| *** tristan has joined #buildstream | 01:30 | |
| gitlab-br-bot | push on buildstream@safe-hardlinks (by Tristan Van Berkom): 1 commit (last: Added fusepy dependency to setup.py) https://gitlab.com/BuildStream/buildstream/commit/22cbe23dca7d28c02b350df1e3a10e20489ecc5e | 01:33 |
|---|---|---|
| gitlab-br-bot | buildstream: Tristan Van Berkom deleted branch safe-hardlinks | 06:26 |
| gitlab-br-bot | push on buildstream@safe-hardlinks (by Tristan Van Berkom): 12 commits (last: Added fusepy dependency to setup.py) https://gitlab.com/BuildStream/buildstream/commit/711f9716d88cd8ff31d4c5845faa8393aea23fba | 06:26 |
| *** tristan has quit IRC | 06:33 | |
| *** tristan has joined #buildstream | 07:05 | |
| *** jonathanmaw has joined #buildstream | 08:02 | |
| *** tiagogomes has joined #buildstream | 08:44 | |
| *** tiagogomes has quit IRC | 08:52 | |
| *** tiagogomes has joined #buildstream | 08:55 | |
| *** jude has joined #buildstream | 09:04 | |
| *** ssam2 has joined #buildstream | 09:08 | |
| jonathanmaw | hmm, I think I'm doing something wrong when trying to use debootstrap-ostree. I always end up with "error: openat: No such file or directory" when creating the commit | 09:11 |
| jonathanmaw | the script usually continues from there, but it looks like this wasn't an acceptable error, as I don't see the branch "debian/testing/amd64" in the ostree repo that was created. | 09:12 |
| *** zza has joined #buildstream | 09:12 | |
| *** zza has quit IRC | 09:14 | |
| jonathanmaw | tristan: my command-line was "./debootstrap-ostree.sh --arch amd64 --ostree /home/jonathanmaw/workspace/buildstream/debian-ostree/ --config data/multistrap.conf --workdir $PWD/work", am I missing anything? | 09:15 |
| jonathanmaw | I'd thought it might have been running out of disk space again, but watching disk usage closely, that seems not to be the case | 09:43 |
| *** tristan has quit IRC | 10:11 | |
| *** tristan has joined #buildstream | 10:21 | |
| *** ChanServ sets mode: +o tristan | 10:22 | |
| tristan | jonathanmaw, so actually, what is /home/jonathanmaw/workspace/buildstream/debian-ostree/ ? | 11:12 |
| tristan | jonathanmaw, it can either be A.) An ostree repo it created in the last run... or B.) A non-existing directory that it will create | 11:13 |
| tristan | jonathanmaw, looks like that was your issue | 11:13 |
| tristan | if problem persists, try unquoting "${gpg_args}" and "${gpg_homedir}" (and let me know if that was needed, cause we'll update the repo) | 11:13 |
| tristan | I'm not sure I've tested it without passing the gpg stuff since I added the gpg stuff, so that might be broken | 11:14 |
| tristan | Ok so, I guess "Here goes nothing" | 11:15 |
| * tristan lands the branch, after much local testing, and hopes very much that it works on other peoples computers | 11:15 | |
| tristan | Or, jonathanmaw wanna test it once for me before I push the branch ? | 11:16 |
| gitlab-br-bot | buildstream: issue #38 ("Lost file metadata in artifacts and images") changed state ("opened") https://gitlab.com/BuildStream/buildstream/issues/38 | 11:39 |
| *** tristanmaat has joined #buildstream | 12:15 | |
| jonathanmaw | tristan: ok | 12:23 |
| jonathanmaw | you mean the branch for lost file metadata? | 12:23 |
| jonathanmaw | also, the build seemed to finish after removing the quotes, but I'm not sure what else I changed to cause that | 12:24 |
| jonathanmaw | I'll run it again to see what happens | 12:24 |
| tristan | jonathanmaw, there is no such branch no, that is a bit bigger of a project (See #38 I filed above) | 12:26 |
| tristan | jonathanmaw, the branch is called safe-hardlinks | 12:26 |
| jonathanmaw | ok | 12:27 |
| tristan | jonathanmaw, anything that you build which runs integration commands is good enough to test it | 12:27 |
| tristan | jonathanmaw, yeah regarding the script and quotes, I think the lack of quotes means that bash is not explicitly passing zero length strings in the ostree argv | 12:28 |
| *** tristanmaat has quit IRC | 12:28 | |
| jonathanmaw | tristan: that's been my experience with them in the past | 12:28 |
| tristan | nod, it's confusing | 12:28 |
| tristan | I usually quote almost everything at first, and then remove quotes where they dont make sense, seems to be decent rule of thumb that's working for me so far | 12:29 |
| tristan | I'll remove them in the repo | 12:29 |
| *** zvb has joined #buildstream | 12:30 | |
| zvb | amdj: FUNDRAISER AND PROMOTIONAL WEEK!!! | 12:30 |
| zvb | +-+-+-+-+-+-+-+-+-+ | 12:30 |
| zvb | |c|h|a|r|y|b|d|i|s| | 12:30 |
| zvb | +-+-+-+-+-+-+-+-+-+ | 12:31 |
| zvb | WE ARE TRYING TO GET FUNDING TO GET WEBSITE BACK UP!! | 12:31 |
| zvb | CHECK OUT #CHARYBDIS | 12:31 |
| zvb | IRC.FREENODE.NET | 12:31 |
| zvb | tristan ssam2 jude tiagogomes jonathanmaw violeta juergbi mattiasb jjardon[m] albfan[m] waltervargas[m] ironfoot persia gitlab-br-bot brlogger hergertme Laurence dabukalam | 12:31 |
| *** zvb has left #buildstream | 12:31 | |
| tristan | Nice ! | 12:31 |
| tristan | I cant wait to fund them ! | 12:31 |
| jonathanmaw | he did a driveby in another server as well | 12:31 |
| ironfoot | we are having a lot of spam :/ | 12:32 |
| tristan | ironfoot, I wouldnt actually call it that | 12:32 |
| tristan | ironfoot, but here on GIMPNet we seem to have a noticeable amount compared to the almost nothing we had few months ago | 12:32 |
| *** tristanmaat has joined #buildstream | 12:33 | |
| tristan | jonathanmaw, so if you run a build with that... if you type `mount` in another terminal you should see one SafeHardlinks mounted for every build that is running integration commands | 12:35 |
| tristan | jonathanmaw, I'm mostly just concerned that your build does not completely break for some reason | 12:35 |
| tristan | jonathanmaw, also you will require a re-setup... like a `pip uninstall buildstream && pip install --user -e .` | 12:36 |
| *** tristanmaat_ has joined #buildstream | 12:36 | |
| *** tristanmaat has quit IRC | 12:36 | |
| tristan | jonathanmaw, oh; and you did remember to do that part I didnt warn you about right ? | 12:38 |
| jonathanmaw | tristan: I can't say I did | 12:38 |
| tristan | jonathanmaw, with debootstrap-ostree, you'll have to manually make the substitution for multistrap.conf.in :) | 12:39 |
| tristan | heh | 12:39 |
| tristan | I just assumed you noticed that | 12:39 |
| jonathanmaw | tristan: I found that out relatively quickly :P | 12:39 |
| tristan | yeah, just checking :) | 12:39 |
| *** tristanmaat_ has quit IRC | 12:44 | |
| *** tristanmaat has joined #buildstream | 12:55 | |
| *** tristanmaat has quit IRC | 13:22 | |
| jonathanmaw | tristan: I seem to be getting https://pastebin.com/6TRBYRyH whenever I try to stage anything | 13:26 |
| *** tristanmaat has joined #buildstream | 13:28 | |
| jonathanmaw | (full build at https://pastebin.com/xPuAq8Ni) | 13:30 |
| tristan | thats not while staging | 13:31 |
| tristan | odd though, havent seen that one before :-S | 13:31 |
| tristan | hmmm, ok so that is from the unlink | 13:31 |
| tristan | jonathanmaw, I have a feeling that that was happening anyway :-S | 13:32 |
| tristan | regardless of copy-on-write hardlinks | 13:32 |
| tristan | hrm... | 13:32 |
| tristan | jonathanmaw, what does bwrap --version tell you ? | 13:35 |
| jonathanmaw | bubblewrap 0.1.7 | 13:35 |
| tristan | looks like mine is older | 13:37 |
| * tristan tries with new one, maybe introduced a bug ? | 13:37 | |
| tristan | jonathanmaw, the good news, or the strange news, is that the part where you use the new hardlinks filesys, doesnt break | 13:39 |
| tristan | jonathanmaw, what does `mount` tell you, after everything ? | 13:39 |
| tristan | Are there leaked mount points maybe ? | 13:39 |
| jonathanmaw | tristan: I have a lot of mounts for /var/lib/mount/baserock-vm-*/dev/pts. I think that's something separate from bst though | 13:40 |
| tristan | so integration commands work, but when you run with the root readonly and without the mount point, it breaks | 13:41 |
| tristan | yeah looks like | 13:41 |
| tristan | jonathanmaw, what is your filesystem ? | 13:41 |
| tristan | heh | 13:41 |
| tristan | oddly it seems that removing the dev file when mounted on *top* of our fuse, has no problem, but when we bind mount the device in the namespace on top of your real filesys, that's where it is "busy" | 13:42 |
| jonathanmaw | tristan: ext4 | 13:42 |
| tristan | it *shouldnt* be a problem | 13:43 |
| tristan | jonathanmaw, no llvm ? | 13:43 |
| tristan | not sure that changes anything | 13:43 |
| jonathanmaw | llvm, too | 13:43 |
| tristan | I am ext4 also on llvm with weird luks setup | 13:43 |
| tristan | cryptfs { llvm { ext4 } } I think it is | 13:44 |
| jonathanmaw | that sounds like mine, too | 13:44 |
| *** xzq has joined #buildstream | 13:47 | |
| tristan | ok it's not bwrap | 13:48 |
| tristan | mine still works :-S | 13:48 |
| tristan | jonathanmaw, ok well it's late for me, thanks for trying it out, I guess good thing I didnt push right away | 13:48 |
| jonathanmaw | I've cleared up all my weird mounts, I'll give it another go now | 13:50 |
| jonathanmaw | hrm, still happening | 13:52 |
| tristan | well, it might tell you if it leaves a mount behind ? | 13:52 |
| tristan | jonathanmaw, and what if you do rm path/to/dev/full ? | 13:52 |
| tristan | in the failed build directory ? | 13:52 |
| tristan | jonathanmaw, possibly we need a busy wait on EBUSY for the weird cases ? | 13:53 |
| tristan | I guess you must be able to remove it | 13:53 |
| tristan | probably that'll do it | 13:54 |
| jonathanmaw | tristan: after the sandbox has raised its error, I can remove the /dev/full file | 13:56 |
| jonathanmaw | hrm, I have some mounts hanging around this time | 13:56 |
| tristan | you have extra mounts ? | 13:56 |
| tristan | or are you in an interactive mode ? | 13:56 |
| tristan | and there are other builds sleeping ? | 13:56 |
| jonathanmaw | three that look like "SafeHardlinkOps 151204036 129689628 15152968 90% /home/jonathanmaw/.cache/buildstream/build/tools-u-boot-zw2zvr9w/scratch/root" | 13:56 |
| jonathanmaw | my build's finished | 13:56 |
| tristan | And it left that behind | 13:57 |
| jonathanmaw | was running in interactive, I terminated as soon as I got a problem | 13:57 |
| tristan | Odd | 13:57 |
| tristan | Ok that's a bug | 13:57 |
| jonathanmaw | ah, they've gone now | 13:57 |
| tristan | by themselves | 13:57 |
| jonathanmaw | yep | 13:57 |
| tristan | Right, but they should be joined | 13:57 |
| tristan | ok not a serious bug, I think I've had that bug for a while | 13:57 |
| tristan | difficult to reap everything on termination :-S | 13:58 |
| tristan | jonathanmaw, anyway, tomorrow I'll add a little busy wait loop with maybe a 5ms sleep in the case that it's busy, and only raise the BUG after a certain number of tries | 14:00 |
| tristan | then you can try it, it will certainly work since you were able to remove it, and then I can merge :) | 14:01 |
| tristan | At least it didnt lock up your kernel :D | 14:01 |
| tristan | hehehehe | 14:01 |
| tristan | it was doing that to me on saturday | 14:01 |
| jonathanmaw | :) | 14:01 |
| tristan | Had to syncrhonize the fuse mount so that it's not in the middle of mounting when bwrap tries to mount on top of it | 14:02 |
| ssam2 | just noticed that buildstream _frontend/widget.py contains a miniature GUI toolkit ;-) | 14:03 |
| tristan | ssam2, height for width task list ! | 14:04 |
| tristan | haha | 14:04 |
| tristan | I think that is in status.py | 14:04 |
| tristan | ssam2, it's a bit sloppy, but objectifying the columns of the log lines saves code, and makes it a bit easier to manage | 14:05 |
| tristan | that said, it was done in a hurry, could be much better :) | 14:06 |
| * tristan changes location to home, and obtains a spicy chicken | 14:06 | |
| *** tristan has quit IRC | 14:10 | |
| *** xlo has joined #buildstream | 14:10 | |
| xlo | amdj: FUNDRAISER AND PROMOTIONAL WEEK!!! | 14:12 |
| xlo | +-+-+-+-+-+-+-+-+-+ | 14:12 |
| xlo | |c|h|a|r|y|b|d|i|s| | 14:12 |
| xlo | +-+-+-+-+-+-+-+-+-+ | 14:12 |
| xlo | WOW THE PROMOTION WEEK IS A SUCCESS SO FAR!!! | 14:12 |
| xlo | THANKS GUYS!! FOR PUTTING UP WITH THE SPAM!! | 14:12 |
| xlo | WE ARE TRYING TO GET FUNDING TO GET WEBSITE BACK UP!! | 14:12 |
| xlo | CHECK OUT #CHARYBDIS | 14:12 |
| xlo | IRC.FREENODE.NET | 14:12 |
| xlo | tristanmaat ssam2 jude tiagogomes jonathanmaw violeta juergbi mattiasb jjardon[m] albfan[m] waltervargas[m] ironfoot persia gitlab-br-bot brlogger hergertme Laure | 14:12 |
| *** xlo has left #buildstream | 14:12 | |
| jonathanmaw | that's a novel opinion of spam | 14:14 |
| *** tristan has joined #buildstream | 14:46 | |
| *** jonathanmaw has quit IRC | 16:18 | |
| *** jonathanmaw has joined #buildstream | 16:21 | |
| *** ssam2 has quit IRC | 16:30 | |
| *** ssam2 has joined #buildstream | 16:31 | |
| *** jonathanmaw has quit IRC | 16:58 | |
| *** ssam2 has quit IRC | 17:01 | |
| *** ssam2 has joined #buildstream | 17:01 | |
| *** jude has quit IRC | 17:34 | |
| *** tristanmaat has quit IRC | 18:28 | |
| *** ssam2 has quit IRC | 18:37 | |
| *** jude has joined #buildstream | 19:26 | |
| *** jude has quit IRC | 19:35 | |
| *** jude has joined #buildstream | 19:48 | |
| *** jude has quit IRC | 20:24 | |
| *** tristan has quit IRC | 21:04 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!