| *** alatiera_ has joined #buildstream | 00:03 | |
| *** bochecha has quit IRC | 02:37 | |
| *** alatiera_ has quit IRC | 02:37 | |
| *** xjuan has quit IRC | 02:48 | |
| *** Guest45614 has joined #buildstream | 04:45 | |
| *** jo_sch has joined #buildstream | 06:39 | |
| *** finn has joined #buildstream | 08:54 | |
| *** toscalix has joined #buildstream | 08:59 | |
| *** toscalix has quit IRC | 09:00 | |
| *** alatiera_ has joined #buildstream | 09:21 | |
| *** toscalix has joined #buildstream | 09:23 | |
| *** benschubert has joined #buildstream | 09:31 | |
| *** jonathanmaw has joined #buildstream | 09:38 | |
| *** rdale has joined #buildstream | 10:02 | |
| gitlab-br-bot | jonathanmaw approved MR !918 (valentindavid/fix-recursive-strip-debug->master: Fix infinite recursion in default strip debug command) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/918 | 10:11 |
|---|---|---|
| gitlab-br-bot | juergbi approved MR !927 (jmac/make_writable_combination->master: Use _force_rmtree instead of custom code.) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/927 | 10:13 |
| *** alatiera__ has joined #buildstream | 10:14 | |
| adds68 | It seemed the answer to my question yesterday, was to call the "translate_url" | 10:15 |
| adds68 | source.translate_url(source.mirror.url, alias_override=None, | 10:15 |
| adds68 | primary=source.mirror.primary) | 10:15 |
| adds68 | returns the resolved URL \o/ | 10:15 |
| *** alatiera_ has quit IRC | 10:16 | |
| *** raoul has joined #buildstream | 10:19 | |
| *** alatiera_ has joined #buildstream | 10:22 | |
| *** alatiera__ has quit IRC | 10:24 | |
| *** lachlan has joined #buildstream | 10:27 | |
| Nexus | Is it possible to get use a string of an element "foo.bst" in _stream.py and get back the element of that name as an Element object? | 10:39 |
| jennis | Hi guys, I've got a MR ready for review, it's a docs patch regarding our examples of declaring the remote caches in project.conf and user config, would someone have some time to review this? | 10:39 |
| jennis | adds68? | 10:39 |
| jennis | https://gitlab.com/BuildStream/buildstream/merge_requests/932 | 10:39 |
| *** jonathanmaw_ has joined #buildstream | 10:41 | |
| *** jonathanmaw has quit IRC | 10:41 | |
| juergbi | Nexus: that's normally loader territory, not sure whether we have access to that in stream. | 10:42 |
| juergbi | is this related to the validation issue? | 10:42 |
| Nexus | yep | 10:42 |
| Nexus | i've bee ntrying various ways of getting access to the node data earlier, and that seemed lik ethe only way to me | 10:42 |
| juergbi | Nexus: have you considered my suggestion of simply copying the node_validate to the first load method in project? (instead of actual copy it should be in a helper function) | 10:43 |
| Nexus | yes, but i still need the node data, which requires a project object | 10:43 |
| Nexus | which isn't generated until after the elements have been loaded afaik | 10:44 |
| juergbi | Nexus: does pre_config_node not work for this? | 10:44 |
| juergbi | I'm talking about Project._load() | 10:45 |
| juergbi | nothing in Stream | 10:45 |
| Nexus | ah ok, i'll have to look | 10:45 |
| Nexus | i've only considered stream._load(0 | 10:45 |
| *** lachlan has quit IRC | 10:45 | |
| adds68 | jennis, you link to 2 documents that explain how the artifact cache is configured, but they are not linked in the MR | 10:46 |
| adds68 | jennis, where can i find them? | 10:46 |
| juergbi | Nexus: my guess is that this issue was introduced by commit 171e803f (@ include support) where node_validate was moved from Project._load() to the new Project._load_second_pass(). maybe it should have been copied instead of moved | 10:47 |
| juergbi | Nexus: in _load() node_validate should be right after the format_version check | 10:48 |
| jennis | adds68, huh? | 10:48 |
| gitlab-br-bot | aevri opened MR !935 (aevri/contributing_gitlab->master: WIP: Add more to GitLab-relevant parts of contributing) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/935 | 10:48 |
| jennis | the links our for our current documentation | 10:48 |
| jennis | And that's what I've replaced | 10:48 |
| *** alatiera_ has quit IRC | 10:49 | |
| jennis | Essentially I've got rid of the examples in 'Configuring an artifact cache', and added the examples regarding how to declare a cache in the project config documentation and the user config documentation. As there are different ways to declare the caches and also different bst behaviours depending on where you declare it | 10:49 |
| adds68 | jennis, i wanted to review the new links, but it seems the bst docs generate the links to the docs on the fly | 10:50 |
| adds68 | jennis, these documents for example: 1. The :ref:`project configuration <project_essentials_artifacts>`, or | 10:50 |
| adds68 | 2. The :ref:`user configuration <config_artifacts>`. | 10:50 |
| jennis | ahh yeah, you'll have to clone the branch, `make -C doc` | 10:51 |
| jennis | *checkout | 10:51 |
| *** alatiera_ has joined #buildstream | 10:54 | |
| *** alatiera__ has joined #buildstream | 10:57 | |
| *** alatiera_ has quit IRC | 10:58 | |
| Nexus | juergbi: \/ woo | 11:00 |
| Nexus | \o/ * | 11:00 |
| juergbi | :) | 11:00 |
| *** alatiera_ has joined #buildstream | 11:01 | |
| *** alatiera__ has quit IRC | 11:03 | |
| *** alatiera_ has quit IRC | 11:03 | |
| *** alatiera_ has joined #buildstream | 11:03 | |
| gitlab-br-bot | richardmaw-codethink approved MR !932 (jennis/docs_patch->master: Update our instructions/examples regarding how we declare remote caches) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/932 | 11:06 |
| *** alatiera_ has quit IRC | 11:08 | |
| toscalix | tristan will not be working this week. He will come back the next one. | 11:09 |
| gitlab-br-bot | adds68 approved MR !932 (jennis/docs_patch->master: Update our instructions/examples regarding how we declare remote caches) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/932 | 11:12 |
| jennis | two approvals :O thanks skullman, adds68 | 11:13 |
| adds68 | jennis, for some weird reason when i pulled your branch it was still using the old documentation | 11:13 |
| jennis | mhmm how did you open the documentation? | 11:13 |
| jennis | s/open/view | 11:13 |
| adds68 | jennis, i ran the make file then opened the html folder in firefox | 11:14 |
| jennis | ahh weird, did the `make -C doc` defiitely complete? | 11:14 |
| jennis | Sometimes it fails if you've got some legacy stuff lying around in the doc/ dir | 11:14 |
| adds68 | jennis, yea after a few attemps due to missing dependencies | 11:15 |
| jennis | That's odd | 11:15 |
| adds68 | jennis, once i rebase the branch it seemed to work though | 11:15 |
| jennis | right, ok glad to hear it | 11:15 |
| benschubert | juergbi: would you have time to look into https://gitlab.com/BuildStream/buildstream/merge_requests/921 ? There are performance implications on this test so I would like a maintainer to have a look at it :) | 11:17 |
| juergbi | sure, will take a look | 11:17 |
| tiagogomes | I can't disable the use of artifact servers through command line right | 11:28 |
| gitlab-br-bot | jonathanmaw approved MR !933 (danielsilverstone-ct/further-optimisations->master: WIP: Some more microoptimisations) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/933 | 11:29 |
| jennis | tiagogomes, not AFAIK | 11:29 |
| *** lachlan has joined #buildstream | 11:30 | |
| jennis | The only remote options we have are for pull and push. E.g. bst pull --remote <url> foo.bst | 11:30 |
| jennis | Again, AFAIK ^ | 11:31 |
| gitlab-br-bot | richardmaw-codethink merged MR !932 (jennis/docs_patch->master: Update our instructions/examples regarding how we declare remote caches) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/932 | 11:32 |
| *** rdale has quit IRC | 11:32 | |
| *** alatiera__ has joined #buildstream | 11:35 | |
| laurence | toscalix, we have the BuildStream gitlab group 'inheriting' BuildGrid milestones at this level - https://gitlab.com/groups/BuildStream/-/milestones | 11:36 |
| laurence | Which doesn't seem right | 11:36 |
| laurence | looking at BuildGrid's gitlab here - https://gitlab.com/BuildGrid/buildgrid/project_members, it tells us that BuildStream is a 'Groups with access to buildgrid' | 11:36 |
| laurence | That's the only place I can find where the two are linked at group level | 11:37 |
| laurence | So I will remove that link later today unless anyone has any objections | 11:37 |
| *** lachlan has quit IRC | 11:37 | |
| gitlab-br-bot | LaurenceUrhegyi approved MR !935 (aevri/contributing_gitlab->master: WIP: Add more to GitLab-relevant parts of contributing) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/935 | 11:38 |
| *** rdale has joined #buildstream | 11:47 | |
| toscalix | laurence: go ahead | 11:49 |
| valentind | jjardon, I have set up the aarch64 builders for BuildStream. I wonder where I can put the ansible configuration. In codethink's gitlab maybe? It contains some private information. | 11:50 |
| valentind | What do you think? | 11:50 |
| toscalix | valentind: we do not have a good place for that. | 11:51 |
| laurence | toscalix, ok, have done and yes, it worked - https://gitlab.com/groups/BuildStream/-/milestones | 11:51 |
| toscalix | I will think/ask how to do it so we do not need a whole new set up internally | 11:52 |
| toscalix | laurence: great | 11:52 |
| valentind | toscalix, OK thanks. | 11:52 |
| toscalix | laurence: buildgrid repo is not listed here anymore: https://gitlab.com/BuildStream | 11:53 |
| laurence | ah, was it previously? I didn't know | 11:54 |
| laurence | toscalix, I will think/ask how to do it so we do not need a whole new set up internally - what will this impact? something i've overlooked? | 11:54 |
| toscalix | laurence: it was an answer to valentind | 11:55 |
| jjardon | valentind: createa a https://gitlab.com/BuildStream/infrastructure/ group and create a repo there with the ansible? | 11:56 |
| toscalix | jjardon: let's not call it infrastructure please | 11:57 |
| *** alatiera_ has joined #buildstream | 11:57 | |
| toscalix | there are already other projects with the same name | 11:57 |
| toscalix | in other groups | 11:57 |
| jjardon | valentind: then try to abstract the secret parts and document where you can get them if needed (probably basically the contact for the infra team) | 11:58 |
| toscalix | I would also like to have everything related with the infra and associated servidces in a subgroup | 11:58 |
| jjardon | toscalix: what do you mean? | 11:58 |
| *** alatiera__ has quit IRC | 11:58 | |
| jjardon | toscalix: yeah, that is the idea a infrastructure subgroup in the Buildstream group | 11:58 |
| benschubert | Could someone point me where the actual caching of artifacts is done in the code? I've got trouble finding that | 11:59 |
| toscalix | infrastructure-buildstream repo or a name that is nor infrastructure. those who work on several gitlab.com groups have other repos called infrastructure | 11:59 |
| *** bochecha has joined #buildstream | 11:59 | |
| laurence | toscalix, ok. | 12:01 |
| jjardon | toscalix: so you suggest to create a https://gitlab.com/BuildStream/buildstream-infrastructure/ group instead https://gitlab.com/BuildStream/infrastructure/ ? | 12:02 |
| toscalix | jjardon: we can even expand that subgroup beyond infrastructure like for instance, other services we might need in the future | 12:04 |
| toscalix | maybe buildstream-services-infra as subgroup name? | 12:04 |
| toscalix | and inside the subgroup.... buildstream-infrastructure as repo ? | 12:05 |
| *** alatiera__ has joined #buildstream | 12:05 | |
| toscalix | valentind: https://gitlab.com/BuildStream/nosoftware/alignment/issues/41 | 12:06 |
| jjardon | toscalix: we should move the buildstream-docker-image repo to that subgroup, for example | 12:06 |
| toscalix | is that a plugin? | 12:07 |
| jjardon | no | 12:07 |
| *** alatiera_ has quit IRC | 12:07 | |
| toscalix | deployment? | 12:07 |
| jjardon | is the repo that generate the docker images we use to the buildstream CI | 12:07 |
| jjardon | the images are pushed here: https://hub.docker.com/u/buildstream/ | 12:08 |
| *** finn has quit IRC | 12:08 | |
| toscalix | maybe the name I proposed is not good then because I was thinking as services things like AWS storage or a forum service....things like that | 12:08 |
| jjardon | toscalix: ot sure those things will be in a git repo | 12:09 |
| toscalix | configs | 12:09 |
| toscalix | credentials..... | 12:10 |
| *** alatiera_ has joined #buildstream | 12:10 | |
| toscalix | let's not stop the creation of the subgroup | 12:11 |
| toscalix | I will open a ticket about buildstream-docker-image repo | 12:11 |
| toscalix | and think about it. We need to put some order/structure in our repos | 12:12 |
| jmac | benschubert: Artifacts are added to the cache with a call to commit at https://gitlab.com/BuildStream/buildstream/blob/master/buildstream/element.py#L1683 | 12:12 |
| toscalix | specially when thinking about the coming plugins discussion | 12:12 |
| jmac | benschubert: Is that what you were looking for, or something else like pushing? | 12:12 |
| *** alatiera__ has joined #buildstream | 12:13 | |
| *** alatiera_ has quit IRC | 12:13 | |
| *** alatiera_ has joined #buildstream | 12:14 | |
| *** alatiera__ has quit IRC | 12:16 | |
| *** alatiera__ has joined #buildstream | 12:17 | |
| toscalix | jjardon: https://gitlab.com/BuildStream/nosoftware/alignment/issues/42 | 12:17 |
| *** finn has joined #buildstream | 12:18 | |
| *** alatiera_ has quit IRC | 12:18 | |
| benschubert | jmac: that was it thanks! | 12:18 |
| jennis | tlater[m] around? | 12:20 |
| *** joshproehl has joined #buildstream | 12:25 | |
| *** lachlan has joined #buildstream | 12:25 | |
| gitlab-br-bot | jmacarthur closed issue #735 (Combine _force_rmtree and make_dir_writable) on buildstream https://gitlab.com/BuildStream/buildstream/issues/735 | 12:26 |
| gitlab-br-bot | jmacarthur merged MR !927 (jmac/make_writable_combination->master: Use _force_rmtree instead of custom code.) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/927 | 12:26 |
| *** lachlan has quit IRC | 12:29 | |
| gitlab-br-bot | jjardon opened MR !936 (jjardon/aarch64->master: WIP: .gitlab-ci.yml: Add jobs to run tests in aarch64) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/936 | 12:37 |
| *** lachlan has joined #buildstream | 12:51 | |
| tiagogomes | Does anyone tried to run buildstream inside bwrap? I am getting a problem with the FUSE layer | 12:57 |
| tpollard | skullman? ^ | 12:57 |
| tpollard | sounds familiar with the dogfooding stuff | 12:58 |
| tiagogomes | /dev/fuse didn't make it to the chroot | 13:02 |
| tiagogomes | I guess bwrap has a whitelist of device nodes to bind on --dev | 13:03 |
| skullman | yeah, and you need to specify `--dev-bind /dev/fuse /dev/fuse` to get it to include fuse | 13:06 |
| skullman | https://gitlab.com/BuildStream/buildstream-integration/blob/master/README.md has some options for doing it | 13:07 |
| tiagogomes | Already passed that stage. I get now a fusermount: mount failed: Operation not permitted | 13:08 |
| skullman | are you running as root or with a user namespace? | 13:08 |
| tiagogomes | user namespace | 13:08 |
| skullman | is your version of linux >= 4.18? | 13:09 |
| tiagogomes | yes | 13:09 |
| tiagogomes | 4.18.16-300 | 13:09 |
| skullman | and bubblewrap --version > 0.3.0 ? | 13:10 |
| tiagogomes | yes 0.3.1 | 13:10 |
| skullman | I'm out of ideas | 13:12 |
| * skullman had some other mount -EPERM problems, but that was from the system already having stuff mounted over /proc | 13:13 | |
| skullman | what's the full command-line you're running? | 13:14 |
| tiagogomes | `bwrap --bind / / --dev /dev --dev-bind /dev/fuse /dev/fuse --proc /proc --unshare-net /bin/bash` | 13:16 |
| * tiagogomes points out the ultimate goal here is to disable the use of artifact servers, even on parent projects | 13:17 | |
| tiagogomes | So if there's an easy way of achieving that… | 13:17 |
| gitlab-br-bot | knownexus opened MR !937 (element-path_not_validated->master: Element path not validated before use) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/937 | 13:18 |
| gitlab-br-bot | BenjaminSchubert opened MR !938 (bschubert/fix-atomic-move-git-repo->master: WIP: Fix os.rename in git source element to correctly handle error codes) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/938 | 13:18 |
| Nexus | juergbi: Please coule you quickly review https://gitlab.com/BuildStream/buildstream/merge_requests/937 for me? | 13:18 |
| skullman | tiagogomes: not sure, jennis was looking at artifact server configuration recently, so might have some ideas | 13:19 |
| tpollard | I don't think it's currently possible, but it sounds like something bst should be able to do | 13:20 |
| *** finn_ has joined #buildstream | 13:24 | |
| tpollard | tiagogomes: you could remove the check for 'if self._artifacts.has_fetch_remotes()' in the build function of _stream.py | 13:24 |
| tpollard | so it won't spawn any pull queues, and only fetch | 13:24 |
| tpollard | I think | 13:24 |
| *** finn has quit IRC | 13:25 | |
| tpollard | sames goes for has_push_remotes() | 13:26 |
| tiagogomes | Modifying buildstream is the last solution | 13:26 |
| tpollard | then my suggestion would be to file an issue to implement a user option/ cli option for doing that | 13:27 |
| * skullman wonders if adding `scheduler: { fetchers: 0 }` would work | 13:28 | |
| skullman | to your buildstream.conf I mena | 13:28 |
| skullman | *mean | 13:28 |
| tiagogomes | hm | 13:29 |
| tpollard | you can set fetchers as a main cli option to bst too | 13:30 |
| tpollard | same for pushers | 13:31 |
| tiagogomes | Doesn't work :/ | 13:33 |
| tiagogomes | Ok: `sudo unshare -n; su $me will do it` | 13:38 |
| skullman | fair, bwrap needs to do a lot more heavy lifting mount-namespace-wise which can cause other complications, which if you just want to turn networking off are unnecessary | 13:41 |
| tpollard | would you like me to raise an issue for adding it as an option to buildstream tiagogomes? to me it has merit to check that a project can always be built from source, especially if you don't have control over your junctions | 13:41 |
| *** jonathanmaw_ is now known as jonathanmaw | 13:50 | |
| tpollard | valentind: if my comments about the cached buildtree workspaces are not clear on the ticket, I can try to discuss on here or another method | 13:51 |
| jonathanmaw | hrm, I'm looking at using a .bstproject.yaml file in a workspace to hold information about the project | 13:52 |
| jonathanmaw | Is there a way to exclude a particular file when mounting a workspace? | 13:52 |
| jonathanmaw | I've tried mounting the subdirectories of the sandbox and excluding .bstproject.yaml, but that a) doesn't seem to let the changes while it was mounted get back to the workspace, and b) Isn't going to be helpful if you add files to the root of the workspace | 13:53 |
| jonathanmaw | is there an alternative I've missed? | 13:53 |
| laurence | benschubert, just an FYI Daniel isn't around today | 13:53 |
| laurence | benschubert, nor is tristan this week, in fact | 13:54 |
| jonathanmaw | if not, I might be stuck removing the .bstproject.yaml file for the duration of the build, and that's not going to play well if buildstream is suddenly terminated | 13:54 |
| benschubert | laurence: oh thank you! Will Daniel be there tomorrow? | 13:54 |
| skullman | jonathanmaw: not really, the way directories and mounting work in Unix-likes. The closest you could get is an overlayfs mount of some form | 13:55 |
| laurence | benschubert, he will, yes | 13:55 |
| skullman | though that would make propagating the changes back harder | 13:55 |
| laurence | np ! | 13:55 |
| jjardon | tpollard: please, I'm interested on the feature tiagogomes is requesting as well | 13:55 |
| * jonathanmaw wonders what the SafeHardlinks FUSE layer gets up to, and whether that'd be useful | 13:56 | |
| benschubert | Thanks! | 13:56 |
| skullman | you could theoretically use it as your overlayfs to remove the file, but then you'd have to copy the changes back on top of your workspace yourself | 13:56 |
| tpollard | jonathanmaw: is there any dependence on how specific source plugins implement init_workspace? | 13:58 |
| jonathanmaw | tpollard: I haven't touched init_workspace. my attempts here have been in Element._stage_sources_in_sandbox() | 13:59 |
| tpollard | ah, I presumed Element._open_workspace | 14:01 |
| tpollard | jjardon: ok sure | 14:02 |
| tiagogomes | tpollard yes please | 14:13 |
| tiagogomes | tpollard maybe we don't need a new option, but rather that `num-fetchers: 0` disables the artifact caches | 14:18 |
| gitlab-br-bot | tpollard opened issue #752 (Add ability to build without interacting with available artifact servers) on buildstream https://gitlab.com/BuildStream/buildstream/issues/752 | 14:19 |
| tpollard | tiagogomes: that would still leave pushers, but yes extending that could be the option | 14:20 |
| tpollard | made note of it | 14:21 |
| *** lachlan has quit IRC | 14:27 | |
| Nexus | my tests are failing and i'm not sure why, i didn't touch the areas that are failing https://gitlab.com/BuildStream/buildstream/-/jobs/117520183 | 14:28 |
| *** lachlan has joined #buildstream | 14:31 | |
| benschubert | https://gitlab.com/BuildStream/buildstream/merge_requests/895 for review :D | 14:32 |
| tpollard | Looks like you've broken something that handles the @ include from configuration Nexus | 14:32 |
| Nexus | and once again in english? :) | 14:33 |
| tpollard | precisely that | 14:36 |
| Nexus | tpollard: unless there's a file called "configuration" that i can't see, i need more information | 14:38 |
| Nexus | "precisely that" isn't helpful if i don't understand what you're referring to | 14:38 |
| raoul | Another poke to get https://gitlab.com/BuildStream/buildstream/merge_requests/900/ merged. Laurence said this may be your area juergbi | 14:39 |
| tpollard | Nexus: near enough every failure on those tests is coming from failing to parse project.conf '@' | 14:40 |
| *** lachlan has quit IRC | 14:42 | |
| tpollard | Nexus: which I'm presuming is https://docs.buildstream.build/format_intro.html?highlight=append#include | 14:43 |
| juergbi | raoul: adding it to my list | 14:54 |
| raoul | thanks :) | 14:54 |
| juergbi | benschubert: commented on !921, hope this makes sense to you as well | 14:54 |
| gitlab-br-bot | MR !921: test behavior when there is no sandbox on Linux https://gitlab.com/BuildStream/buildstream/merge_requests/921 | 14:54 |
| *** bjurr has quit IRC | 14:55 | |
| Nexus | juergbi: could yougive me a hand please? | 14:56 |
| Nexus | for some reason, adding _yaml.node_validate to Project._load(), stops buildstream being able to parse '@' | 14:57 |
| juergbi | Nexus: right, so as this is before include processing, (@) keys are still in there | 14:57 |
| juergbi | not quite sure what the best approach is here | 14:57 |
| juergbi | we might be able to simply add '(@)' to the list of allowed keys in the first check | 14:58 |
| benschubert | juergbi: thanks a lot! I will move the check in linux.py and remove that cache! It will indeed be way cleaner | 14:58 |
| juergbi | however, it feels a bit odd | 14:58 |
| juergbi | Nexus: maybe valentind has a better idea (or alternatively, can confirm that simply adding '(@)' makes sense) as he wrote support for '(@)' | 15:01 |
| skullman | huh, that's weird, `kind: manual` isn't flagged as supporting `BST_VIRTUAL_DIRECTORY` | 15:02 |
| Nexus | juergbi: i would need to add both '(@)' and 'siurces' | 15:02 |
| Nexus | sources* | 15:02 |
| *** lachlan has joined #buildstream | 15:03 | |
| juergbi | Nexus: right, but 'sources' feels less odd. and actually, we already include 'elements' even though it's processed the same way as 'sources', afaict | 15:05 |
| Nexus | that seems to have fixed it | 15:05 |
| juergbi | valentind: can you please provide your input on the above discussion between Nexus and me with regards to !937 (related to include work) | 15:05 |
| gitlab-br-bot | MR !937: Element path not validated before use https://gitlab.com/BuildStream/buildstream/merge_requests/937 | 15:05 |
| *** lachlan has quit IRC | 15:06 | |
| juergbi | Nexus: ok, good, that should at least avoid blocking you in this regard | 15:07 |
| Nexus | juergbi: kk, pushed the fix up, please could you review https://gitlab.com/BuildStream/buildstream/merge_requests/925 for me? :) | 15:09 |
| juergbi | put it on my list | 15:10 |
| Nexus | thanks :) | 15:10 |
| juergbi | Nexus: is it still WIP or should it be unmarked? | 15:10 |
| Nexus | well i kindof make things complicated for myself | 15:10 |
| Nexus | because it;s dependent on https://gitlab.com/BuildStream/buildstream/merge_requests/937 | 15:10 |
| Nexus | which is the fix we just did | 15:11 |
| Nexus | but i rebased 925 off of it and pushed that up too | 15:11 |
| Nexus | so assuming 937 is ok, 925 is no longer wip | 15:11 |
| Nexus | if that makes sense? | 15:11 |
| juergbi | ok, I see, makes sense to keep it WIP (unmergable) for now, then | 15:11 |
| juergbi | but there should be a comment in the MR to make this clear | 15:11 |
| juergbi | ah, there is, sorry for the noise | 15:12 |
| Nexus | np :) | 15:12 |
| juergbi | (was only checking the description) | 15:12 |
| Nexus | i'll add it to the desc | 15:12 |
| tpollard | jjardon: OOI with !936, are we expecting the aarch64 testing to act as gating? | 15:20 |
| gitlab-br-bot | MR !936: WIP: .gitlab-ci.yml: Add jobs to run tests in aarch64 https://gitlab.com/BuildStream/buildstream/merge_requests/936 | 15:20 |
| jjardon | tpollard: yes | 15:21 |
| tpollard | interesting | 15:22 |
| juergbi | if this increases CI test time, it might make sense to setup a merge bot | 15:22 |
| *** lachlan has joined #buildstream | 15:27 | |
| jjardon | about that, can anyone please review https://gitlab.com/BuildStream/buildstream-docker-images/merge_requests/70 ? | 15:31 |
| *** lachlan has quit IRC | 15:43 | |
| *** lachlan has joined #buildstream | 15:44 | |
| benschubert | https://gitlab.com/BuildStream/buildstream/merge_requests/938 for review :) | 15:46 |
| laurence | jjardon - I am going through some old non-WIP MRs in the buildstream backlog - feels like this one on docs licensing may just be a small adjustment required to get it to render properly? - https://gitlab.com/BuildStream/buildstream/merge_requests/336 | 15:46 |
| jjardon | laurence: not sure how we can do that copyright shorter; ideas welcomed | 15:48 |
| jjardon | laurence: (and not sure it matters that much) | 15:48 |
| *** lachlan has quit IRC | 15:49 | |
| jjardon | actually, maybe we can say "CC BY 4.0" | 15:49 |
| jjardon | instead "Creative Commons Attribution 4.0" ? | 15:49 |
| jjardon | would that work? | 15:49 |
| toscalix | commentyed on the MR | 15:51 |
| benschubert | jjardon: I see that you have a few branches testing pycodestyle, are they stale or still in progress? | 15:51 |
| jjardon | benschubert: stale ,waiting for fixes in pytest-pycodestyle | 15:51 |
| benschubert | Ok! thanks a lot | 15:52 |
| toscalix | jjardon: if the code is in a specific license and you distribute the doc together with it, the idea is that the docu should have the same license | 15:52 |
| toscalix | a different story is comlpementary docu | 15:52 |
| toscalix | that is how the license section is explained on the website | 15:52 |
| jjardon | benschubert: https://gitlab.com/BuildStream/buildstream/merge_requests/638#note_113559848 | 15:52 |
| laurence | jjardon, hhhmm, looking again, the ellipsis doesn't even cut off any actual text. | 15:52 |
| jjardon | benschubert: of course feel free to submit patches to fix pycodestyle issues you found, they can land independenly | 15:53 |
| Nexus | What's with the "tmp_path" error? is that still a global thing? | 15:53 |
| skullman | I thought I submitted a MR to fix that | 15:54 |
| jjardon | toscalix: oh, so the docs are under LGPL2.1 ? weird. I will update the MR | 15:54 |
| benschubert | jjardon: sure! | 15:54 |
| Nexus | yup, hadn't updated, ta skullman | 15:55 |
| toscalix | jjardon: weird, agree, but if you make code proprietary, like for instance if you intend to charge for a license of the code, you would not be able to distribute the docu | 15:56 |
| toscalix | so you would have to separate them | 15:56 |
| toscalix | the easy path is to license the code and the doc associated under the same licenese if they are meant to be distributed together | 15:56 |
| jjardon | toscalix: you can not make LGPL code propietary | 15:57 |
| toscalix | let me rephrase... if you want to double license the code.... | 15:57 |
| toscalix | you would like to dounble license the docu | 15:59 |
| toscalix | in any case, jjardon that was a decision I took based on previous experience. It can be discussed and changed | 16:01 |
| toscalix | previous experience....normal practice | 16:01 |
| jjardon | sure, I do not mind; It's only a bit weird license doc as LGPL but I see your point (a bit extreme but I guess it can happen) | 16:02 |
| gitlab-br-bot | jonathanmaw approved MR !934 (valentindavid/script-artifact-corruption->master: Fix cache corruption by scripts when layout and integration commands are used) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/934 | 16:02 |
| toscalix | jjardon: with LGPL it does. With some other licenses with some restrictions on the distribution or the usage side (freedem 0) it does too | 16:03 |
| toscalix | LGPL is not an easy license model at all | 16:04 |
| jjardon | toscalix: how so? | 16:04 |
| laurence | jennis, is there a reason a project might not use the YAML cache? | 16:11 |
| laurence | wrt https://gitlab.com/BuildStream/buildstream/merge_requests/933#note_115250626 | 16:11 |
| *** lachlan has joined #buildstream | 16:12 | |
| gitlab-br-bot | jennis opened issue #753 (Cache size logging is very spammy) on buildstream https://gitlab.com/BuildStream/buildstream/issues/753 | 16:19 |
| *** lachlan has quit IRC | 16:25 | |
| gitlab-br-bot | richardmaw-codethink approved MR !934 (valentindavid/script-artifact-corruption->master: Fix cache corruption by scripts when layout and integration commands are used) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/934 | 16:28 |
| gitlab-br-bot | jennis opened issue #754 (Local cache cleaning/expiry takes a really long time and cache quota is not strict) on buildstream https://gitlab.com/BuildStream/buildstream/issues/754 | 16:32 |
| tiagogomes | aah | 16:37 |
| tiagogomes | `fetchers: 0` means no limit | 16:37 |
| *** WSalmon has quit IRC | 16:45 | |
| *** alatiera_ has joined #buildstream | 16:48 | |
| gitlab-br-bot | jjardon closed issue #754 (Local cache cleaning/expiry takes a really long time and cache quota is not strict) on buildstream https://gitlab.com/BuildStream/buildstream/issues/754 | 16:48 |
| *** alatiera__ has quit IRC | 16:50 | |
| gitlab-br-bot | valentindavid closed issue #749 (Script element corrupts cache when using layout and enables integration commands) on buildstream https://gitlab.com/BuildStream/buildstream/issues/749 | 16:59 |
| gitlab-br-bot | richardmaw-codethink closed issue #749 (Script element corrupts cache when using layout and enables integration commands) on buildstream https://gitlab.com/BuildStream/buildstream/issues/749 | 17:00 |
| gitlab-br-bot | richardmaw-codethink merged MR !934 (valentindavid/script-artifact-corruption->master: Fix cache corruption by scripts when layout and integration commands are used) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/934 | 17:00 |
| skullman | juergbi: I could do with an API review on https://gitlab.com/BuildStream/buildstream/merge_requests/909 to make sure I've not done something with my BST_GRANULAR_STAGE flag that's going to cause compatibility problems | 17:04 |
| juergbi | will take a look | 17:05 |
| gitlab-br-bot | martinblanchard approved MR !933 (danielsilverstone-ct/further-optimisations->master: WIP: Some more microoptimisations) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/933 | 17:10 |
| *** alatiera__ has joined #buildstream | 17:13 | |
| *** alatiera_ has quit IRC | 17:15 | |
| *** abderrahim3 has joined #buildstream | 17:27 | |
| *** abderrahim2 has quit IRC | 17:28 | |
| *** alatiera_ has joined #buildstream | 17:31 | |
| toscalix | laurence: resolved comments on the evaluation | 17:31 |
| toscalix | thanks for taking the time to review | 17:31 |
| gitlab-br-bot | jennis opened MR !939 (jennis/quota_declaration_fix->master: Ensure that specifying a percentage of disk space as a cache quota will start cleaning once we hit that percentage.) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/939 | 17:31 |
| *** alatiera__ has quit IRC | 17:33 | |
| toscalix | sadly I haven't found a bulk way to move the tickets from one milestone to another one. I believe github just implemented recently. It is an obvious feature. I am unsure it can be done through the API. I would expct so | 17:33 |
| toscalix | So the best way I can think of is to create a specific board for this and simply move the tickets from one column to another one, one by one. | 17:34 |
| toscalix | I tried it already in nosoftware and it works well | 17:35 |
| toscalix | gitlab boards to the rescue, once again | 17:35 |
| juergbi | toscalix: you can click on "Edit issues" in an issue list to change the milestone for multiple issues at once | 17:40 |
| benschubert | juergbi: I updated the PR : https://gitlab.com/BuildStream/buildstream/merge_requests/921 is that good for you now? | 17:40 |
| laurence | toscalix, np | 17:43 |
| toscalix | juergbi: I am sinking on a glass of water. Of course ! Thanks. I used before. Simply did not remember | 17:43 |
| laurence | juergbi, toscalix, I never knew that. Nice ! | 17:44 |
| toscalix | we filter the list first with bst 1.2 milestone and then we tell to update to 1.4 | 17:45 |
| toscalix | I will notify through the ML the change | 17:48 |
| toscalix | ups, 110 tickets | 17:52 |
| toscalix | will be updated | 17:53 |
| toscalix | ah, no, 98 the other are from other repos | 17:54 |
| *** lachlan has joined #buildstream | 17:54 | |
| toscalix | laurence: it does the bulk change.... per page so not all at once | 17:56 |
| toscalix | 98 tickets correspond to 5 pages... so I need to do it 5 times | 17:56 |
| toscalix | the conversion | 17:57 |
| *** lachlan has quit IRC | 17:58 | |
| *** lachlan has joined #buildstream | 17:59 | |
| juergbi | benschubert: mostly fine now, see comments | 17:59 |
| benschubert | thanks! | 17:59 |
| *** jonathanmaw has quit IRC | 18:00 | |
| gitlab-br-bot | jennis approved MR !933 (danielsilverstone-ct/further-optimisations->master: WIP: Some more microoptimisations) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/933 | 18:03 |
| *** lachlan has quit IRC | 18:03 | |
| *** alatiera_ has quit IRC | 18:15 | |
| *** lachlan has joined #buildstream | 18:23 | |
| benschubert | juergbi: would you mind having a last look? thanks a lot for your help! https://gitlab.com/BuildStream/buildstream/merge_requests/921/ | 18:24 |
| gitlab-br-bot | juergbi approved MR !921 (bschubert/tests-no-chroot-linux->master: test behavior when there is no sandbox on Linux) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/921 | 18:27 |
| laurence | toscalix, I actually did not receive any email about the milestone changes | 18:27 |
| laurence | which was cool | 18:28 |
| toscalix | I never did | 18:28 |
| laurence | so, applying labels and milestone changes doesn't create an email notification, but changing assignee does | 18:28 |
| toscalix | but hey.... just in case | 18:28 |
| benschubert | thanks! | 18:28 |
| laurence | yeah :) | 18:28 |
| juergbi | laurence, toscalix: my guess is that gitlab skips notifications for bulk changes. not sure, though | 18:29 |
| *** raoul has quit IRC | 18:29 | |
| juergbi | because I've seen milestone change emails before | 18:29 |
| toscalix | hummm... could be. I am not receiving notifications of my own changes | 18:29 |
| toscalix | juergbi: could it be that, together with the milestone, something else has changed? | 18:30 |
| toscalix | like labels? | 18:30 |
| toscalix | I tend to add the milestone together with labels | 18:30 |
| toscalix | but could be that the bulk change does not generate notifications | 18:31 |
| laurence | juergbi, do you subscribe to BuildGrid notifications? Because BuildGrid often changes milestones - wondering if you get notifications there? | 18:31 |
| toscalix | it would be weird though | 18:31 |
| laurence | as in, often moves tickets from one milestone to another | 18:31 |
| juergbi | it's possible that it was milestone change and something else | 18:32 |
| toscalix | laurence: I was wondering if you have tried to create team boards | 18:32 |
| toscalix | in buildstream we have status boards, severity boards but not user or team boards | 18:32 |
| valentind | toscalix, I will create a simple repository in buildstream to contain the ansible configuration for the sleds. We can always rename and put it into a subgroup later. But I would like to commit the files now. That I do not forget it. | 18:32 |
| toscalix | valentind: I would like to do it properly. I need some time to go over the configurations of the subgroup | 18:33 |
| toscalix | I cannot now | 18:33 |
| toscalix | I will remind you tomorrow, do not worry | 18:33 |
| valentind | toscalix, Sure. But you can just move the project I created. | 18:34 |
| toscalix | please add the repo here so I remember tomorrow: https://gitlab.com/BuildStream/nosoftware/alignment/issues/42 | 18:35 |
| toscalix | valentind: ^ | 18:35 |
| valentind | toscalix, commented. | 18:36 |
| toscalix | thanks | 18:37 |
| toscalix | it is a shame that freedesktop-sdk, buildstream and buildgrid do not share the same scales and label priorities so the ToDo and Issues view of each user's profile would show the list of tickets assigned to us by priority label | 18:39 |
| toscalix | anyway | 18:39 |
| *** xjuan has joined #buildstream | 18:40 | |
| *** toscalix has quit IRC | 18:55 | |
| *** lachlan has quit IRC | 18:57 | |
| *** lachlan has joined #buildstream | 19:22 | |
| *** lachlan has quit IRC | 19:36 | |
| *** lachlan has joined #buildstream | 19:39 | |
| *** xjuan has quit IRC | 20:15 | |
| *** toscalix has joined #buildstream | 20:24 | |
| *** lachlan has quit IRC | 20:45 | |
| *** benschubert has quit IRC | 21:18 | |
| *** xjuan has joined #buildstream | 21:33 | |
| *** juanalday has joined #buildstream | 21:40 | |
| *** bochecha has quit IRC | 22:01 | |
| *** bochecha has joined #buildstream | 22:01 | |
| *** juanalday has quit IRC | 22:24 | |
| *** juanalday has joined #buildstream | 22:25 | |
| *** JackWinter has joined #buildstream | 22:28 | |
| gitlab-br-bot | jjardon opened issue #755 (Some tests are failing on aarch64) on buildstream https://gitlab.com/BuildStream/buildstream/issues/755 | 22:33 |
| *** Juan_ has joined #buildstream | 22:52 | |
| *** juanalday has quit IRC | 22:54 | |
| *** Juan__ has joined #buildstream | 22:55 | |
| *** bochecha has quit IRC | 22:57 | |
| *** Juan_ has quit IRC | 22:57 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!