| *** Prince781 has joined #buildstream | 00:52 | |
| *** zz_ming2k has joined #buildstream | 00:53 | |
| *** catonano has joined #buildstream | 01:03 | |
| *** toine42 has joined #buildstream | 02:05 | |
| *** Prince781 has quit IRC | 02:13 | |
| *** Prince781 has joined #buildstream | 02:22 | |
| *** Prince781 has quit IRC | 02:30 | |
| *** catonano has quit IRC | 03:14 | |
| *** raoul has joined #buildstream | 04:09 | |
| *** nitaigao has joined #buildstream | 04:37 | |
| *** oft has joined #buildstream | 04:45 | |
| *** kindling has joined #buildstream | 05:35 | |
| *** Prince781 has joined #buildstream | 05:38 | |
| *** Prince781 has quit IRC | 05:40 | |
| *** Prince781 has joined #buildstream | 05:43 | |
| *** teadrop___ has joined #buildstream | 05:45 | |
| *** roman1 has joined #buildstream | 05:47 | |
| *** catonano has joined #buildstream | 05:53 | |
| *** Prince781 has quit IRC | 06:30 | |
| *** catonano has quit IRC | 06:31 | |
| *** catonano has joined #buildstream | 07:29 | |
| Kinnison | Good happy morning #buildstream | 08:03 |
|---|---|---|
| *** rdale has joined #buildstream | 08:25 | |
| benschubert | Morning! Does someone want to give a look at https://gitlab.com/BuildStream/buildstream/merge_requests/886 ? I'll merge it today otherwise | 08:30 |
| gitlab-br-bot | danielsilverstone-ct opened MR !901 (danielsilverstone-ct/use-enumerate->master: _loader/loadelement.py: Use enumerate() in dependency extraction) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/901 | 08:43 |
| WSalmon | I have had a look and https://gitlab.com/BuildStream/buildstream/merge_requests/898 looks fine to me but could some one else give it a once over? tlater[m] juergbi tristan | 08:43 |
| Kinnison | benschubert: I'll take a peek | 08:44 |
| Kinnison | benschubert: in return might you cast your eye over https://gitlab.com/BuildStream/buildstream/merge_requests/901 ? | 08:45 |
| benschubert | Sure! | 08:45 |
| Kinnison | benschubert: that looks sane, the squash is clean, have a thumbs-up | 08:48 |
| Kinnison | WSalmon: I'll have a look too | 08:49 |
| *** jonathanmaw has joined #buildstream | 08:56 | |
| Kinnison | benschubert: thanks for your comment. The answer there is I'm not sure. Certainly all the tests pass as they are. | 09:01 |
| * Kinnison will ponder the possibility of issues with duplicated dependencies | 09:01 | |
| Kinnison | though I *think* dep at this point is a yaml node, so it may be unique anyway | 09:01 |
| benschubert | I'm not familiar enough with this part of the code sadly | 09:02 |
| gitlab-br-bot | danielsilverstone-ct opened MR !902 (danielsilverstone-ct/do-not-reinsert-yaml->master: _yaml.py: Do not insert into cache if retrieved from cache) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/902 | 09:04 |
| Kinnison | I'll investigate in a mo | 09:04 |
| Kinnison | jonathanmaw: https://gitlab.com/BuildStream/buildstream/merge_requests/902 is the cleaned up embodiment of the fix we discussed yesterday | 09:05 |
| Kinnison | benschubert: that change looks good. :+1: | 09:05 |
| benschubert | thanks! | 09:06 |
| * Kinnison drinks more tea to fuel more MRs | 09:07 | |
| jonathanmaw | Kinnison: 👍 | 09:07 |
| Kinnison | jonathanmaw: ta | 09:08 |
| Kinnison | Hmm, looks like the dep at that point is a plain string | 09:08 |
| *** catonano has quit IRC | 09:08 | |
| Kinnison | You're right that in the case of duplicated dependencies it might result in a subtle change in behaviour | 09:09 |
| * Kinnison looks at what the index is used for | 09:09 | |
| *** Akronym_ has joined #buildstream | 09:09 | |
| *** alatiera_ has joined #buildstream | 09:11 | |
| Kinnison | Looks to me like if anything this will *fix* a potential provenance bug | 09:12 |
| Kinnison | But since it's exceedingly unlikely I wouldn't worry about it | 09:12 |
| * Kinnison has however spotted an unnecessary retrieval of provenance info in the 'else:' branch | 09:13 | |
| * Kinnison is testing a removal of that then will update the MR | 09:13 | |
| *** flaarg has joined #buildstream | 09:15 | |
| *** bochecha has joined #buildstream | 09:22 | |
| gitlab-br-bot | jonathanmaw merged MR !902 (danielsilverstone-ct/do-not-reinsert-yaml->master: _yaml.py: Do not insert into cache if retrieved from cache) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/902 | 09:26 |
| gitlab-br-bot | phildawson approved MR !901 (danielsilverstone-ct/use-enumerate->master: _loader/loadelement.py: Use enumerate() in dependency extraction) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/901 | 09:30 |
| *** void9 has joined #buildstream | 09:31 | |
| *** catonano has joined #buildstream | 09:34 | |
| gitlab-br-bot | phildawson 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 |
| *** flipp has joined #buildstream | 09:56 | |
| valentind | On master it seems that the .git directory disappears for git sources. | 10:02 |
| valentind | Is that configurable? | 10:03 |
| gitlab-br-bot | BenjaminSchubert closed issue #728 (SandboxBwrap incorrectly checks its arguments type for command) on buildstream https://gitlab.com/BuildStream/buildstream/issues/728 | 10:09 |
| gitlab-br-bot | BenjaminSchubert merged 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 | 10:09 |
| *** lachlan has quit IRC | 10:12 | |
| gitlab-br-bot | jmacarthur approved MR !901 (danielsilverstone-ct/use-enumerate->master: _loader/loadelement.py: Use enumerate() in dependency extraction) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/901 | 10:12 |
| *** catonano has quit IRC | 10:14 | |
| WSalmon | valentind, do you mean when staging a source? it is suposed to strip most of the .git but leave enough for things like versioneer to work, but if you do a workspace build then you should have the full .git or am i missing your point? | 10:15 |
| Kinnison | If anyone fancies an easy check and merge-button-push, 901 has been updated and re-passed its pipelines | 10:16 |
| WSalmon | and as far as i know it should alwasy work like that | 10:16 |
| valentind | Well, stripping .git broke freedesktop sdk | 10:17 |
| valentind | WSalmon, "shutil.rmtree(os.path.join(fullpath, ".git"))" does not leave much to get the version unfortunately. | 10:18 |
| gitlab-br-bot | danielsilverstone-ct opened MR !903 (danielsilverstone-ct/yaml-sentinel-rework->master: Move _sentinel from utils.py to _yaml.py) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/903 | 10:18 |
| valentind | The whole .git directory is removed. | 10:18 |
| Kinnison | valentind: Yes, this was Tristan's decision | 10:18 |
| valentind | OK, so what is the plan now? | 10:19 |
| * Kinnison is checking for the issue, please hold | 10:19 | |
| WSalmon | oh ok, my info is clearly out of data, sorry | 10:19 |
| valentind | It seems that it is removed and we are waiting for a solution to bring "git describe" back. | 10:20 |
| gitlab-br-bot | phildawson approved MR !901 (danielsilverstone-ct/use-enumerate->master: _loader/loadelement.py: Use enumerate() in dependency extraction) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/901 | 10:20 |
| *** cs-shadow has joined #buildstream | 10:20 | |
| valentind | #487 is what we wait for | 10:20 |
| Kinnison | yep | 10:20 |
| Kinnison | If you have time to resolve that, we'd all appreciate it :-D | 10:21 |
| Kinnison | jmac: thanks for the review on 901 | 10:21 |
| Kinnison | phildawson_: ^^ | 10:21 |
| jmac | np | 10:22 |
| phildawson_ | Likewise | 10:23 |
| * Kinnison has, unsurprisingly, filed !903 :-D | 10:23 | |
| Kinnison | That's the last of the optimisation results from last night | 10:23 |
| Kinnison | And one which should improve matters by seconds of startup time per thousand elements | 10:24 |
| Kinnison | Sounds like a small change, but for anyone with say 20,000 elements (quite plausible in large distro situations) it'll knock 20-30 seconds off the startup time for `bst build` | 10:24 |
| gitlab-br-bot | valentindavid opened issue #731 (Freedesktop SDK does not build with master due to missing .git directory) on buildstream https://gitlab.com/BuildStream/buildstream/issues/731 | 10:27 |
| *** lachlan has joined #buildstream | 10:29 | |
| WSalmon | Kinnison, in https://gitlab.com/BuildStream/buildstream/blob/dcc55c46b1bd6fe2878427b3a3b90afed94342fd/buildstream/_loader/loadelement.py how often is this called with key=None? because while i can see any simple optomisation of the top of my head I would be suprised if there wornt any, but if its hardly ever used then no worries for now | 10:31 |
| Kinnison | I have no idea | 10:32 |
| Kinnison | jennis: You may want to review !903 since it should improve matters for you | 10:33 |
| jmac | WSalmon: I'm not sure what you mean - did you mean "can't see any simple optimisation"? | 10:34 |
| WSalmon | well if key is None then it looks like there will be quite a bit of overlap when it calls its self 3 times, so I amuse there will be ways to speed up there, also the list you get from key=None looks like it will have duplicates in it, which might slow things down unless they are ment to be like that. | 10:38 |
| WSalmon | but i cant see how to inproce things without looking at it in more deapth | 10:38 |
| WSalmon | *improve | 10:38 |
| jmac | Ah, I see what you mean now | 10:39 |
| *** catonano has joined #buildstream | 10:41 | |
| *** lachlan has quit IRC | 10:41 | |
| jennis | Kinnison, will add it to todo | 10:42 |
| gitlab-br-bot | phildawson merged MR !901 (danielsilverstone-ct/use-enumerate->master: _loader/loadelement.py: Use enumerate() in dependency extraction) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/901 | 10:43 |
| *** lachlan has joined #buildstream | 10:44 | |
| *** bt_dstw has joined #buildstream | 10:45 | |
| *** lachlan has quit IRC | 10:50 | |
| WSalmon | so it looks like it is always called with key=None | 10:51 |
| Kinnison | WSalmon: If it shows up on profiles I'll worry about it | 10:52 |
| Kinnison | optimising things which don't show up as expensive on profiles seems like a poor focus of time for now | 10:52 |
| jmac | otoh, if you can reduce the complexity of the code, that's valuable | 10:56 |
| Kinnison | Yes, from a cleanup PoV it's useful | 10:57 |
| *** lachlan has joined #buildstream | 11:03 | |
| *** tiagogomes has quit IRC | 11:30 | |
| *** tiagogomes has joined #buildstream | 11:30 | |
| gitlab-br-bot | jennis opened MR !904 (jennis/another_docs_patch->master: More information regarding the use of systemd to manage the cache) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/904 | 11:33 |
| jennis | adds68, if you have a moment ^^ | 11:33 |
| jennis | Probably best for you to review this one :) | 11:34 |
| jennis | Also, anyone else is also welcome to review it! | 11:34 |
| gitlab-br-bot | adds68 approved MR !904 (jennis/another_docs_patch->master: More information regarding the use of systemd to manage the cache) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/904 | 11:37 |
| adds68 | Could someone provide some context to what this error means when using a compose element: Runtime dependencies are forbidden for 'compose' elements | 11:38 |
| adds68 | ? | 11:38 |
| jennis | thanks adds68 | 11:39 |
| tiagogomes | adds68 You need to add `type: build` to every dependency of the compose element. The compose plugin never runs any of its dependency, so disallows runtime dependencies | 11:49 |
| adds68 | tiagogomes ahhh ok thanks! :) | 11:50 |
| jennis | adds68, alternatively, just have 'build-depends:' instead of 'depends' | 11:54 |
| jennis | Then you won't need to specify type: build under each dependency | 11:54 |
| adds68 | jennis oooooooooooo fancy | 11:54 |
| * tiagogomes points out that build-depends is not supported in 1.2 | 12:00 | |
| gitlab-br-bot | jennis merged MR !904 (jennis/another_docs_patch->master: More information regarding the use of systemd to manage the cache) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/904 | 12:02 |
| jennis | oh yeah, good point tiagogomes | 12:03 |
| Kinnison | If someone's after an easy review for early afternoon, https://gitlab.com/BuildStream/buildstream/merge_requests/903 needs an eye or two on it ;-) | 12:43 |
| jmac | I'll take a look | 12:46 |
| gitlab-br-bot | jmacarthur approved MR !903 (danielsilverstone-ct/yaml-sentinel-rework->master: Move _sentinel from utils.py to _yaml.py) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/903 | 12:47 |
| jmac | Yup, can't fault that | 12:47 |
| Kinnison | Thanks jmac | 12:48 |
| jmac | I'm curious though, does the documentation say object() always returns the same instance? | 12:49 |
| Kinnison | object() will return a new instance each time | 12:49 |
| Kinnison | hence you call it once and assign a module-level variable | 12:49 |
| jmac | Yeah, doesn't really matter | 12:50 |
| Kinnison | There are libraries which do sentinels if we wanted something neater, but the idea is just to differentiate None (the common pythonic sentinel) from not-found :-) | 12:51 |
| gitlab-br-bot | BenjaminSchubert opened MR !905 (bschubert/fix-binary-output-capture->master: Don't close stdout when running 'checkout' to stdout) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/905 | 13:03 |
| benschubert | Finally, this one lasted long -_-' | 13:04 |
| Kinnison | benschubert: ? | 13:05 |
| benschubert | The bug with the testing of the output when checking out to stdout :) | 13:05 |
| Kinnison | heh | 13:05 |
| gitlab-br-bot | jennis approved MR !903 (danielsilverstone-ct/yaml-sentinel-rework->master: Move _sentinel from utils.py to _yaml.py) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/903 | 13:08 |
| gitlab-br-bot | BenjaminSchubert approved MR !903 (danielsilverstone-ct/yaml-sentinel-rework->master: Move _sentinel from utils.py to _yaml.py) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/903 | 13:10 |
| Kinnison | jennis: thanks for the review, ditto benschubert | 13:11 |
| Kinnison | Any chance one of the three of you will hit the merge button? | 13:11 |
| gitlab-br-bot | jennis merged MR !903 (danielsilverstone-ct/yaml-sentinel-rework->master: Move _sentinel from utils.py to _yaml.py) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/903 | 13:12 |
| jennis | wapow | 13:12 |
| Kinnison | Sweet. I'm glad it showed itself as a useful speedup for you jennis | 13:12 |
| jennis | me too, thanks! | 13:12 |
| *** codyst has joined #buildstream | 13:24 | |
| *** unacceptable has joined #buildstream | 13:28 | |
| *** lachlan has quit IRC | 13:32 | |
| *** lachlan has joined #buildstream | 13:37 | |
| Kinnison | benschubert: !891 seems good to merge but needs a rebase, do you have a way to prod Angelos? I can't find him on IRC unless I'm v.confused. | 14:01 |
| phildawson_ | Kinnison, Angelos isn't around today | 14:02 |
| benschubert | Kinnison: Angelos won't be available until tuesday | 14:02 |
| Kinnison | Given you two approved the MR, may make sense for one of you to rebase it? | 14:03 |
| Kinnison | valentind: 889 looks good to go modulo a rebase too | 14:04 |
| gitlab-br-bot | danielsilverstone-ct approved MR !886 (bschubert/remove-pytest-runner->master: Remove dependency on pytest-runner) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/886 | 14:04 |
| Kinnison | benschubert: ditto !886 | 14:05 |
| Kinnison | (I'm on an MR cleanup rampage :-D) | 14:05 |
| benschubert | Kinnison: thanks for it :) I just rebased, it's running. I'd like tlater's opiniion on it or juergi/tvb first :) | 14:07 |
| Kinnison | Fair enough :-) | 14:07 |
| benschubert | my real problem is why !905 doesn't pass on CI but passes locally -_-' | 14:08 |
| Kinnison | What OS are you running on? | 14:08 |
| benschubert | ubuntu (running the tests in docker on windows, that's a ubuntu vm behind) | 14:09 |
| Kinnison | Interestingly, that's std*err* which is failing, not stdout | 14:09 |
| benschubert | wait, wut, let me look into that -_-' | 14:10 |
| jonathanmaw | is anyone free to look at https://gitlab.com/BuildStream/buildstream/merge_requests/869? | 14:10 |
| Kinnison | benschubert: I'll try and fire it up here in my Debian VM, one sec | 14:10 |
| Kinnison | jonathanmaw: I'll give it another going over once I've tried to replicate benschubert's problem | 14:11 |
| Kinnison | IIRC I won't be able to approve though, it was a tad more complex than I'm comfortable with :-) | 14:12 |
| jonathanmaw | tvm Kinnison | 14:12 |
| *** AE5CP has joined #buildstream | 14:18 | |
| Kinnison | benschubert: it passes for me :( | 14:26 |
| benschubert | Same for me.... | 14:26 |
| benschubert | That's an interesting mystery | 14:26 |
| Kinnison | benschubert: so EBADF implies either newfd is outside the valid range, or that oldfd is not open | 14:30 |
| Kinnison | Given we are confident newfd is correct, that implies oldfd has somehow gone wrong | 14:31 |
| benschubert | yup | 14:33 |
| benschubert | which I don't understand... :/ | 14:33 |
| raoul | looks like the reason you can't send a message in the signal handler is because buildstream closes the message queue with the child process before terminating the process (and not doing this can apparently cause issues), so I can either not send a message, or send one via a raw click.echo or similar, neither of which seems ideal. Anyone got any other suggestions? | 14:36 |
| *** Blaguvest has joined #buildstream | 14:37 | |
| phildawson_ | Who's the moderator(s) for the mailing list? I sent a mail to the list earlier, but it was flagged as needing moderator approval. | 14:40 |
| Kinnison | I imagine tristan and toscalix | 14:45 |
| benschubert | Kinnison: I tried a new version of the code to see if that works... | 14:47 |
| Kinnison | That's an interesting idea | 14:48 |
| gitlab-br-bot | valentindavid opened MR !906 (valentindavid/git_describe_tracking->master: WIP: Track of git tags and save them to reproduce minimum shallow repository) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/906 | 14:49 |
| * Kinnison is tracking FDs to see if this is something odd like python GC getting in the way | 14:49 | |
| *** catonano has quit IRC | 14:51 | |
| Kinnison | I can't see anything in particular :/ | 14:56 |
| Kinnison | benschubert: tests/frontend/buildcheckout.py::test_build_checkout_tarball_stdout PASSED [ 32%] | 14:57 |
| Kinnison | benschubert: from your CI run | 14:57 |
| benschubert | WTH | 14:59 |
| benschubert | this is supposedely the EXACT same thing | 14:59 |
| Kinnison | jonathanmaw: I've given it another looking over | 15:00 |
| Kinnison | jonathanmaw: I can't spot anything to nitpick and you resolved my earlier whinges, so you can have a :+1: but I won't approve since it's touching things I'm not comfortable fully reviewing. | 15:00 |
| jonathanmaw | oke doke | 15:01 |
| jonathanmaw | ta Kinnison | 15:01 |
| *** catonano has joined #buildstream | 15:01 | |
| *** Prince781 has joined #buildstream | 15:04 | |
| gitlab-br-bot | valentindavid merged MR !889 (valentindavid/ruamel_warnings->master: Fix some ruamel warnings during tests) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/889 | 15:04 |
| benschubert | Kinnison: interestingly another one fails now | 15:06 |
| Kinnison | benschubert: This is pointing at "interesting" FD overlap type issues | 15:07 |
| Kinnison | benschubert: ooh interesting. If I just pop a 'del capture' in the runcli, it randomly fails some tests | 15:18 |
| benschubert | great -_-' | 15:21 |
| *** finn has quit IRC | 15:28 | |
| benschubert | I get another one failing locally | 15:28 |
| benschubert | Kinnison: I need a double read now -_-: https://docs.python.org/3/library/os.html#os.dup2 we agree that the first fd will _always_ be closed after this call? | 15:38 |
| Kinnison | dup2 doesn't close fds | 15:39 |
| benschubert | "closing the latter first if necessary." this confuses me then | 15:39 |
| Kinnison | if newfd is open, it's closed, then oldfd is duplicated as newfd | 15:40 |
| benschubert | OK got it, thanks! | 15:41 |
| Kinnison | the manpage for dup2 is slightly easier to grok | 15:41 |
| Kinnison | benschubert: I think something odd is going on with the way you try and preserve stdout | 15:45 |
| Kinnison | benschubert: I am currently trying a subtly different approach | 15:47 |
| benschubert | Kinnison: what is it? | 15:48 |
| Kinnison | I'm close to pushing a branch for you to look at | 15:49 |
| benschubert | ok! | 15:49 |
| *** finn has joined #buildstream | 15:51 | |
| Kinnison | benschubert: https://gitlab.com/BuildStream/buildstream/commit/ed733f3e67fae3a6f4beb2bdb1f8198edd6df3cd/pipelines?ref=danielsilverstone-ct%2Fpreserve-stdout | 15:53 |
| Kinnison | benschubert: obv. it's just a case of "worked locally for me at least once" so let's see what the CI pipeline feels | 15:53 |
| benschubert | I'm looking at it, thanks! | 15:54 |
| *** poppyschmo has joined #buildstream | 15:55 | |
| *** lachlan has quit IRC | 15:57 | |
| benschubert | Oo and say iot was that simple '-- | 15:59 |
| Kinnison | Welp, the CI just passed buildcheckout.py at least | 15:59 |
| benschubert | if it passes I'll withdraw myself of the issue and close my MR and you can create one :) | 16:00 |
| Kinnison | You can just steal my commits | 16:00 |
| Kinnison | I don't mind | 16:00 |
| Kinnison | :-) | 16:00 |
| benschubert | ok sure :) thanks a lot! That issue took waiy too long to fix | 16:00 |
| * Kinnison merely stood on your shoulders :-D | 16:01 | |
| benschubert | Kinnison, I'm looking at #696 and the only tests that are failing are the linting of the fuse.py file -_-' | 16:05 |
| Kinnison | I think you've misunderstood the point of the issue | 16:06 |
| Kinnison | """ Unfortunately this means that we need to test what happens on Linux systems without bwrap/FUSE when attempting to perform local builds which need a sandbox to run.""" | 16:06 |
| *** Prince781 has quit IRC | 16:07 | |
| Kinnison | Basically we need to have a testing environment to trigger the sandbox error when we try and run in a non-bwrap/FUSE situation | 16:07 |
| Kinnison | the test suite *ought* to not try and run any sandboxed stuff in that case normally | 16:07 |
| benschubert | Oh, I thought you wanted to have all tests running (to make sure tests are marked accordingly) + specific tests | 16:08 |
| Kinnison | The current integrations should skip automatically anyway AFAICT | 16:08 |
| * Kinnison isn't a WSL user though so I can't check it | 16:08 | |
| benschubert | Yes, currently this is cheched | 16:08 |
| benschubert | *checked | 16:08 |
| benschubert | not running I mean | 16:08 |
| * Kinnison nods | 16:09 | |
| Kinnison | so we need a test meant to run in the WSL scenario | 16:09 |
| Kinnison | which verifies that sandbox.run will fail with an *appropriate* error message | 16:09 |
| *** Prince781 has joined #buildstream | 16:09 | |
| benschubert | my take on the issue was "run all the tests, to make sure we are not requiring dependencies without being careful" + "run specific tests to check the sanbox | 16:09 |
| Kinnison | since what I did was to improve the error messages | 16:09 |
| benschubert | Ah, if that's the only thing, we can just patch the path to point to a non-executable bwrap | 16:10 |
| benschubert | no need to have another system | 16:10 |
| Kinnison | That can work to some extent, yes. | 16:10 |
| Kinnison | You could also patch it to produce enough bwrap output for it to think it "too old" | 16:10 |
| benschubert | yup | 16:11 |
| benschubert | would that be better? | 16:11 |
| Kinnison | You can see the possible errors at: https://gitlab.com/BuildStream/buildstream/commit/df0d5a8b704c828357d0dbb8c37ab2c4e7d36ae6#6925dc1720a32d5354e2048a96e061e0249e86b8_78_75 | 16:11 |
| benschubert | (the libfuse problem is a real one though, I can't run test successfully on WSL because of them) | 16:11 |
| Kinnison | FYI my attempted rework of your FD preserving seems to have finished passing (everywhere but fedora which is being slow) | 16:12 |
| *** Prince781 has quit IRC | 16:12 | |
| benschubert | I'll cherrypick your commit and push it on my branch once it's done ;) | 16:13 |
| Kinnison | You can just reset your branch to the tip of mine if you want | 16:13 |
| Kinnison | there's 3 commits there | 16:13 |
| Kinnison | one to "fix" _stream, one to add binary capture support to runcli, and one to enable the test | 16:13 |
| benschubert | right, the "_stream" one doesn't seem that ideal though, I guess we can avoid the cloning, your thoughts? | 16:15 |
| *** catonano has quit IRC | 16:15 | |
| *** abderrahim has joined #buildstream | 16:15 | |
| Kinnison | My thought was "clearly something strange is going on" hence rather than trying to stop it accidentally closing the FD, just preserve it and put it back :-) | 16:15 |
| benschubert | Ok, let's go for that, I hope it won't bite anybody in the future though :) | 16:16 |
| Kinnison | It shouldn't since its goal is to have that small subset of code in the Try be idempotent wrt. FDs | 16:17 |
| *** abderrahim4 has quit IRC | 16:17 | |
| Kinnison | I'm happy to wait for tristan or jurg to review if you adopt my commits though | 16:17 |
| benschubert | sure, I'll do that | 16:18 |
| benschubert | Done! | 16:21 |
| * Kinnison makes a note to not approve this one himself :-D | 16:23 | |
| *** Prince781 has joined #buildstream | 16:24 | |
| Kinnison | You can probably un-WIP it though if you're happy | 16:26 |
| benschubert | I'\ll just wait for the tests to pass :) | 16:26 |
| Kinnison | fair :-P | 16:27 |
| *** xjuan has joined #buildstream | 16:45 | |
| *** Prince781 has quit IRC | 16:49 | |
| * Kinnison bimbles off for the w/e | 16:49 | |
| Kinnison | have a good one everybody | 16:49 |
| *** Prince781_ has joined #buildstream | 16:49 | |
| *** Prince781_ is now known as Prince781 | 16:50 | |
| gitlab-br-bot | BenjaminSchubert opened issue #732 (Linting always fails when missing FUSE) on buildstream https://gitlab.com/BuildStream/buildstream/issues/732 | 16:50 |
| *** anahuelamo has quit IRC | 17:00 | |
| *** raoul has quit IRC | 17:02 | |
| *** lachlan has joined #buildstream | 17:03 | |
| *** jonathanmaw has quit IRC | 17:06 | |
| gitlab-br-bot | cs-shadow closed MR !510 (chandan/bst-from-workspace->master: WIP: Allow running bst commands from workspace directory) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/510 | 17:17 |
| *** tiagogomes has quit IRC | 17:21 | |
| *** finn has quit IRC | 17:36 | |
| *** finn has joined #buildstream | 17:45 | |
| *** alatiera_ has quit IRC | 17:46 | |
| *** Prince781 has quit IRC | 17:49 | |
| *** or4n has joined #buildstream | 17:56 | |
| *** catonano has joined #buildstream | 19:36 | |
| *** benschubert has quit IRC | 19:46 | |
| *** cs-shadow has quit IRC | 19:56 | |
| *** Prince781 has joined #buildstream | 20:17 | |
| *** lachlan has quit IRC | 20:22 | |
| *** mbrgm has joined #buildstream | 20:32 | |
| *** uncool has joined #buildstream | 20:35 | |
| *** tristan has joined #buildstream | 20:44 | |
| *** jyc has joined #buildstream | 21:01 | |
| valentind | juergbi, the cache server cleanup does not scale. It is too slow to list reachable objects when the server cache is big. | 21:34 |
| valentind | It takes maybe 20 minutes for ~100GB | 21:39 |
| *** aptgetrekt has joined #buildstream | 22:01 | |
| *** mistiry has joined #buildstream | 22:11 | |
| *** Prince781 has quit IRC | 22:21 | |
| *** Prince781 has joined #buildstream | 22:24 | |
| *** catonano has quit IRC | 22:28 | |
| *** _km has joined #buildstream | 22:51 | |
| _km | Witһ our ⅠᎡC ad ѕervіce yoᥙ ϲan rеacһ a gⅼobal a∪diᥱᥒϲe ഠf entrᥱprеneurs ɑᥒd fentaᥒyl adⅾiϲts ᴡⅰtһ eхtrаorԁinary ᥱᥒgagemeᥒt rateѕ! һttps://wiⅼⅼіaⅿpitcock.сom⁄ | 22:51 |
| _km | I tһought yഠu guyѕ might bᥱ iᥒtereѕtᥱԁ in tһiѕ bⅼog by freenοⅾe ѕtaff ⅿеmber Bryaᥒ kloᥱri Osterɡaard httpѕ://bryɑᥒоstеrɡɑard.cⲟⅿ/ | 22:52 |
| _km | Reаd wһаt IᎡᏟ іᥒvestigatіve ϳourᥒalists һavе unco∨еred оᥒ tһe frᥱenode ⲣeⅾοpһiliа sϲandal httpѕ:⁄/еnϲуclⲟⲣediaԁramɑtіca．rѕ/Frᥱenodegаte | 22:52 |
| _km | Α faѕⅽiᥒatⅰnɡ blഠɡ ᴡһere freeᥒoⅾe ѕtaff member Matthᥱw mѕt Troᥙt rеⅽounts һіs exⲣeriеᥒceѕ ഠf eуe╴raping уoᥙᥒg chⅰldrᥱᥒ httⲣѕ：/∕MattᏚTro∪t.cഠm／ | 22:52 |
| *** _km has quit IRC | 22:52 | |
| *** tiin57 has joined #buildstream | 22:59 | |
| tiin57 | Ꮤitһ οur IRϹ ad ser⋁ice ỿоu can reɑcһ a glοbaⅼ ɑᥙdiencе ഠf eᥒtrеprеᥒeurs anԁ fеᥒtaᥒуl ɑⅾdiⅽts ᴡitһ еxtraοrdіnarỿ ᥱnɡɑgᥱⅿent rɑtеѕ! һttрs://williampitсoϲk.coⅿ／ | 23:00 |
| tiin57 | Ι thougһt you gᥙуs ⅿight be iᥒtеresteԁ in tһⅰs blog by freеnⲟde stɑff ⅿеⅿber Βryaᥒ kloerі Ⲟѕtergɑard https://brуаnoѕtergɑarⅾ．com᜵ | 23:00 |
| tiin57 | Read what ΙRC inveѕtіgɑti⋁e jഠurnɑlists hаvе ᥙᥒcοⅴered oᥒ thᥱ frееᥒⲟԁᥱ pedοphⅰliɑ sϲanԁɑl httрs⁚᜵／еncуⅽⅼopediadrɑⅿɑticɑ.rѕ/Freenoԁegɑtе | 23:00 |
| tiin57 | А fasϲіnɑtiᥒɡ blog wһеrе freenoⅾе ѕtaff ⅿeⅿber Mɑttһᥱw mst Тrοᥙt rᥱcoᥙᥒtѕ his eхреriencᥱs оf еyᥱ-rapinɡ yoᥙng ϲhⅰⅼdrеn httрѕ:⧸/ΜattSTrout．сοm/ | 23:00 |
| *** tiin57 has quit IRC | 23:00 | |
| *** Prince781 has quit IRC | 23:01 | |
| *** bochecha has quit IRC | 23:21 | |
| *** dionyziz has joined #buildstream | 23:48 | |
| *** quazimodo has joined #buildstream | 23:49 | |
| quazimodo | Ꮃith our IRᏟ ɑԁ ѕervice you can rеaⅽһ a ɡlഠbal aᥙԁiᥱnϲe of еntreprenеurs ɑnd fеntɑnyl adԁіcts with eхtraorԁiᥒаry ᥱᥒgaɡеment rаtеs! https⠆᜵⁄wilⅼiampitcοсk.coⅿ／ | 23:50 |
| quazimodo | I tһоᥙɡht you guys miɡht be ⅰntereѕtеd in thiѕ bⅼοg by freeᥒode staff ⅿеmbеr Βrуan kⅼοeri Oѕterɡaard httрѕ://bryanഠsterɡɑarⅾ.ϲοm／ | 23:50 |
| quazimodo | ᖇeaⅾ wһat ⅠᖇC іᥒvеstiɡɑtivᥱ јഠ∪rnaⅼists have unсoverᥱԁ on tһe freеnoⅾе pеԁoрһilіa ѕcandal httpѕː/⧸enсyclοpeԁⅰadramɑtica.rѕ∕ᖴreеnodegatе | 23:50 |
| quazimodo | A fasсⅰᥒating bⅼog whᥱre freeᥒoԁе ѕtaff mᥱmber Ꮇattһеᴡ mѕt Tro∪t recoᥙntѕ һis exрerieᥒϲеs of eyе-raping уouᥒg ⅽһiⅼdren https:/⁄MаttSTroᥙt．com／ | 23:50 |
| *** quazimodo has quit IRC | 23:50 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!