before the pipeline is scheduled. | 10:04 |
| cs-shadow | Kinnison: I think that's because we are also missing conftest.py from the MANIFEST.in | 10:07 |
| Kinnison | cs-shadow: I think MANIFEST.in is seriously underpopulated | 10:07 |
| Kinnison | the docs build is failing because none of the documentation is copied in either | 10:08 |
| phildawson | thanks valentind | 10:08 |
| valentind | adds68, I do not think it is easy. But if you plan to make some unit testing tools for plugins, that can be interesting, but it is some work. Otherwise I would recommend just to test with .bst files that instantiate your plugin. | 10:08 |
| cs-shadow | Kinnison: I agree. Essentially, at some point setuptools included a lot of stuff from the top-level without them being in the MANIFEST.in, which I think was a bug. And I guess it's been fixed now and we are seeing the fallout from that | 10:08 |
| * Kinnison sets about sorting MANIFEST.in out | 10:08 | |
| adds68 | valentind, yea it seems extremely abstract, so i guess the pain is not worth it | 10:09 |
| valentind | adds68, there are already .bst files using the collect-manifest plugin in the fd.o sdk branch, have you seen them? | 10:10 |
| adds68 | valentind, yea but this method of working is a bit verbose, i usually like to load stuff into iPython and play around with the objects | 10:11 |
| tlater[m] | Kinnison: That's pretty odd, the branch is essentially just master with that one commit on top. | 10:12 |
| Kinnison | tlater[m]: yay for non-strictly-constrained versioning | 10:12 |
| * Kinnison can replicate the sdist issues so is updating MANIFEST.in to try and fix | 10:13 | |
| valentind | adds68, real hackers debug with printf! | 10:13 |
| adds68 | valentind, hahaha why was python chosen, we may as well of used C! :P | 10:13 |
| skullman | argp *is* a better command-line parser than click | 10:16 |
| skullman | but no, I wouldn't suggest BuildStream even in jest | 10:17 |
| skullman | s/BuildStream/& be written in C/ | 10:17 |
| valentind | adds68, Be happy it is not written in Pascal. | 10:17 |
| * skullman locates more caffeine to reduce future typoing | 10:17 | |
| gitlab-br-bot | aevri closed issue #452 (Mistyped urls can cause a stack trace) on buildstream https://gitlab.com/BuildStream/buildstream/issues/452 | 10:17 |
| gitlab-br-bot | aevri merged MR !893 (aevri/unknown-uri-452->master: _downloadablefilesource: handle ValueError-s) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/893 | 10:17 |
| gitlab-br-bot | jmacarthur approved MR !886 (bschubert/remove-pytest-runner->master: Remove dependency on pytest-runner) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/886 | 10:17 |
| *** lachlan has quit IRC | 10:18 | |
| * tlater[m] picks up a new bit of `sed` magic | 10:18 | |
| *** lachlan has joined #buildstream | 10:19 | |
| skullman | tlater[m]: if you want a bit of bash magic to go with it, {filedescriptor}<$filepath opens a file and puts the file descriptor number in $filedescriptor, instead of having to guess at which file descriptor number may be available | 10:21 |
| skullman | handy for if you're using flock | 10:21 |
| * tlater[m] revels in newly acquired knowledge | 10:24 | |
| *** Prince781 has quit IRC | 10:34 | |
| tiagogomes | What "Conflicting junction bootstrap-junction.bst in subprojects, define junction in xpad" exactly means? | 10:41 |
| adds68 | do plugins only get called at build time? | 10:45 |
| tiagogomes | valentind if project B junctions from project A, and project C junctions from project B, that's not allowed? | 10:49 |
| *** phildawson has quit IRC | 10:49 | |
| adds68 | tiagogomes i think at the moment nested junctions are not supported? | 10:49 |
| valentind | tiagogomes, this is allowed. | 10:50 |
| valentind | But you need to make sure that junctions of the same name are the same. | 10:50 |
| valentind | So if A has B as junction and B as C as junction. If A also has a C as junction, it should be the same as the one in B. If you want to access to C from A directly, you should also have it in A. | 10:51 |
| valentind | tiagogomes, and the name is actually the name in project.conf I think. | 10:53 |
| valentind | Or no, I think you need to have matching names in project.conf for matching .bst filename for junctions. | 10:54 |
| valentind | The error message however is not very good. | 10:56 |
| WSalmon | i think i have upset my system and am having this issue when i try to build the doc's https://hastebin.com/yupapaqeja.bash which is funny cos i was building them fine a few weeks ago, any one seen this one before and can tell me the issue before i spend time fixing it? | 11:00 |
| gitlab-br-bot | BenjaminSchubert opened MR !898 (bschubert/fix-command-sandbox->master: Check is command is a str and replace by list before checking existence) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/898 | 11:00 |
| valentind | tiagogomes, I think I know. You have 2 subprojects using a junction of the same name (the name of the .bst file). But this junction is not found in the top project. This seems to not be allowed. | 11:01 |
| jmac | WSalmon: Not seen that one, sorry. element_enums was changed a few weeks ago, but I've not seen that error, I'm afraid | 11:01 |
| valentind | And the matching project.conf name, I was wrong. It does not matter, we use only the .bst file name of the junction as identifier. | 11:02 |
| WSalmon | jmac, thanks | 11:02 |
| tiagogomes | valentind actually I renamed the bst file on C to have the same filename as in B, and I think it worked | 11:02 |
| tiagogomes | Which is odd, because the filename shouldn't matter, but rather the URL | 11:02 |
| WSalmon | i tried reinstalling to see if there was a problem, i suspect something is caching a old thing but i cant workout what i need to do to rebuild just that | 11:02 |
| WSalmon | yep a git clean -fxd fixed it | 11:04 |
| jjardon | how can I force buildstream to rebuild a element? | 11:19 |
| jjardon | also, is there any way to show all the configure / make logs when building? | 11:19 |
| tiagogomes | jjardon delete the ref from rm -rf ~/.cache/buildstream/artifacts/cas/refs | 11:20 |
| *** Prince781 has joined #buildstream | 11:21 | |
| jjardon | tiagogomes: thanks, any idea about the second? | 11:21 |
| tiagogomes | --verbose or --debug? | 11:22 |
| jjardon | tiagogomes: those options doesnt seem to exist | 11:23 |
| tiagogomes | jjardon bst --verbose --debug build … | 11:23 |
| tiagogomes | They need to be placed before the subcommand | 11:23 |
| jjardon | ah rigth | 11:23 |
| jjardon | tiagogomes: still, no configure and make output after passing those | 11:24 |
| jjardon | only debug output of bst itself | 11:25 |
| tpollard | you'll want bst show | 11:26 |
| tpollard | I believe | 11:26 |
| tpollard | it won't give logs but can show you what it's going to do | 11:27 |
| jjardon | tpollard: I need to see the output of running configure and make over a element; not sure that would help | 11:28 |
| coldtom | jjardon: you can find the logs in the cache, but it's a bit of a pain | 11:30 |
| jjardon | indeed | 11:30 |
| jjardon | tiagogomes: tiagogomes coldtom thanks, found it | 11:35 |
| tiagogomes | how did you do it | 11:36 |
| jjardon | searching on the ~/.cache/buildstream/logs/freedesktop-sdk/desktop-fftw/ folder | 11:36 |
| jjardon | see which one is the newest as there are several files there | 11:37 |
| jjardon | so yeah not the best UX :) | 11:37 |
| gitlab-br-bot | jennis opened MR !899 (jennis/patch_remote_cache_docs->master: Improve our documentation for configuring an artifact server) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/899 | 11:55 |
| jennis | Kinnison, regarding yesterday, are you able to review this small docs patch? | 11:58 |
| jennis | https://gitlab.com/BuildStream/buildstream/merge_requests/899/ | 11:58 |
| * Kinnison takes a peek | 11:58 | |
| jennis | The line lengths may need changing | 11:59 |
| Kinnison | reviewed | 12:01 |
| jennis | Kinnison: regarding the unrelated fix, I did add this to the commit message? | 12:02 |
| jennis | or should we still have this separate | 12:03 |
| Kinnison | Separate commits please | 12:03 |
| Kinnison | same MR is fine | 12:03 |
| Kinnison | cs-shadow: this MANIFEST.in problem is making me very sad. I'm half tempted to replace the entire thing with "include-recursive . *" | 12:06 |
| *** catonano has joined #buildstream | 12:07 | |
| jennis | Kinnison, changes made | 12:07 |
| * Kinnison dislikes the whitespace cleanup being in a link correction commit | 12:08 | |
| Kinnison | think about it, git annotate now says that the whitspace cleanup is related to a link correction | 12:08 |
| Kinnison | I'd avoid the whitespace cleanup unless changing the text around it | 12:09 |
| Kinnison | jennis: also you don't appear to have taken on board my comment about rewording the explanation of /CN | 12:09 |
| *** lachlan has quit IRC | 12:20 | |
| *** phildawson has joined #buildstream | 12:28 | |
| tiagogomes | When stating multiple artifacts, if a path in one artifact is a symbolic link, and it is something else in another artifact; what happens? | 12:28 |
| tiagogomes | *staging | 12:29 |
| laurence | do people get an email notification when someone gives the 'thumbs up' to a gitlab comment?> | 12:33 |
| laurence | have been giving a few recently and am hoping i'm not creating spam mail | 12:33 |
| cs-shadow | laurence: I don't think that it generates email notifications, at least I don't see them in my inbox | 12:41 |
| cs-shadow | Kinnison: It's rather sad but once it is fixed, it should be relatively easier to keep it up to date | 12:42 |
| Kinnison | cs-shadow: yeah except to avoid a screenful of crud, I ended up with things like "recursive-include tests *" which is a bit sadmaking | 12:43 |
| valentind | tiagogomes, That depends. | 12:45 |
| valentind | tiagogomes, Is it a directory in the other artifact? | 12:45 |
| valentind | tiagogomes, And does it contain something? | 12:46 |
| valentind | If it is not a directory it will be an overlap. If it is a directory and the symlink is staged first, then no problem will happen. If it is a directory containing files and those files are staged before the symlink, then you will get an error message. | 12:48 |
| laurence | cs-shadow, cool, thanks | 12:51 |
| Kinnison | cs-shadow: I think I've finally bottomed the MANIFEST.in stuff, I've pushed what I hope will be the last time I need to frob it | 12:51 |
| Kinnison | cs-shadow: If you're okay with the changes, I'd appreciate a final OK on https://gitlab.com/BuildStream/buildstream/merge_requests/892 (obv. it needs to finish the tests successfully too) | 13:01 |
| cs-shadow | Kinnison: was just looking at it :) | 13:01 |
| Kinnison | \o/ | 13:01 |
| jennis | Kinnison, !899 now includes your comments | 13:02 |
| * Kinnison goes to re-review | 13:02 | |
| jennis | Thanks | 13:02 |
| Kinnison | jennis: You can have a thumbsup | 13:03 |
| jennis | why thanks | 13:03 |
| * Kinnison is impressed that laurence understood the SSL implications well enough to review too :-D | 13:03 | |
| jmac | tiagogomes, valentind: I've tried to capture the rules for this at https://wiki.gnome.org/Projects/BuildStream/ImportRules | 13:03 |
| tiagogomes | thanks jmac and valentind | 13:04 |
| jennis | I can see laurence as a participant but can't see his participation | 13:05 |
| jennis | oh, a thumbs up | 13:05 |
| *** lachlan has joined #buildstream | 13:10 | |
| Kinnison | Oh for pity's sake | 13:15 |
| Kinnison | the coverage isn't working | 13:15 |
| * Kinnison sobs | 13:15 | |
| Kinnison | tlater[m], cs-shadow: If either of you have any idea how to resolve the newly exposed coverage issue (https://gitlab.com/BuildStream/buildstream/-/jobs/112652068) I'd appreciate pointers. I am entirely beyond my comfort zone here :( | 13:20 |
| * cs-shadow is taking a look but not sure what's going wrong | 13:26 | |
| tlater[m] | Kin | 13:27 |
| tlater[m] | whoops | 13:27 |
| tlater[m] | Heh | 13:27 |
| tlater[m] | Kinnison: I won't have time today, but I'll try to merge it over the weekend if nobody else comes up with a solution by then. | 13:28 |
| jennis | Has anyone seen an error like this before: OPENSSL_internal:WRONG_VERSION_NUMBER when trying to pull from a remote cache | 13:28 |
| Kinnison | tlater[m]: Okay, thanks. Hopefully someone will spot it before then :/ | 13:28 |
| jennis | Handshake failed with fatal error SSL_ERROR_SSL: error:100000f7:SSL routines:OPENSSL_internal:WRONG_VERSION_NUMBER | 13:29 |
| Kinnison | jennis: That sounds like the python ssl library you have is built against a different libssl to the one found at runtime | 13:29 |
| jennis | uhh ok | 13:29 |
| tlater[m] | jennis: That can only happen if you're not using distro packages for either of the two - I'd say that is more likely to be an issue on your server. | 13:30 |
| jennis | tlater[m], distro packages of BuildStream? | 13:31 |
| tlater[m] | jennis: Distro packages for openssl/python | 13:32 |
| jennis | oh | 13:32 |
| jennis | (: | 13:32 |
| gitlab-br-bot | jennis merged MR !899 (jennis/patch_remote_cache_docs->master: Improve our documentation for configuring an artifact server) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/899 | 13:35 |
| jennis | mhmm tlater server has it | 13:35 |
| tlater[m] | jennis: By "it" you mean the error? | 13:36 |
| jennis | tlater[m], no, `apt install python-openssl` -> nothing needed to be done | 13:37 |
| cs-shadow | Kinnison: I suspect we should be copying ".coverage" instead of/in addition to ".coverage.*" but I am trying to understand how it was working before. | 13:39 |
| tlater[m] | jennis: The issue is that `python-openssl` uses openssl libraries, and therefore needs to be compiled against something on your machine. Distro packages are downloaded pre-compiled, and it is assumed that you only use other distro packages. If you use a non-distro package, for example for openssl, then suddenly `python-openssl` will try to talk to something it doesn't know how to talk to. | 13:39 |
| tlater[m] | And that's what is probably causing your error. | 13:39 |
| tlater[m] | My other guess is mismatching protocols (i.e. one side uses an ancient version of a library involved), but I think that would be a different error - also I trust Kinnison knows enough about these errors to make an accurate guess ;) | 13:40 |
| jennis | considering the client I'm using, I wouldn't be surprised if it's the latter | 13:41 |
| jennis | thanks tlater[m] | 13:41 |
| tlater[m] | jennis: To figure out if they mismatch, write a small python script that does some openssl thing and see if it breaks | 13:41 |
| cs-shadow | Kinnison: can you please try changing "cp .coverage.* ..." to "cp .coverage ..." here: https://gitlab.com/BuildStream/buildstream/blob/4247cef3985b4fa0a15ed02b7beae1dc1a9cd8f1/.gitlab-ci.yml#L81 | 13:44 |
| Kinnison | cs-shadow: I just rebooted, let me get my VM back up | 13:45 |
| cs-shadow | sure | 13:45 |
| Kinnison | Okay I've pushed that, let's cross fingers :-D | 13:47 |
| * Kinnison nips to a brief meeting while CI runs | 13:47 | |
| * cs-shadow will keep an eye on the CI | 13:48 | |
| gitlab-br-bot | BenjaminSchubert opened issue #728 (SandboxBwrap incorrectly checks its arguments type for command) on buildstream https://gitlab.com/BuildStream/buildstream/issues/728 | 13:51 |
| *** lachlan has quit IRC | 14:02 | |
| raoul | Think I've got remote job cancellation working, but it probably needs an integration test. Would having an integration test that runs if buildgrid is installed be alright? | 14:03 |
| *** toscalix has joined #buildstream | 14:04 | |
| mablanch | raoul: At the moment, the REAPI client is not tested, here is the opened issue tracking that: https://gitlab.com/BuildStream/buildstream/issues/629 | 14:06 |
| *** brlogger has joined #buildstream | 15:08 | |
| *** gitlab-br-bot has joined #buildstream | 15:16 | |
| *** lachlan has joined #buildstream | 15:16 | |
| Kinnison | Nice | 15:17 |
| * Kinnison hi5s cs-shadow | 15:17 | |
| * cs-shadow hi5s Kinnison | 15:17 | |
| Kinnison | As a word of warning to anyone with open MRs, when you rebase for merge, be aware you may get new lint warnings \o/ | 15:18 |
| gitlab-br-bot | raoul.hidalgocharman opened MR !900 (725-job-cancellation-on-remote-builds->master: WIP: _sandboxremote.py: Add sigterm handler that sends CancelOperation) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/900 | 15:19 |
| *** anahuelamo- has joined #buildstream | 15:23 | |
| *** anahuelamo has quit IRC | 15:23 | |
| WSalmon | this looks quite interesting anyone else have any feed back for ben? https://gitlab.com/BuildStream/buildstream/issues/728 tlater[m] skullman tiagogomes | 15:29 |
| *** lachlan has quit IRC | 15:32 | |
| bochecha | I have a question which isn't directly related to buildstream, but I hope nobody will mind, it's about how builstream uses bwrap in its CI :) | 15:34 |
| bochecha | whenever i try using bwrap in docker, I get "bwrap: No permissions to creating new namespace, likely because the kernel does not allow non-privileged user namespaces. On e.g. debian this can be enabled with 'sysctl kernel.unprivileged_userns_clone=1'." | 15:34 |
| bochecha | does the Buildstream CI do anything special to make this work? | 15:34 |
| skullman | I think it may run as root, since we use the same kind of container to run the UNIX tests, and they *do* have to run as root. | 15:41 |
| jjardon | bochecha: the docker image runs with privileged permissions | 15:54 |
| jjardon | bochecha: this is the config of the runners: | 15:56 |
| jjardon | https://www.irccloud.com/pastebin/MqOwaTa0/ | 15:56 |
| *** lachlan has joined #buildstream | 15:58 | |
| *** lachlan has quit IRC | 16:11 | |
| *** abderrahim4 has joined #buildstream | 16:15 | |
| *** abderrahim3 has quit IRC | 16:16 | |
| *** lachlan has joined #buildstream | 16:17 | |
| *** xjuan has joined #buildstream | 16:31 | |
| benschubert | Would people be against using pytest-timeout (https://pypi.org/project/pytest-timeout/) to set a per-test timeout? Some tests do block indefinitely for some reasons and that's quite hard to debug :/ | 16:33 |
| benschubert | I'll open an issue if no strong opposition from here :) | 16:33 |
| tpollard | looks interesting for debugging | 16:35 |
| bochecha | jjardon: ok, that explains it then… unfortunately I don't control those runners, so I guess I just won't be able to do what I needed :( | 16:37 |
| bochecha | jjardon: thanks for confirming | 16:37 |
| *** bochecha has quit IRC | 16:41 | |
| gitlab-br-bot | tacgomes approved MR !898 (bschubert/fix-command-sandbox->master: WIP: Check is command is a str and replace by list before checking existence) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/898 | 16:42 |
| raoul | I'm trying to add messages to a SIGTERM handler in _sandboxremote when it's called, via the context and then the message function, but they don't seem to get to output. Anyone got any idea why and how to get around it? | 16:44 |
| *** tpollard has quit IRC | 16:46 | |
| *** raoul has quit IRC | 16:50 | |
| *** toscalix has quit IRC | 16:54 | |
| *** Prince781 has joined #buildstream | 16:57 | |
| gitlab-br-bot | cs-shadow closed issue #713 (Don't redundently document directory configuration in plugins) on buildstream https://gitlab.com/BuildStream/buildstream/issues/713 | 16:59 |
| gitlab-br-bot | cs-shadow merged MR !896 (chandan/source-doc-directory->master: plugins: Don't redundantly document common configurations) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/896 | 16:59 |
| *** jonathanmaw has quit IRC | 17:03 | |
| *** phildawson_ has joined #buildstream | 17:04 | |
| *** phildawson has quit IRC | 17:05 | |
| *** anahuelamo- has quit IRC | 17:05 | |
| *** anahuelamo has joined #buildstream | 17:06 | |
| *** anahuelamo has quit IRC | 17:07 | |
| *** anahuelamo has joined #buildstream | 17:07 | |
| *** Prince781 has quit IRC | 17:10 | |
| *** catonano has quit IRC | 17:12 | |
| *** loganlee has joined #buildstream | 17:13 | |
| *** Prince781 has joined #buildstream | 17:15 | |
| *** catonano has joined #buildstream | 17:19 | |
| *** Prince781 has quit IRC | 17:33 | |
| *** Prince781 has joined #buildstream | 17:40 | |
| gitlab-br-bot | BenjaminSchubert opened issue #730 (Show each warning only once when running tests) on buildstream https://gitlab.com/BuildStream/buildstream/issues/730 | 17:56 |
| *** Prince781 has quit IRC | 18:02 | |
| *** tuxcanfly has joined #buildstream | 18:06 | |
| *** Prince781 has joined #buildstream | 18:10 | |
| *** Prince781 has quit IRC | 18:18 | |
| *** alatiera_ has quit IRC | 19:01 | |
| *** catonano has quit IRC | 19:01 | |
| *** catonano has joined #buildstream | 19:17 | |
| *** Prince781 has joined #buildstream | 20:34 | |
| *** cs-shadow has quit IRC | 20:40 | |
| *** bochecha has joined #buildstream | 21:26 | |
| *** xjuan has quit IRC | 21:44 | |
| *** Prince781 has quit IRC | 21:46 | |
| *** Prince781 has joined #buildstream | 22:39 | |
| *** Prince781 has quit IRC | 23:14 | |
| *** catonano has quit IRC | 23:35 | |
| *** Caerus has joined #buildstream | 23:44 | |
| *** bochecha has quit IRC | 23:53 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!