| *** Prince781 has quit IRC | 00:20 | |
| *** Prince781 has joined #buildstream | 02:33 | |
| *** mohan43u has quit IRC | 04:01 | |
| *** Prince781 has quit IRC | 05:14 | |
| *** mohan43u has joined #buildstream | 06:26 | |
| paulsherwood | morning | 06:55 |
|---|---|---|
| paulsherwood | i'm thinking about setting up a new pipeline to build some arm-target things... the current implmenetation is using buildroot for cross-compile | 06:57 |
| paulsherwood | could i use buildstream for this? (the targets are linux applications, not gnome, to run on a small multicore SoC | 06:58 |
| *** Prince781 has joined #buildstream | 06:58 | |
| *** Prince781 has quit IRC | 07:07 | |
| *** Prince781 has joined #buildstream | 07:08 | |
| *** mohan43u has quit IRC | 07:12 | |
| *** Phil has joined #buildstream | 07:27 | |
| *** coldtom has joined #buildstream | 08:03 | |
| *** toscalix has joined #buildstream | 08:08 | |
| *** bethw has joined #buildstream | 08:12 | |
| *** bethw has quit IRC | 08:16 | |
| *** bethw has joined #buildstream | 08:36 | |
| *** jonathanmaw has joined #buildstream | 08:42 | |
| laurence | paulsherwood, a lot of contributors to the project are travelling to GUADEC at the moment, so the channel may be a little quiet | 08:44 |
| finn | jmac, juergbi, is there any metadata you can attach to a command to tell it what directory to chdir into? | 08:58 |
| *** noisecell has joined #buildstream | 08:59 | |
| jmac | Yeah, it's 'build-root' in buildstream | 09:08 |
| tiago | paulsherwood my understanding is that you can only cross-compile until stage 2. The rest, needs to be natively built | 09:10 |
| finn | Maybe you could stick it in the EnvironmentVariable part of Command? | 09:11 |
| jmac | Yes, EnvironmentVariable looks like the best place | 09:11 |
| jmac | `PWD` seems unlikey to clash with a real variable on unix | 09:12 |
| juergbi | it'd be a bit against the spec, though | 09:13 |
| juergbi | according to the spec, the working directory is always the input root | 09:13 |
| juergbi | but that's a bit inconvenient | 09:14 |
| finn | have you any other suggestions? | 09:14 |
| juergbi | we could insert an appropriate 'cd' command into the command line | 09:15 |
| finn | that would also work | 09:15 |
| finn | that would need to be done on bst client of course | 09:15 |
| finn | the addition of the command that is | 09:15 |
| jmac | Possible but I think there might be corner cases where that wouldn't work | 09:16 |
| juergbi | we could also raise it upstream, proposing a spec extension | 09:17 |
| jmac | Remote execution commands are meant to be single commands, not shell scripts, so you'd have to wrap the whole thing in a shell | 09:17 |
| jmac | Shell and directory specification are platform-specific | 09:17 |
| finn | juergbi, raising this upstream could take time and could be a long-term fix | 09:18 |
| finn | I still don't think the Watcher API has even been replaced yet... | 09:18 |
| finn | For short term, environment variable until we get a spec change? | 09:19 |
| jmac | Environment variable feels like a better hack to me than adding a cd | 09:19 |
| finn | I should send an email out about our progress to the community. I could include our thoughts about raising an API change too | 09:20 |
| juergbi | we anyway need to wrap it in a shell line, though, afaict | 09:21 |
| juergbi | e.g., bst commands expect to respect PATH while REAPI doesn't | 09:21 |
| juergbi | and with possible future command batching it's also required in any case | 09:22 |
| juergbi | or am I missing something? | 09:22 |
| paulsherwood | tiago: so the short answer is 'no', not possible to build arm things on x86 via bst? | 09:22 |
| jmac | I noticed all the commands we send are wrapped, didn't know if it was a bst thing | 09:22 |
| jmac | If so, then adding a CD wouldn't be a problem | 09:22 |
| jmac | Apart from needing to know the build system path rules which I'm not sure we handle at the moment | 09:23 |
| juergbi | paulsherwood: short answer is 'no', correct. that said, I've successfully done full cross compilation using buildstream with a scratchbox-like approach (cross compiler + qemu-user with binfmt-misc for the rest) | 09:23 |
| * juergbi has to board the plane now | 09:24 | |
| *** Prince781 has quit IRC | 09:25 | |
| finn | see you in Spain juergbi :) | 09:26 |
| paulsherwood | safe travels all | 09:26 |
| *** toscalix has quit IRC | 09:45 | |
| * tlater notes that tiago's auto-generated profile picture is probably the coolest of them all | 09:48 | |
| tlater | Also since when to commits link to their parent MRs? | 09:48 |
| * tiago looks | 09:49 | |
| jmac | I think we've had that rule for a while tlater, since we don't have merge commits in master | 09:51 |
| * tiago deslikes merge commits | 09:52 | |
| tlater | I know we have that rule, but I've never seen gitlab make your commit message a *link* to the MR | 09:52 |
| jmac | Oh, I see | 09:52 |
| tiago | It makes it easier to revert if necessary, but they add too much noise to the git log. Particular in MR with only 1 commit | 09:53 |
| jmac | That's quite clever, but also a bit annoying | 09:53 |
| tlater | I think it's cute :3 | 09:53 |
| *** mohan43u has joined #buildstream | 09:57 | |
| tiago | valentind, adds68 are you guys responsible for the freedesktop-sdk? | 10:21 |
| adds68 | tiago, yes and jjardon also | 10:22 |
| adds68 | and abderrahim[m] | 10:22 |
| tiago | On aarch64, I get this broken symbolic link for base-sdk-image.bst: | 10:22 |
| gitlab-br-bot | buildstream: merge request (sam/pushing-fix->master: WIP: _artifactcache/pushreceive.py: Avoid premature 'done' messages) #532 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/532 | 10:22 |
| tiago | /home/ubuntu/.cache/buildstream/build/gnu-toolchain-stage1-binutils-89gn_ln4/root/usr/files/bin/sh -> ../../../bin/bash | 10:23 |
| adds68 | tiago, is this when you check it out? | 10:26 |
| adds68 | I am not too clued up on the bootstrapping phase, as valentind did a lot of that work, and he is currently travelling to GUADEC | 10:26 |
| *** mohan43u has quit IRC | 10:27 | |
| jmac | Unfortunately, buildbox doesn't seem to handle 'cd' well: https://paste.gnome.org/pkxwaxdwj | 10:33 |
| finn | try removing the --chdir option? | 10:35 |
| finn | ah also | 10:35 |
| finn | I'm guessing this is command line right? | 10:35 |
| finn | sh -c -e "cd /buildstream " >> try that in quotes? | 10:35 |
| jmac | Hmm, that works | 10:36 |
| jmac | That's odd | 10:36 |
| finn | It's the way it interprets the command line args | 10:37 |
| finn | You can see what happens if you do: ls; ls | 10:37 |
| finn | the first one will work in BuildBox, the second ls will list your actual directory unless you put it in quotes | 10:37 |
| *** WSalmon has joined #buildstream | 10:38 | |
| jmac | Yes, I thought that would happen, but why is "cd /buildstream" treated differently? | 10:38 |
| finn | I'm wondering if sh -c -e is executed followed by cd | 10:39 |
| * finn shrugs | 10:40 | |
| WSalmon | I have been looking at https://buildstream.gitlab.io/buildstream/examples/flatpak-autotools.html it mentions gnome-sdk.gpg but a quick google dose not give a down load link, i presume i am looking in a silly place.. | 10:40 |
| finn | it should download it from the alias right> | 10:41 |
| finn | aliases: | 10:41 |
| finn | gnomesdk: https://sdk.gnome.org/ | 10:41 |
| finn | there should be a keys alias I'm guessing since: | 10:42 |
| finn | gpg-key: keys/gnome-sdk.gpg | 10:42 |
| WSalmon | found it, thanks finn! now to work out how to update the docs :) | 10:43 |
| finn | Phil is doc king | 10:43 |
| finn | jmac, I've extracted the built directory using your tool and gone into: buildstream-install/usr/bin | 10:45 |
| finn | There is a hello executable in there | 10:45 |
| finn | can't run it | 10:45 |
| finn | hello: ELF 64-bit LSB executable, x86-64, version 1 (SYSV), dynamically linked, interpreter /lib/ld-musl-x86_64.so.1, stripped | 10:45 |
| jmac | Ok, is it saying 'file not found'? i.e. failing to find the interpreter? | 10:45 |
| jmac | Do you have /lib/ld-musl-x86_64.so.1 ? | 10:45 |
| finn | bash: ./hello: No such file or directory | 10:45 |
| finn | ta, it needed musl | 10:47 |
| finn | cool, all works then | 10:47 |
| jmac | That's the same error I saw yesterday, anyway. Something in buildbox must cause it to link with that. | 10:48 |
| *** mohan43u has joined #buildstream | 10:55 | |
| *** mohan43u has quit IRC | 11:03 | |
| gitlab-br-bot | buildstream: merge request (caching_build_trees->master: Caching buildtrees) #474 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/474 | 11:11 |
| *** mohan43u has joined #buildstream | 11:13 | |
| *** mohan43u has quit IRC | 11:15 | |
| laurence | finn, let's move the unit-tests for buildgrid back to 'Todo' - https://gitlab.com/BuildGrid/buildgrid/issues/25 | 11:39 |
| * laurence is trying to tidy up all gitlab instances ahead of guadec | 11:39 | |
| laurence | the board should show an accurate picture of the current items that are being worked on | 11:40 |
| laurence | finn, can we please raise an issue to cover the work done so far and the next steps re the demo? | 11:40 |
| *** bethw has quit IRC | 11:45 | |
| noisecell | hi! Im getting this build script: https://paste.baserock.org/himixijide from this element: https://gitlab.com/baserock/definitions/blob/master/elements/gnu-toolchain/base.bst when running the following command: 'bst --option arch armv8l64 source-bundle bootstrap/stage3-sysroot.bst --except gnu-toolchain/stage2.bst' https://gitlab.com/baserock/definitions/blob/master/elements/bootstrap/stage3-sysroot.bst for reference | 12:04 |
| noisecell | my question is: in the autogeneraged script why (set -ex; mv /buildstream/baserock/gnu-toolchain/base.bst/gnu-toolchain-base /buildstream-install) || exit 1) is looking for gnu-toolchain-base inside of base.bst ? | 12:05 |
| noisecell | it is clear that previously gnu-toolchain-base contents have been copied/moved into base.bst | 12:06 |
| noisecell | cp -a "$SRCDIR/gnu-toolchain-base/." '/buildstream/baserock/gnu-toolchain/base.bst' <-- line ^^ | 12:06 |
| noisecell | Im following the instructions of how to bootstrap sysroots: https://gitlab.com/baserock/definitions/blob/master/README.md and more specifically http://wiki.baserock.org/guides/how-to-cross-bootstrap/ trying to crossboostrap to aarch64 | 12:09 |
| noisecell | any help would be appreciated | 12:09 |
| *** bochecha has joined #buildstream | 13:19 | |
| tlater | Umm | 13:42 |
| tlater | Is gitlab broken? I'm trying to comment on an MR, and when I click the little comment button it creates a comment input box for *every single line* in the MR | 13:42 |
| * tlater waves his fist at UIs | 13:43 | |
| * noisecell tries to recreate his question | 13:49 | |
| noisecell | does someone know where to look at the commands that an import element will run when building? | 13:50 |
| laurence | noisecell, apologies for the lack of responses today, some people are travelling to guadec and many others are busy with preparation. | 13:51 |
| laurence | If you still don't get answers, the ML is an option | 13:51 |
| laurence | and also if you are going to guadec you can obvioiusly ask in person :) | 13:51 |
| noisecell | laurence, ok, thank you for the heads up :) | 13:51 |
| jmac | noisecell: Import shouldn't run any commands itself, but integration commands for each element it uses will | 13:52 |
| Phil | noisecell, For integration commands, if you haven't found it already there is https://buildstream.gitlab.io/buildstream/tutorial/integration-commands.html | 13:54 |
| noisecell | jmac, https://paste.baserock.org/himixijide and https://gitlab.com/BuildStream/buildstream/blob/master/buildstream/data/build-module.sh.in suggests that at least creates a directory and moves the content (or something else) there | 13:56 |
| jmac | I'm not familiar with build-module.sh, sorry | 13:59 |
| noisecell | Phil, the integration commands are not failing here https://gitlab.com/baserock/definitions/blob/master/elements/gnu-toolchain/base.bst in relation with the URLs linked before hand. I think the issue is either the autogeneration of that script is nor right or base.bst should never passed as element for a crosscompilation source tarball | 14:00 |
| noisecell | jmac, no worries, thanks for replying | 14:00 |
| *** coldtom has quit IRC | 14:18 | |
| *** coldtom has joined #buildstream | 14:18 | |
| tiago | Is there any documentation about the meaning of (Pulled:0 79 0)→ (Fetched:0 79 0)→ (Built:1 0 0). I failed to decode this | 14:22 |
| jmac | The numbers are the number of jobs processed (succeded), skipped, and failed respectively | 14:24 |
| jmac | There's an expanded version of that printed out when a job finishes | 14:24 |
| gitlab-br-bot | buildstream: merge request (willsalmon/documentation_link->master: Adding a helpful link to the example) #533 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/533 | 14:25 |
| tiago | jmac makes sense, thanks :) | 14:25 |
| gitlab-br-bot | buildstream: merge request (phil/437-workspaces-tutorial->master: Phil/437 workspaces tutorial) #519 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/519 | 14:39 |
| *** finn_ has joined #buildstream | 15:01 | |
| *** Phil has quit IRC | 15:01 | |
| *** Phil has joined #buildstream | 15:02 | |
| *** jonathanmaw has quit IRC | 15:02 | |
| *** finn has quit IRC | 15:03 | |
| *** WSalmon has quit IRC | 15:04 | |
| *** WSalmon has joined #buildstream | 15:09 | |
| *** jonathanmaw has joined #buildstream | 15:18 | |
| *** finn_ has quit IRC | 15:45 | |
| *** Prince781 has joined #buildstream | 15:51 | |
| *** noisecell has quit IRC | 15:57 | |
| *** Prince781 has quit IRC | 16:05 | |
| *** Phil has quit IRC | 16:08 | |
| *** finn has joined #buildstream | 16:09 | |
| *** Phil has joined #buildstream | 16:10 | |
| *** Phil has quit IRC | 16:11 | |
| *** jennis has quit IRC | 16:19 | |
| *** finn has quit IRC | 16:25 | |
| laurence | tlater, please update this ticket with a comment as per our chat earlier - https://gitlab.com/BuildStream/buildstream/issues/135 | 16:47 |
| laurence | and the MR, if you wish - https://gitlab.com/BuildStream/buildstream/merge_requests/347 | 16:47 |
| tlater | laurence: Thanks for the reminder, was just about to update that :) | 16:52 |
| tlater | I also have a few other issues that have been assigned to me, I'll figure out the ones I'm actually working on | 16:53 |
| laurence | thanks a lot | 16:57 |
| *** coldtom has quit IRC | 16:59 | |
| laurence | jonathanmaw, please update the source mirroring ticket - https://gitlab.com/BuildStream/buildstream/issues/328 | 17:06 |
| laurence | i know you've been updating the MR recently too | 17:06 |
| gitlab-br-bot | buildstream: merge request (328-support-for-downloading-sources-from-mirrors->master: Resolve "Support for downloading sources from mirrors") #404 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/404 | 17:08 |
| gitlab-br-bot | buildstream: issue #458 ("Update test suite images") changed state ("opened") https://gitlab.com/BuildStream/buildstream/issues/458 | 17:10 |
| laurence | thanks jonathanmaw | 17:14 |
| *** jonathanmaw has quit IRC | 17:15 | |
| laurence | juergbi, if you have a min, please could you update the ticket for the CAS cache - https://gitlab.com/BuildStream/buildstream/issues/387 | 17:27 |
| laurence | just a few lines summarising the current status | 17:27 |
| *** coldtom has joined #buildstream | 17:29 | |
| juergbi | laurence: the pending points are in the MR, does it really make sense to repeat that (producing noise for subscribers)? | 17:42 |
| gitlab-br-bot | buildstream: merge request (phil/436-add-ubuntu-install-intructions->master: Phil/436 add ubuntu install intructions) #525 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/525 | 18:27 |
| *** Prince781 has joined #buildstream | 18:40 | |
| *** Prince781 has quit IRC | 19:05 | |
| *** toscalix has joined #buildstream | 19:11 | |
| *** Prince781 has joined #buildstream | 19:32 | |
| *** bochecha has quit IRC | 19:36 | |
| *** bochecha has joined #buildstream | 19:39 | |
| *** Prince781 has quit IRC | 19:40 | |
| *** Prince781 has joined #buildstream | 20:33 | |
| *** bochecha has quit IRC | 20:51 | |
| *** WSalmon has quit IRC | 21:45 | |
| *** Prince781 has quit IRC | 21:51 | |
| *** Prince781 has joined #buildstream | 22:29 | |
| *** Prince781 has quit IRC | 22:34 | |
| *** Prince781 has joined #buildstream | 23:05 | |
| *** Prince781 has quit IRC | 23:07 | |
| *** Prince781 has joined #buildstream | 23:08 | |
| *** toscalix has quit IRC | 23:41 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!