| *** tristan has joined #buildstream | 04:03 | |
| *** dominic has joined #buildstream | 06:54 | |
| juergbi | tristan: gitlab is acting up again, so commenting here. mcatanzaro mentioned he is using 1.1.2, so I assume it can be closed | 07:06 |
|---|---|---|
| tristan | juergbi, aha... but we have bigger problems... | 07:43 |
| tristan | juergbi, where did mcatanzaro say that ? | 07:44 |
| tristan | here in channel maybe ? | 07:44 |
| tristan | I am unhappy with the state of things today, anyway :'( | 07:45 |
| tristan | It all started, when I tried to reproduce his problem... went ahead with the most straight forward use case | 07:45 |
| tristan | open a workspace for glib-networking.bst... build glib-networking.bst | 07:46 |
| juergbi | tristan: it's in the issue description | 07:46 |
| juergbi | > bst 1.1.2 crashes | 07:46 |
| tristan | Go ahead and modify a file in glib-networking checkout | 07:46 |
| *** toscalix has joined #buildstream | 07:46 | |
| tristan | And build the workspace again, let's see if there's going to be a crash right ? | 07:46 |
| tristan | But worse than that | 07:46 |
| tristan | It just happily reports "cached" | 07:46 |
| tristan | So, surely this most basic case must be covered right ? no... checked tests/frontend/workspace.py... | 07:47 |
| juergbi | oh :/ | 07:47 |
| tristan | it only has one test, which adds/removes files | 07:47 |
| juergbi | I thought we did have coverage for such basic workspace build tests | 07:47 |
| tristan | I added a new test now, which has 3 different paths; one add, one remove, one modify | 07:48 |
| tristan | modify fails | 07:48 |
| tristan | So then... I thought lets fix it... | 07:48 |
| *** toscalix has quit IRC | 07:49 | |
| tristan | Workspaces are element wide now... but... we have tons of spaghetti bloat which delegates calls from Element -> Source, which swings back to the Element-wide workspace | 07:49 |
| *** toscalix has joined #buildstream | 07:49 | |
| tristan | So I guess, when calculating a cache key for a workspaced element with more than one source, we just blindly checksum the whole directory recursively... *twice* | 07:49 |
| tristan | juergbi, I'm in the middle of factoring out all the junk from Source | 07:49 |
| tristan | that should clear the muddy waters a bit, and help me get to the bottom of solving the failing test case | 07:50 |
| tristan | So far I now have Element._update_source_state() which does the additional workspace stuff, and I've removed the __track state from Source to replace with more consistent __tracking_scheduled / __tracking_done in Element | 07:51 |
| tristan | At the end of it, there will no longer be any __workspace pointer in Source | 07:51 |
| * tristan is doing operation by operation in separate commits | 07:51 | |
| tristan | one commit is: Removing pointless Element._workspaced(), since we have Element._get_workspace() | 07:52 |
| tristan | Can you believe that was somehow there ? Yay lets ask every source if we gave it a workspace in a loop, to find out if the element is "workspaced" ! | 07:53 |
| tristan | Man, what a morning | 07:53 |
| juergbi | well, leftovers from per-source workspaces are not too surprising | 07:53 |
| juergbi | I'm more surprised that we didn't have a test case for this | 07:54 |
| tristan | I am disappointed in both | 07:54 |
| tristan | The leftovers left behind when workspaces were made to be per-element, should really never have happened, and it appears that the situation has just worsened since then | 07:56 |
| tristan | instead of setting things straight, people "follow whats there" and add growth to existing warts | 07:56 |
| tristan | it's not the right way :-S | 07:56 |
| juergbi | it would obviously be better if such things never happened, however, the larger a project the more likely such things happen. we just have to make sure to fix it as soon as we notice it | 08:00 |
| gitlab-br-bot | buildstream: issue #347 ("AssertionError when caching artifact") changed state ("closed") https://gitlab.com/BuildStream/buildstream/issues/347 | 08:26 |
| *** noisecell has joined #buildstream | 08:36 | |
| tristan | Ummm, tlater... is there any reason that you decided that we dont use the mounting codepath for sources if they cannot be "built incrementally" ? | 08:40 |
| tristan | https://gitlab.com/BuildStream/buildstream/blob/master/buildstream/element.py#L1547 | 08:41 |
| tristan | I think that we still want to mount the workspace into the sandbox, even if the artifacts cannot be "diffed" right ? | 08:41 |
| tristan | Or we explicitly want to disallow the build from modifying the workspace here ? | 08:42 |
| tlater | tristan: I don't quite remember the context, I think this was part of that one bug where we had to ensure that stuff like `make` doesn't recompile everything? | 08:42 |
| tristan | 989bb3fe000dc8c843b7a8293f9cf0a3293b8d39 | 08:42 |
| tristan | element.py: Disallow incremental builds for caches that can't diff | 08:42 |
| tlater | Ah, right | 08:43 |
| tristan | I just dont understand exactly why | 08:43 |
| tristan | I guess I'll leave this part in place whilst removing the silly ridiculous loops around there | 08:43 |
| tlater | I thought mounting the workspace into a sandbox was equivalent to making it build incrementally? | 08:43 |
| tlater | I might have misunderstood that, though | 08:44 |
| tristan | Well, it's *part* of it | 08:44 |
| tristan | But what was the motivation to disable that when the artifact cache could not assist in the diffs, is rather my question | 08:44 |
| tlater | tristan: It was this bug: https://gitlab.com/BuildStream/buildstream/issues/215 | 08:45 |
| tlater | It's a bit hard to wrap my head around this again right now, for some reason | 08:46 |
| juergbi | tristan: if we can't track changes in dependencies, we can't allow incremental builds | 08:46 |
| juergbi | due to incorrect timestamp | 08:46 |
| tlater | juergbi: I'm trying to remember why the incorrect time stamp means that we *can't* allow incremental builds | 08:47 |
| tlater | What would be incorrect about the build? | 08:47 |
| juergbi | updated header file will still have the same timestamp | 08:47 |
| juergbi | so make won't do anything | 08:47 |
| tlater | Ah, right | 08:47 |
| tlater | And that would cause the updated dependency to not have an effect on the resulting artifact | 08:47 |
| tlater | Which is a big no-no because it would be a broken artifact | 08:47 |
| juergbi | correct | 08:47 |
| tristan | juergbi, Right so... whether we can do them incrementally or not, I dont think is related to whether we want a workspace to be *mounted* into the sandbox rather than recursively copied | 08:48 |
| tristan | right ? | 08:48 |
| tristan | that's rather my angle | 08:48 |
| tristan | it wont be incremental anyway | 08:48 |
| juergbi | tristan: if we mount it read-write, we'd have to be able to clean it | 08:48 |
| juergbi | we can't force non-incremental build | 08:48 |
| tristan | Hmmm | 08:48 |
| juergbi | as I've mentioned before, long term I want to keep clean workspace dir and build dir completely separate | 08:49 |
| tristan | Ok anyway, I'll treat this separately, although I believe I'm steering directly towards it | 08:49 |
| tristan | because of the bug, I will probably have to discover the mystery of how we determine the cache key of the workspace; which has probably changed | 08:50 |
| juergbi | that hasn't changed that recently, iirc | 08:50 |
| juergbi | we added the two stage approach when introducing incremental workspace builds but you're already aware of this | 08:50 |
| *** jonathanmaw has joined #buildstream | 08:51 | |
| tristan | Yes, but it's broken so something has changed :-/ | 08:52 |
| tristan | juergbi, I can only surmise it has to do with how we have recorded the latest successful build in the workspace metadata | 08:52 |
| juergbi | possibly, yes | 08:53 |
| juergbi | maybe git bisect would help together with the new test case? | 08:53 |
| tristan | but I have to finish cleaning the lens here to see it more clearly | 08:53 |
| tristan | it could, but I'm almost finished completing the work of "Making workspaces element wide instead of source specific" | 08:54 |
| tristan | So I'll finish that first and then probably debugging it will be easy | 08:54 |
| gitlab-br-bot | buildstream: merge request (135-expire-artifacts-in-local-cache->master: WIP: Resolve "Expire artifacts in local cache") #347 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/347 | 08:58 |
| gitlab-br-bot | buildstream: merge request (135-expire-artifacts-in-local-cache->master: WIP: Resolve "Expire artifacts in local cache") #347 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/347 | 09:11 |
| *** tiago has joined #buildstream | 09:15 | |
| gitlab-br-bot | buildstream: merge request (135-expire-artifacts-in-local-cache->master: WIP: Resolve "Expire artifacts in local cache") #347 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/347 | 09:18 |
| tristan | tlater, f761140f18a7d54caf6e6dba8a722b9ff1f4430e left Element._set_source_workspaces() deadcode behind, btw | 09:20 |
| tlater | That wasn't called anymore? Frankly, I just assumed that we needed to call into sources still for some reason. | 09:21 |
| tlater | That was probably the time to refactor this mess, I should have gone a bit deeper. | 09:21 |
| gitlab-br-bot | buildstream: merge request (jmac/performance-ci->master: Add benchmark to CI test and document CI.) #341 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/341 | 09:22 |
| tristan | tlater, it was a mess before you arrived, I mean; you changed it so that Workspace.init() calls source._set_workspace() directly, bypassing Element._set_source_workspaces(), without removing it :) | 09:23 |
| tlater | Ah o\. Yeah, that's a little stupid. | 09:23 |
| tristan | I will make that cache the workspace on Element, and make Element._get_workspace() use a cached value, instead of doing the lookup it currently still does | 09:24 |
| tristan | tlater, it's amazing how much of a mess this has become :-S | 09:24 |
| * tristan almost finished removing __workspace pointer from Source | 09:24 | |
| tristan | would be good if the linter could help us find private methods which are not reachable by public ones, at least | 09:30 |
| * tristan understands that it's difficult to find deadcode with a linter, but at least private methods should be doable | 09:31 | |
| tlater | tristan: We disabled a bunch of the private/public helpers because we follow a different-than-usual convention | 09:31 |
| tlater | Not sure it could find that kind of deadcode in python, though. | 09:32 |
| tristan | tlater, we are now *getting* close to the pep8 standards, I think the only exception really will be Sandbox() | 09:33 |
| tristan | well, even then | 09:33 |
| tristan | probably we implicitly follow the pep8 standard for underscores, after we finished the overhaul | 09:34 |
| tristan | Ummm | 09:36 |
| tlater | Oh, that would be neat. From my understanding the public/private API marking we do isn't as usual, but I haven't looked closely at what pep8 says about it. | 09:37 |
| tristan | So, can anyone tell me why we have this check in source.py: https://gitlab.com/BuildStream/buildstream/blob/master/buildstream/source.py#L615 ? | 09:37 |
| tristan | tlater, maybe you know about that "invalidate the current workspace key when setting a new workspace" thing ? | 09:37 |
| tristan | I am moving that to Element, and I am going to turn it into "assert self.__workspace is None" | 09:38 |
| tlater | Hm | 09:38 |
| tlater | It might be that this is "setting" it in memory, rather than creating it | 09:38 |
| tristan | it is just caching the workspace close to the source (not the element) | 09:39 |
| tristan | What I dont understand is, how can the same element ever have more than one workspace in a given session ? | 09:39 |
| tlater | I think there's the workspace reset command that kind of requires this sort of thing | 09:39 |
| tristan | Ummm | 09:39 |
| tristan | that will first close and then open the workspace, though. | 09:40 |
| tristan | I'll see what happens | 09:40 |
| tlater | Yeah, that will require changing what the source thinks is the current workspace | 09:40 |
| * tlater isn't sure he introduced workspace keys though, so not entirely certain this is required | 09:41 | |
| *** aday has joined #buildstream | 09:43 | |
| tristan | if the workspace reset test fails, then I can't put the assert and have to treat it differently | 09:44 |
| tristan | but I think in *any* case, it's unset, and set; separately; it is never *changed* | 09:45 |
| tlater | Certainly not the end of the world to poke it ;) | 09:45 |
| tristan | Now the question of cache keys, I guess I'll be ugly and make cache key calculation compatible | 09:45 |
| tristan | tlater, not the end of the world, but misleading to the reader | 09:46 |
| * tlater was referring to seeing if an assert breaks it | 09:47 | |
| tristan | it invites speculation that weird things might happen during the life cycle of an element, whereas an assert reads "This cannot happen" | 09:47 |
| tristan | Ah | 09:47 |
| gitlab-br-bot | buildstream: merge request (jmac/artifact_cache_error_message->master: Artifact cache: Mention the remote URL when we fail to fetch remote refs) #354 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/354 | 09:49 |
| tristan | hmmm, it does raise the assert; maybe I'll keep the hash lookup instead | 09:50 |
| gitlab-br-bot | buildstream: merge request (tristan/element-wide-workspaces->master: Untangle sphaghetti workspace element/source mess) #368 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/368 | 10:11 |
| gitlab-br-bot | buildstream: merge request (tristan/element-wide-workspaces->master: Untangle sphaghetti workspace element/source mess) #368 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/368 | 10:12 |
| tristan | juergbi, if you want to proof read the above ^^^^^ feel free | 10:14 |
| jmac | Could I have a review or merge of !341 & !354 please? | 10:20 |
| gitlab-br-bot | buildstream: merge request (jmac/artifact_cache_error_message->master: Artifact cache: Mention the remote URL when we fail to fetch remote refs) #354 changed state ("merged"): https://gitlab.com/BuildStream/buildstream/merge_requests/354 | 10:25 |
| gitlab-br-bot | buildstream: merge request (tristan/element-wide-workspaces->master: Untangle sphaghetti workspace element/source mess) #368 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/368 | 10:25 |
| *** dominic has quit IRC | 10:36 | |
| *** dominic has joined #buildstream | 10:36 | |
| gitlab-br-bot | buildstream: merge request (tristan/element-wide-workspaces->master: Untangle sphaghetti workspace element/source mess) #368 changed state ("merged"): https://gitlab.com/BuildStream/buildstream/merge_requests/368 | 10:41 |
| tristan | tlater, all; if you have any core Element/Source especially workspace related stuff... -> Rebase | 10:44 |
| tristan | juergbi, I did bisect after all | 11:06 |
| tristan | tlater, you broke it in: f761140f | 11:06 |
| tristan | lets see how it broke... | 11:06 |
| juergbi | oh, that was supposed to be a no-behavior-change refactoring | 11:07 |
| tristan | fwiw, it's this: https://bpaste.net/show/be95b73bb664 | 11:07 |
| juergbi | iirc, I even diff'd the moved code to verify but apparently not sufficient | 11:07 |
| tristan | juergbi, the above is a no-behavior-change | 11:08 |
| tristan | I ran face first into a wall of spaghetti standing in between me and figuring out how things broke | 11:08 |
| tristan | so I ate the spaghetti first | 11:08 |
| tristan | The above paste is the test case showing breakage | 11:09 |
| tristan | the non-strict side of the test case breaks because of the bug we fixed last week | 11:09 |
| tristan | the strict side shows whether it passes or fails | 11:09 |
| tristan | f761140f is the commit which causes the file modifications to no longer be considered in "whether a workspace has changed" | 11:10 |
| tristan | As I suspected, that happens *directly* after adding ArtifactCache.diff() | 11:10 |
| tristan | Not sure it's related though | 11:10 |
| juergbi | I'm confused | 11:10 |
| juergbi | you say f761140f broke the issue but at the same time you're saying it's a no-behavior-change | 11:11 |
| juergbi | (broke the feature, not the issue, of course) | 11:12 |
| tristan | juergbi, MR !368 which I merged above, is a no-behavior-change refactor | 11:14 |
| tristan | juergbi, https://bpaste.net/show/be95b73bb664 is a test case which shows broken behavior, which I have bisected down to f761140f | 11:14 |
| tristan | juergbi, clear ? | 11:14 |
| juergbi | yes, I misinterpreted 'the above' | 11:15 |
| *** aday has quit IRC | 11:15 | |
| tristan | it's actually very recently broken | 11:16 |
| tristan | Which would explain why people havent been shouting about it, and why mcatanzaro should not be using master but should stick with the other bugs in 1.1.2 | 11:16 |
| *** aday has joined #buildstream | 11:17 | |
| juergbi | I don't see a significant difference in the get_workspace_key => get_key move | 11:17 |
| juergbi | unless it's a workspace path issue | 11:17 |
| juergbi | i.e., it's checksumming the wrong directory. but then added files would also not work | 11:18 |
| tristan | I see the bug | 11:18 |
| tristan | juergbi, tlater; when translating the cache key algo, it became the following: | 11:19 |
| tristan | self.__key = [(relpath, unique_key(self.path)) for relpath, fullpath in filelist] | 11:19 |
| tristan | That should be unique_key(fullpath) I think | 11:19 |
| tristan | yup, that fixes it | 11:21 |
| tristan | The workspace key has turned into a list of paths plus the checksum of the base directory for each path | 11:21 |
| tristan | instead of ever checksumming the files | 11:21 |
| tlater | Oh | 11:21 |
| * tlater apologizes for blowing that up | 11:22 | |
| juergbi | oh, right | 11:22 |
| tristan | so not related to diffing artifacts after all | 11:24 |
| jmac | Do we have any means of specifying multiple mirrors for a given source? | 11:24 |
| tristan | jmac, read https://gitlab.com/BuildStream/buildstream/issues/328 carefully | 11:25 |
| jmac | Right, thanks | 11:26 |
| tristan | The answer is no (or yes if you have used separate aliases, when it gets implemented) | 11:26 |
| gitlab-br-bot | buildstream: merge request (tristan/fix-workspace-change-detections->master: Fix regression in workspace cache key calculation) #369 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/369 | 11:32 |
| jonathanmaw | tristan: in issue https://gitlab.com/BuildStream/buildstream/issues/328 (downloading sources from mirrors), do we want to support mirror aliases that are defined for subprojects to go more than one subproject deep? | 11:47 |
| jonathanmaw | e.g. being able to define the aliases for foo:baz:bar:bar-git? | 11:47 |
| tristan | jonathanmaw, I think these are separate activities, but I certainly think we want that kind of addressing yes | 11:48 |
| tristan | jonathanmaw, addressing is one part; and it unlocks some features, like explicit cross junction tracking (which can be used with project.refs) | 11:49 |
| tristan | jonathanmaw, addressing also unlocks the mirroring details; I think it makes more sense to implement solid addressing of elements and fixing the loader to be able to deal with loading explicitly addressed targets in this way *first* before diving into other features which require it | 11:50 |
| tristan | jonathanmaw, however I would understand if you had incentives to implement source mirroring without cross-junction support first | 11:50 |
| tristan | hmmm, recession of gitlab runners ? https://gitlab.com/BuildStream/buildstream/pipelines/20192642 | 11:51 |
| jonathanmaw | tristan: okay, I'll consider it more seriously once we have a general solution for cross-junction support | 11:53 |
| jonathanmaw | Is there an issue for that? | 11:54 |
| tristan | not that I know of no | 11:54 |
| tristan | there might be | 11:54 |
| gitlab-br-bot | buildstream: merge request (tristan/fix-workspace-change-detections->master: Fix regression in workspace cache key calculation) #369 changed state ("merged"): https://gitlab.com/BuildStream/buildstream/merge_requests/369 | 11:55 |
| *** cs_shadow has joined #buildstream | 12:11 | |
| jonathanmaw | tristan: yeah, looks like the closest we've got is only tangentially-related, where https://gitlab.com/BuildStream/buildstream/issues/331 refers to your E-mail proposal https://mail.gnome.org/archives/buildstream-list/2018-March/msg00030.html, which mentions the junction path prefix | 12:11 |
| jonathanmaw | and that they're currently only used for display purposes, and need some minor work to support using them to address elements. | 12:14 |
| tristan | right | 12:18 |
| tristan | so it was a big day | 12:18 |
| tristan | jmac, I'm skeptical about merging !341 in buildstream, I dont know what it adds; I dont know why we would run the benchmarks against a single version of BuildStream and require a "same hardware" every time, seems less useful than running the benchmarks on a single machine with nothing else running at the time against multiple versions of buildstream, and not caring much about the actual numbers (only caring about the observable *differences*) | 12:21 |
| tristan | If we add that to buildstream repo, as is, I dont see what we get out of it really | 12:21 |
| * tristan is going to eat supper and soothe todays headache with some soju | 12:22 | |
| *** tristan has quit IRC | 12:28 | |
| jmac | We do run the benchmarks on a single machine with nothing else running | 12:38 |
| jmac | I don't see how you can figure out the differences without the absolute numbers | 12:39 |
| jmac | Also, it would have been helpful if you mentioned this two weeks ago when I raised it | 12:39 |
| jmac | It also means the benchmarks script itself is regularly tested against buildstream. We've seen one error last week which would have been caught if benchmarks were running against buildstream regularly. | 12:42 |
| *** bethw has joined #buildstream | 13:38 | |
| paulsherwood | +1 for benchmarks of absolute times | 13:42 |
| paulsherwood | i think tristan has a blindspot about this, tbh... it's the same underlying problem as logs without absolute times imo | 13:42 |
| *** bethw has quit IRC | 14:21 | |
| gitlab-br-bot | buildstream: merge request (311-opening-a-workspace-with-a-cached-build->master: WIP: Resolve "Opening a workspace with a cached build") #370 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/370 | 14:57 |
| gitlab-br-bot | buildstream: merge request (311-opening-a-workspace-with-a-cached-build->master: WIP: Resolve "Opening a workspace with a cached build") #370 changed state ("closed"): https://gitlab.com/BuildStream/buildstream/merge_requests/370 | 14:59 |
| gitlab-br-bot | buildstream: merge request (311-opening-a-workspace-with-a-cached-build->master: WIP: Resolve "Opening a workspace with a cached build") #371 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/371 | 14:59 |
| gitlab-br-bot | buildstream: merge request (issue-21_Caching_build_trees->master: WIP: Issue 21 caching build trees) #372 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/372 | 15:01 |
| gitlab-br-bot | buildstream: merge request (311-opening-a-workspace-with-a-cached-build->master: WIP: Resolve "Opening a workspace with a cached build") #371 changed state ("closed"): https://gitlab.com/BuildStream/buildstream/merge_requests/371 | 15:02 |
| gitlab-br-bot | buildstream: merge request (issue-21_Caching_build_trees->master: WIP: Issue 21 caching build trees) #372 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/372 | 15:17 |
| gitlab-br-bot | buildstream: merge request (issue-21_Caching_build_trees->master: WIP: Issue 21 caching build trees) #372 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/372 | 15:24 |
| *** noisecell has quit IRC | 15:28 | |
| gitlab-br-bot | buildstream: merge request (issue-21_Caching_build_trees->master: WIP: Issue 21 caching build trees) #372 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/372 | 15:29 |
| gitlab-br-bot | buildstream: merge request (issue-21_Caching_build_trees->master: WIP: Issue 21 caching build trees) #372 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/372 | 15:40 |
| *** bethw has joined #buildstream | 15:41 | |
| gitlab-br-bot | buildstream: issue #348 ("[RFE] Add plugin to deploy tarball") changed state ("opened") https://gitlab.com/BuildStream/buildstream/issues/348 | 15:45 |
| gitlab-br-bot | buildstream: issue #349 ("[RFE] Add plugin to generate docker (OCI) images") changed state ("opened") https://gitlab.com/BuildStream/buildstream/issues/349 | 15:45 |
| gitlab-br-bot | buildstream: issue #350 ("[RFE] Add plugin to generate qemu images") changed state ("opened") https://gitlab.com/BuildStream/buildstream/issues/350 | 15:47 |
| gitlab-br-bot | buildstream: issue #351 ("[RFE] Add plugin to deploy to a ostree repo (local and/or remote)") changed state ("opened") https://gitlab.com/BuildStream/buildstream/issues/351 | 15:58 |
| *** dominic has quit IRC | 16:16 | |
| *** toscalix has quit IRC | 16:36 | |
| *** bethw has quit IRC | 16:50 | |
| *** bethw has joined #buildstream | 16:51 | |
| *** bethw has quit IRC | 16:56 | |
| *** bethw has joined #buildstream | 16:56 | |
| *** jonathanmaw has quit IRC | 17:24 | |
| juergbi | tlater: fyi, I wouldn't mind the humanfriendly dependency. My point is more that we need a clearly defined file format, and that aspect may actually be more difficult with humanfriendly | 17:55 |
| *** bethw has quit IRC | 18:20 | |
| *** valentind has joined #buildstream | 18:57 | |
| *** valentind_ has joined #buildstream | 19:19 | |
| *** valentind has quit IRC | 19:20 | |
| *** bethw has joined #buildstream | 19:21 | |
| gitlab-br-bot | buildstream: merge request (jjardon/update_depencies->master: docs: Update list of dependecies) #373 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/373 | 19:59 |
| gitlab-br-bot | buildstream: merge request (jjardon/update_depencies->master: docs: Update list of dependecies) #373 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/373 | 20:00 |
| *** tristan has joined #buildstream | 20:18 | |
| *** xjuan has joined #buildstream | 20:24 | |
| *** bethw has quit IRC | 20:53 | |
| *** aday has quit IRC | 20:55 | |
| *** xjuan has quit IRC | 21:21 | |
| *** bethw has joined #buildstream | 21:38 | |
| gitlab-br-bot | buildstream: merge request (reset-all-workspaces->master: Add option to reset multiple workspaces) #374 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/374 | 22:33 |
| gitlab-br-bot | buildstream: merge request (reset-all-workspaces->master: WIP: Add option to reset multiple workspaces) #374 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/374 | 22:49 |
| *** bethw has quit IRC | 22:58 | |
| gitlab-br-bot | buildstream: merge request (reset-all-workspaces->master: WIP: Add option to reset multiple workspaces) #374 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/374 | 22:58 |
| gitlab-br-bot | buildstream: merge request (reset-all-workspaces->master: Add option to reset multiple workspaces) #374 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/374 | 23:28 |
| gitlab-br-bot | buildstream: merge request (reset-all-workspaces->master: Add option to reset multiple workspaces) #374 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/374 | 23:37 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!