| radiofree | so good news, 15.10 can build 15.10 | 00:10 |
|---|---|---|
| radiofree | and 15.10 can build master, with morph master | 00:11 |
| radiofree | what i don't know is why i have a board that fails to do that.... | 00:12 |
| radiofree | http://cmusphinx.sourceforge.net/wiki/tutorialam#troubleshooting | 00:12 |
| *** wschaller has joined #baserock | 03:00 | |
| *** wschaller has quit IRC | 03:39 | |
| *** zoli__ has joined #baserock | 04:49 | |
| *** dabukalam_ has joined #baserock | 05:06 | |
| *** richard_1aw has joined #baserock | 05:07 | |
| *** wdutch__ has joined #baserock | 05:09 | |
| *** bwh_ has joined #baserock | 05:09 | |
| *** dabukalam has quit IRC | 05:09 | |
| *** wdutch has quit IRC | 05:09 | |
| *** bwh has quit IRC | 05:09 | |
| *** richard_maw has quit IRC | 05:09 | |
| * paulsherwood is tickled by the quote 'Troubleshooting is not rocket science. For all issues you may blame yourself. You are most likely the reason of failure.' | 05:30 | |
| *** zoli__ has quit IRC | 05:38 | |
| *** zoli__ has joined #baserock | 06:05 | |
| rdale | morph is telling me i've no space to build anything: /src/tmp requires 4000000000 bytes free, has 1577066496 | 07:45 |
| rdale | but i have loads of space: /dev/sdb 29.4G 11.0G 16.9G 40% /src | 07:45 |
| Kinnison | morph demands 4G of free space and the mechanism it's using said there's only 1.5G left | 07:46 |
| Kinnison | rdale: is that the output of 'df /src/tmp' ? | 07:46 |
| rdale | ah ok i see: tmpfs 1.5G 0 1.5G 0% /tmp | 07:46 |
| Kinnison | odd that /src/tmp is a tmpfs | 07:48 |
| Kinnison | did you mount one there specially? | 07:48 |
| Kinnison | normally we recommend /src/tmp so that we're using the space in /src | 07:48 |
| rdale | yes, i didn't do anything to create /src/tmp and so i'm not sure what has happened | 07:49 |
| rdale | i'm using an image i built from the definitions master with morph from the master branch yesterday | 07:52 |
| petefoth | I believe all 5 references to 'PYTHONPATH="$morphpath"' on w.b.o are now correct and all unnecessary references to pylru have gone, so all tests on the documentation should now pass :) Though there's clearly some refactoring that could be done to avoid duplication! | 07:55 |
| Kinnison | rdale: is /src/tmp a symlink to /tmp or something odd? | 07:56 |
| rdale | the contents are the same as /tmp | 07:56 |
| rdale | but it isn't a sym link | 07:57 |
| paulsherwood | rdale: i bet /etc/morph.conf is not linked to /src/morph.conf ? | 07:58 |
| rdale | it seems to have the proper link: http://paste.baserock.org/ipuyemofam | 07:59 |
| Kinnison | rdale: can you pastebin the output of 'mount' ? | 08:00 |
| *** a1exhughe5 has joined #baserock | 08:01 | |
| *** gfinney has joined #baserock | 08:01 | |
| paulsherwood | rdale: cat morph.conf ? does it specify tmpdir correctly? | 08:01 |
| rdale | http://paste.baserock.org/refamanepe | 08:01 |
| Kinnison | tmpfs on /src/tmp type tmpfs (rw,relatime) | 08:02 |
| Kinnison | so there is a tmpfs explicitly mounted there | 08:02 |
| Kinnison | rdale: is that in /etc/fstab ? | 08:02 |
| rdale | this is /etc/fstab: http://paste.baserock.org/pofononiqe | 08:04 |
| Kinnison | Okay that's sane | 08:04 |
| * Kinnison wonders where on earth that spare mount for /src/tmp came from | 08:04 | |
| rdale | this is morph.conf: http://paste.baserock.org/doroyukexe | 08:04 |
| Kinnison | your configurations look sane | 08:04 |
| Kinnison | what's odd is that /src/tmp is *mounted* as a tmpfs despite that not being in fstab | 08:04 |
| paulsherwood | rdale: why is your label src2 ? | 08:05 |
| rdale | because i had another partition that i used to mount as /src which had a label of src | 08:05 |
| paulsherwood | rdale: this is a good one. i don't know :/ | 08:10 |
| rdale | my morph is from the master as at fri 6th march | 08:11 |
| Kinnison | rdale: I say again, all the general configuration is sane, I'm confused as to where that tmpfs mount came from | 08:12 |
| paulsherwood | rdale: ah... that tmpfs is locked at 1.5G, wihch is not enough for morph | 08:12 |
| paulsherwood | ie /src/tmp | 08:12 |
| rdale | yes | 08:13 |
| Kinnison | rdale: short term, umount /src/tmp | 08:13 |
| Kinnison | rdale: and then morph should work | 08:13 |
| paulsherwood | i don't know what set that, but i don't have it | 08:13 |
| Kinnison | rdale: longer term, next time you reboot, see if /src/tmp is mounted as a tmpfs again | 08:13 |
| Kinnison | rdale: if it is, we need to work out why | 08:13 |
| Kinnison | rdale: if it's not, then bogons ate your hamster and you can continue baserocking :) | 08:13 |
| rdale | i'm just trying booting my 'factory' kernel instead of 'development' and will see how /src/tmp is mounted | 08:13 |
| paulsherwood | rdale: my current (very recently updated, if not master) system has http://paste.baserock.org/atamanozey | 08:14 |
| rdale | yes, i might have some small bogons hidden under my desk | 08:14 |
| rdale | after rebooting it is gone - perhaps i created it by accident, but i can't think how | 08:16 |
| * paulsherwood wipes his brow, turns off the siren and the flashing lights | 08:17 | |
| * petefoth created https://storyboard.baserock.org/#!/story/26 to cover the --no-git-update issue | 08:27 | |
| *** mdizzle has joined #baserock | 08:37 | |
| *** sambishop has joined #baserock | 08:47 | |
| SotK | petefoth: do you still plan to send patches for that? :) | 08:47 |
| petefoth | SotK: not in a hurry. I was planning to look into the second task on that story 'Check if there are other calls that should honour this parameter but don't' to try and work out what shoudl and should not be possible when working offline. If you've got a patch you're happy with, then don't let me stop you :) | 08:50 |
| jjardon | petefoth: maybe that should be a task of a more general story (like "improve offline mode in morph")? | 08:52 |
| petefoth | jjardon: maybe. I don't feel strongly. I just wanted to investigate how morph *should* behave before submitting patches to change how it *does* behave | 08:54 |
| petefoth | Because if you dig into what should be possible when working offline, you might discover that 'morph edit' is something that should onl work when you are online, in which case you'd have to revert any patches you made to allow it to work offline. | 08:56 |
| petefoth | I was wondering about 'morph checkout', but at the moment I don't know enough about what that command is supposed to do and what it is for - that's why I was going to dig | 08:57 |
| SotK | I feel like we should make as much as possible work offline, on the proviso that you only want to work with things you have in your repo cache | 08:58 |
| * SotK disappears | 09:00 | |
| petefoth | SotK: that's my gut feeling too, I just don't like changing code until I know what I'm trying to do (which is probably why I don't do a lot of coding these days) | 09:00 |
| *** bashrc_ has joined #baserock | 09:02 | |
| jjardon | SotK: exactly. Its not like we are keeping up to date with changes | 09:04 |
| jjardon | I mean, we are not tracking branches in our components, but fixed sha's | 09:05 |
| bashrc_ | that seems reasonable to avoid breakage | 09:05 |
| *** jonathanmaw has joined #baserock | 09:06 | |
| bashrc_ | if everything was on HEAD I imagine that chaos would reign | 09:06 |
| *** gary_perkins has joined #baserock | 09:07 | |
| *** franred has joined #baserock | 09:07 | |
| bashrc_ | I like SotK's hotfix idea. That sounds something like an upgrade. In most cases recreating and deploying an entire new image is not really a practical update method | 09:12 |
| *** tiagogomes_ has joined #baserock | 09:26 | |
| *** CTtpollard has joined #baserock | 09:28 | |
| * straycat thinks that having ps1 set to '% ' and using git push origin HEAD is a bad idea | 09:31 | |
| * straycat also thinks that morph should do as much offline as possible | 09:31 | |
| jjardon | bashrc_: depends on your use case, some people may be interested on detecting bugs as soon as possible, so they can fix them as soon as possible as well | 09:32 |
| straycat | jjardon, having refs everywhere meant that anyone who wanted to do anything needed to run morph petrify first if they didn't want to have everything changing underneath them | 09:33 |
| * petefoth had a converation with paulsherwood this morning about why 'no-git-update' should be the default, and suspects that the 'correct' answer depends on how an individual chooses to use morph and what they use it for | 09:33 | |
| petefoth | Is it something that can be added to morph.conf? | 09:34 |
| * robtaylor thinks you should pull when you want to pull.. | 09:34 | |
| robtaylor | but we seem to be losing morph pull :/ | 09:34 |
| straycat | what is morph pull? | 09:35 |
| straycat | Oh i see | 09:35 |
| petefoth | 'This is a wrapper for the `git pull` command. It also deals with | 09:35 |
| petefoth | pulling any binary files that have been added to the repository using | 09:35 |
| petefoth | git-fat. | 09:35 |
| * straycat never uses that | 09:36 | |
| straycat | *shrug* | 09:36 |
| robtaylor | yeah, its probably wrong for what i mean, so just as well it's going away in in its present state | 09:37 |
| petefoth | straycat: what do you use baserock / morph for? | 09:37 |
| * petefoth wants to understand why people use it, and what for because that's how you work otu what it *should* do | 09:38 | |
| jjardon | straycat: yeah, I know the reasons why we use sha, not arguing against that. But making a point that we should be able to work offline because we do not need to pull the repos so often | 09:38 |
| straycat | robtaylor, yeah what i understood by your use of 'pull' was 'update repo caches' | 09:41 |
| jjardon | straycat: BTW, sadly, 'morph petrify' is gone | 09:41 |
| straycat | jjardon, *nod* | 09:41 |
| straycat | jjardon, oh yes that happened too >.> | 09:43 |
| straycat | petefoth, i don't really know right now tbh | 09:44 |
| petefoth | straycat: :) I suspect you're not the only one! | 09:44 |
| straycat | p.s. i meant that don't know how to usefully discuss this point further right now and should be doing something else, not that i don't know what i use morph for, obviously | 09:46 |
| *** ssam2 has joined #baserock | 09:56 | |
| *** ChanServ sets mode: +v ssam2 | 09:56 | |
| jjardon | Do anyone know if the manifest proposal for definitions is documented anywhere? (A mail with a brief explanation would work) | 09:58 |
| SotK | jjardon: the "store all the shas in one file" thing? | 09:59 |
| ssam2 | there was a thread on baserock-dev a while back where robtaylor proposed the idea. maybe a few actually | 10:01 |
| ssam2 | I don't think anyone's taken the time to write it up on the wiki | 10:01 |
| jjardon | SotK: yep | 10:01 |
| jjardon | ssam2: thanks i will take a look to the ml | 10:01 |
| richard_1aw | it was also revisited at the baserock meet-up | 10:09 |
| richard_1aw | oh, excuse me | 10:09 |
| *** richard_1aw is now known as richard_maw | 10:09 | |
| petefoth | https://storyboard.baserock.org/#!/story/26 is now entitled Improve offline functionality in 'morph' and fixing `morph edit` is now just a task within that story | 10:09 |
| richard_maw | that's better | 10:09 |
| radiofree | ssam2: btw used 15.10 to build 15.10 and master last night | 10:09 |
| radiofree | so now i have no idea what's going on... | 10:09 |
| pedroalvarez | radiofree: good news :) | 10:09 |
| *** edcragg has joined #baserock | 10:13 | |
| jjardon | petefoth: thanks | 10:17 |
| ssam2 | radiofree: thank heavens for that | 10:17 |
| ssam2 | did it build master with a clean cache? | 10:18 |
| ssam2 | i suppose that wouldn't matter probably.. | 10:18 |
| radiofree | yep, morph master worked on master as well | 10:18 |
| radiofree | i'm trying to reproduce the issue i had on my main board | 10:18 |
| radiofree | but i'm guessing i won't be able to now... | 10:19 |
| petefoth | SotK: I am going to try to submit a patch to fix the 'morph edit' taks | 10:19 |
| *** CTtpollard has quit IRC | 10:20 | |
| ssam2 | radiofree: do you still have the broken staging area? you could compare it against the staging area from a sucessful build | 10:20 |
| ssam2 | (except you'll have to make the successful build fail at the last minute so the staging area is kept) | 10:20 |
| *** CTtpollard has joined #baserock | 10:20 | |
| SotK | petefoth: great, I won't write one then :) | 10:20 |
| radiofree | it probably would have been sensible for me to keep a copy of it right | 10:21 |
| * radiofree doesn't have it | 10:21 | |
| ssam2 | that gives us a reason to be happy if we see the issue again :) | 10:23 |
| paulsherwood | radiofree: actually, given your symptoms, maybe it was just corrupted cache? | 10:26 |
| radiofree | but i deleted /src/cache each time | 10:27 |
| paulsherwood | not, then | 10:27 |
| radiofree | and of course i can't reproduce it now | 10:32 |
| paulsherwood | heh | 10:36 |
| jjardon | Hi, Is anyone here building baserock system with multiuser support? I ask that because, even I do not know almost nothing about pam, the configuration of it in our baserock systems seems to be incomplete / broken | 10:50 |
| Kinnison | I think pam is only present because it was build-depended on by things. I don't know of anyone actively using it yet | 10:51 |
| jjardon | Kinnison: maybe, but I think some components can be broken / misconfigured; for example: systemd is in charge of creating XDG_RUNTIME_DIR through the systemd_pam module; currently XDG_RUNTIME_DIR is not created because pam is not correctly configured to be aware of that module (AFAIU) | 10:55 |
| Kinnison | jjardon: Quite likely | 10:55 |
| jjardon | As I said I do not know anything about pam, only suspect this after reading about it yesterday. I was wondering if someone with more experience could take a look | 10:56 |
| Kinnison | Hopefully someone will be able to -- sadly I can't help since I'm also a fairly newbie person when it comes to pam | 10:56 |
| jjardon | I created a storyboard with some of the tasks I think we should do, if anyone interested: https://storyboard.baserock.org/#!/story/21 | 10:58 |
| pdar | tail` | 10:58 |
| pdar | please excuse me | 10:58 |
| Kinnison | :) | 10:59 |
| *** CTtpollard has quit IRC | 11:11 | |
| tlsa | I want to make `morph build` detect when a chunk ref is not anchored, and warn the user in that case | 11:16 |
| tlsa | currently my plan is http://paste.baserock.org/raw/ecemominor | 11:16 |
| tlsa | but that has the side-effect of making unpetrify-ref manditory | 11:16 |
| tlsa | what do people think? Any better ideas? | 11:17 |
| persia | So, I'd like unpetrify-ref to be obsolete, rather than mandatory. | 11:18 |
| Kinnison | the git operations will slow things down -- try not to do that | 11:19 |
| Kinnison | tlsa: I'd go with "Warn if ref: doesn't look like a sha" | 11:19 |
| tlsa | Kinnison: I sent a patch to the list for that | 11:19 |
| Kinnison | tlsa: you could temper that with "and also isn't the current system branch name" | 11:19 |
| Kinnison | But I don't know if system branches remain 'a thing' | 11:19 |
| persia | For the approach, I worry that only validating the unpetrify-ref means that things break if that happened to be a branch, and the commit the user is tracking happens to have been tagged. | 11:19 |
| * Kinnison knows there has been lots of talk about whether the workflow stuff works | 11:19 | |
| persia | Most of the folk who talk about it seem to say "I only run morph {checkout|branch} once per devel system", which implies they aren't using it. | 11:20 |
| paulsherwood | afaict expert users eventually just use git (and i count myself as one of those, radiofree as anonther) | 11:20 |
| tlsa | Kinnison: http://listmaster.pepperfish.net/pipermail/baserock-dev-baserock.org/2015-March/012609.html | 11:20 |
| Kinnison | Right, so the system-branch thing can be ignored | 11:20 |
| persia | paulsherwood seems to be against it, yet has shared that he has tens of system branches. | 11:20 |
| paulsherwood | persia: not any more. | 11:21 |
| paulsherwood | i had hundreds, until i realised it was the wrong way to work | 11:21 |
| persia | paulsherwood: My apologies: I thought I remembered backscroll of you explaining system branches in the last couple weeks. | 11:21 |
| Kinnison | tlsa: so you have a review asking you to consider if the cmdtest tests you updated would be better expressed as yarns or just deleted | 11:21 |
| Kinnison | tlsa: noone seems to dislike your idea though | 11:22 |
| tlsa | Kinnison: yep, but I have been asked to look at reproduciblity guarantees | 11:23 |
| persia | I don't think morph can provide those. | 11:23 |
| tlsa | and to ensure its reproducable, we can't accept a referenceing a dangling commit | 11:23 |
| jjardon | tlsa: unpetrify-ref only have informative value (Its basically a comment of what ref: means), I do not think making it mandatory is a good idea | 11:23 |
| persia | The best we can do is report whether we expect a given definition commit to be unreproducible if someone runs `git gc` | 11:23 |
| Kinnison | tlsa: That's got to be not part of 'morph build' though IMO | 11:24 |
| persia | +1 | 11:24 |
| Kinnison | tlsa: If you do anythign which slows builds, people will be most unhappy | 11:24 |
| tlsa | hmm, ok | 11:24 |
| * Kinnison wonders if his order of fresh fingers is in yet. These ones are getting tired | 11:24 | |
| tlsa | so it would be something like `morph validate-refs`? | 11:25 |
| persia | That seems right. | 11:25 |
| paulsherwood | persia: i have one system branch on each baserock machine i operate | 11:25 |
| persia | And have it check to make sure every ref is anchored: it need not be HEAD of a branch | 11:25 |
| paulsherwood | then i reset the state of that to whatever i need to work on eg git reset --hard jrandom-branch | 11:26 |
| persia | paulsherwood: That makes more sense, considering your statements that branches don't help anymore :) | 11:26 |
| persia | Although I'd prefer the workload to involve `git checkout ${BRANCH}` and `git checkout -b ${BRANCH}`, but that means morph not caring anymore. | 11:26 |
| paulsherwood | branches were just filling up space, and i was losing track, and having to cd all the time | 11:26 |
| paulsherwood | persia: agreed, if what you mean is, can morph just work in a checked out git repo please | 11:27 |
| paulsherwood | with some standard for where it looks for checked out versions of chunks vs that directory | 11:27 |
| persia | That is what I mean by "not caring", but I also think that any workflow that recommends `git reset` is likely to cause a user to shoot themselves in the foot. | 11:28 |
| persia | I actually don't want that to be standard. I arrange my local checkouts in a variety of ways. The location is best determined by the repo: field. | 11:28 |
| persia | We might recommend users typically put stuff in some standard location, but morph should not scold them if they decide something else works better for them (if nothing else, not checking to determine if scolding is necessary is faster) | 11:29 |
| * jjardon still doesnt understand the usefulness of morph checkout/morph branch after the first time you execute them, neither the need of workspaces | 11:33 | |
| ssam2 | i'd like `morph build` and `morph deploy` to not require a workspace, but don't have time to do a patch for that any time soon | 11:33 |
| jjardon | storyboard! ;) | 11:34 |
| jjardon | https://storyboard.baserock.org/#!/story/27 | 11:38 |
| ssam2 | heh, thanks | 11:39 |
| radiofree | ssam2: i looked at that a while ago | 11:39 |
| radiofree | ran away screaming though | 11:39 |
| radiofree | there's a lot of code there that depends on it being in a i) git repo ii) workspace | 11:39 |
| ssam2 | the temporary build branch stuff would have to be disabled in this case, I think | 11:39 |
| ssam2 | and it would still be in a git repo | 11:39 |
| radiofree | yeah git repo is fine, i was looking at it because i wanted to do morph deploy image.raw, in which case it wouldn't really need to care about it being a git repo | 11:40 |
| radiofree | sorry morph upgrade image.raw | 11:40 |
| ssam2 | oh yeah | 11:40 |
| radiofree | which i suppose is morph deploy --upgrade... | 11:40 |
| radiofree | is 30G enough for a /src directory? | 11:42 |
| Kinnison | I'd recommend 50 to 100 | 11:42 |
| Kinnison | but 30 might be enough if you're mostly playing at the leaves | 11:43 |
| radiofree | ah, ok | 11:43 |
| radiofree | i stumbled upon some 30GB ssd's | 11:43 |
| radiofree | i tried upgrading a system that has a /boot and / partition, but using the old deploy cluster (with only ROOT_DEVICE set, incorrectly) | 11:50 |
| radiofree | it works... the boot menu doesn't get updated though | 11:50 |
| radiofree | probably should warn there? | 11:51 |
| ssam2 | where? | 11:52 |
| radiofree | system-version-manager | 11:52 |
| radiofree | if you wanted to upgrade the kernel it wouldn't work, because that wouldn't get copied over to /boot/ | 11:52 |
| radiofree | but you'd end up booting into the new rootfs because /systems/default will point to it on the btrfs partition | 11:53 |
| radiofree | i think erroring would be nicer than having something that "kinda worked" | 11:53 |
| ssam2 | yeah, if it can detect that the user hasn't given the right settings then it should do so | 11:55 |
| *** CTtpollard has joined #baserock | 12:00 | |
| radiofree | i suppose any deploy where ROOT_DEVICE != whatever / is should be treated as an error? | 12:02 |
| radiofree | i can't think why anyone would want to set that incorrectly | 12:03 |
| ssam2 | why do we make people specify ROOT_DEVICE if we can detect it ? | 12:03 |
| radiofree | i think it's only for morph + rawdisk | 12:03 |
| radiofree | i don't actually use ROOT_DEVICE in system-version-manager, i just think it would be a good way to check "you need to change your upgrade cluster now" | 12:03 |
| radiofree | i suppose i could just check for BOOT_DEVICE actually.... which i already do | 12:04 |
| radiofree | ssam2: for rawdisk morph defaults to /dev/sda (i think), so you need to set it there | 12:04 |
| radiofree | it's probably not needed for an upgrade cluster, i think it's better to leave it in though, for clarity | 12:04 |
| ssam2 | right | 12:06 |
| ssam2 | I think if we can make it optional maybe we should, but still let the user pass it, and check if the specified value is wrong | 12:08 |
| ssam2 | it'd definitely be confusing to say 'don't specify this for upgrades but do specify it for some other things' | 12:08 |
| kejiahu | Hi, a newbie question. what kernel parameter should be used to specify initilazation process when netboot baserock? | 12:34 |
| tiagogomes_ | kejiahu, `root=/dev/nfs ip=dhcp nfsroot=$rootfs_nfs_url` | 12:37 |
| kejiahu | tiagogomes_: thanks, but the rootfs was on a local partition. what I am missing is something like 'init='. I don't know what that should be for systemd | 12:38 |
| tiagogomes_ | you shouldn't need to change init= | 12:39 |
| kejiahu | tiagogomes_: but if I didn't specify anything, I got 'Kernel panic - not syncing: No working init found. Try passing init= option to kernel' | 12:40 |
| tiagogomes_ | root=/dev/sda then, or root=/dev/sda2 for moonshot | 12:40 |
| persia | On ROOT_DEVICE: if we can detect it is wrong, and we're checking it, why do we want the user to provide it? | 12:40 |
| kejiahu | tiagogomes_: actually I can boot a cross-bootstarap rootfs with 'root=/dev/sda2 inti=/tools/bin/sh', but for a proper baserock rootfs, I've no idea what it should be | 12:42 |
| tiagogomes_ | kejiahu, just remove the init line then | 12:43 |
| kejiahu | tiagogomes_: then the error: 'Kernel panic - not syncing: No working init found. Try passing init= option to kernel' | 12:43 |
| tiagogomes_ | kejiahu, mmm, is your /dev/sda2 a btrfs partition? Try try add this if so: try add this then `rootfstype=btrfs rootflags=subvol=systems/default/run` | 12:47 |
| kejiahu | tiagogomes_: that works, thanks | 12:57 |
| *** gary_perkins_ has joined #baserock | 13:11 | |
| *** gary_perkins has quit IRC | 13:11 | |
| *** bwh_ is now known as bwh | 13:13 | |
| *** fay_ has quit IRC | 14:05 | |
| *** lachlanmackenzie has joined #baserock | 14:07 | |
| *** fay_ has joined #baserock | 14:20 | |
| petefoth | I'm getting out of my depth with the --no-git-update bugs. morph branch doesn't do the right thing either. I've added a comment to the story, and I'm going to walk away :( | 14:38 |
| * paulsherwood is not entirely surprised... he walked away himself, too :) | 14:40 | |
| ssam2 | anyone used `git-review` before? | 14:42 |
| ssam2 | i'm trying it out for submitting a branch to Gerrit, but everything I try just tells me No changes between HEAD and gerrit/master. Submitting for review would be pointless.' | 14:43 |
| ssam2 | which is not true at all | 14:43 |
| petefoth | ssam2: I played with it a while ago (when we did some work on gerrit) but never used it in anger | 14:43 |
| SotK | ssam2: do you have Change-IDs in the commit messages? | 14:43 |
| ssam2 | no, I don't | 14:44 |
| mwilliams_ct | is w.b.o down for anyone else? | 14:44 |
| ssam2 | I'd hoped git-review would do that for me | 14:44 |
| SotK | maybe it does | 14:44 |
| * SotK has never used git-review | 14:44 | |
| De|ta | yes mwilliams_ct | 14:44 |
| ssam2 | it seems that it gives me that message because `git log --color=always --decorate --oneline HEAD --not --remotes=gerrit` returns no output | 14:44 |
| mwilliams_ct | De|ta: Groovy thanks for confirming | 14:44 |
| ssam2 | but I'm not clear what that command is doing or what it should be doing | 14:44 |
| ssam2 | wiki.baserock.org is hosted by Branchable so there's not much we can do about it being down, except ask Branchable what's going on | 14:45 |
| mwilliams_ct | branchable.com is also not loading here so I guess they're having problems | 14:45 |
| ssam2 | does anyone here speak git manpage? | 14:47 |
| ssam2 | the git log manpage has this for --remotes: | 14:47 |
| * SotK learns about git log --decorate | 14:47 | |
| ssam2 | "Pretend as if all the refs in refs/remotes are listed on the command line as <commit>. If <pattern> is given, limit remote-tracking branches to ones matching given shell glob. If pattern lacks ?, *, or [, /* at the end is implied." | 14:47 |
| SotK | I *assume* that command is showing all commits not from/on your remote called gerrit | 14:49 |
| ssam2 | hmm, I guess the branch I want to send for review *is* in gerrit... just not in 'master' | 14:49 |
| SotK | if I run that command in something without a gerrit remote I get a big log, if I run it with remote=origin I get nothing | 14:50 |
| ssam2 | maybe I need to patch git-review to pass `--not --remotes=gerrit/master` rather than `--not --remotes=gerrit` | 14:51 |
| ssam2 | thanks | 14:51 |
| ssam2 | heh, now it's trying to submit every single commit. I guess that wasn't right either | 14:52 |
| ssam2 | `git log gerrit/master..HEAD` makes it do what I want. | 14:54 |
| ssam2 | but it still didn't add a change-id. seems it'll only do that for me if I only have one new commit in my branch. | 14:59 |
| *** sambishop has quit IRC | 15:02 | |
| radiofree | i updated the flashing instructions for the jetson, feel free to test it! | 15:07 |
| ssam2 | woohoo | 15:08 |
| *** zoli__ has quit IRC | 15:08 | |
| *** zoli__ has joined #baserock | 15:08 | |
| *** zoli__ has quit IRC | 15:13 | |
| pedroalvarez | i think they were duplicated though | 15:16 |
| *** sambishop has joined #baserock | 15:16 | |
| radiofree | pedroalvarez: ? | 15:20 |
| radiofree | the instructions? | 15:20 |
| pedroalvarez | jetson board flashing instructions | 15:20 |
| radiofree | duplicated where? | 15:20 |
| pedroalvarez | e.g. http://wiki.baserock.org/guides/baserock-jetson-genivi/ | 15:20 |
| radiofree | oh :\ | 15:21 |
| radiofree | this seems a bit silly, but ok | 15:21 |
| ssam2 | i didn't realise gerrit forces each change to be a single commit ... that's pretty annoying | 15:22 |
| pedroalvarez | radiofree: hey, you don't have to fix that ;/ | 15:22 |
| pedroalvarez | radiofree: and I do appreciate the flashing guide update :) | 15:23 |
| petefoth | ssam2: can you group changes using 'topics'? | 15:24 |
| ssam2 | I don't know | 15:24 |
| ssam2 | this time, pushing a branch with 2 commits created two change requests that were linked by a topic | 15:35 |
| ssam2 | which is cool, except one of the commits already *had* a change request, and now there are two for the same thing | 15:35 |
| petefoth | ssam2: are you using the gerri ui or git review? | 15:36 |
| petefoth | I'm sure I recall gerrit allowing multiple commits in a single change | 15:36 |
| pedroalvarez | hmm.. anyone knows about selinux and how could we support it in baserock? | 15:40 |
| ssam2 | git-review | 15:40 |
| bashrc_ | is gerrit really all that useful? The UI looks rather klunky | 15:40 |
| petefoth | ssam2: ah OK | 15:40 |
| petefoth | bashrc_: the UI *is* cluncky, but gerrit's functionality is *very* useful IMHO | 15:41 |
| pedroalvarez | I agree | 15:41 |
| bashrc_ | can you use it from the command line? | 15:42 |
| Kinnison | Reasons why Baserock is *really important* number 14576: http://www.vitavonni.de/blog/201503/2015031201-the-sad-state-of-sysadmin-in-the-age-of-containers.html | 15:42 |
| Kinnison | We need to make that kind of crap better | 15:42 |
| petefoth | bashrc_: git review (which ssam2 is looking at) is a command line ui for gerrit I beleieve | 15:44 |
| bashrc_ | https://github.com/terranpro/magit-gerrit | 15:44 |
| * petefoth sees the word 'emacs' and walks away :) | 15:45 | |
| *** petefoth has quit IRC | 15:46 | |
| *** CTtpollard has quit IRC | 16:02 | |
| *** jonathanmaw has quit IRC | 16:16 | |
| bashrc_ | speaking of which, has anyone made a definition for emacs? | 16:17 |
| jmacs | It's been attempted before but not finished AIUI | 16:18 |
| jmacs | I imagine it depends on all the software | 16:18 |
| bashrc_ | yes I imagine | 16:19 |
| Kinnison | IIRC emacs "just works" when built, but it's a long build so it was never added to a standard system | 16:19 |
| jmacs | I'd very much like emacs in baserock, if anyone has the time | 16:19 |
| Kinnison | It's already lorried IIRC | 16:20 |
| *** zoli__ has joined #baserock | 16:24 | |
| jjardon | indeed: http://git.baserock.org/cgi-bin/cgit.cgi/delta/emacs.git/ , but seems there is something wrong with the repo: it doesnt show tags/latest changes | 16:24 |
| Kinnison | looks like it hasn't lorried for a while | 16:25 |
| pedroalvarez | because: error: could not lock config file .git/config: No such file or directory | 16:26 |
| pedroalvarez | given that is a git mirror, I'm tempted to remove its working area | 16:27 |
| Kinnison | pedroalvarez: that's a worrisome error | 16:27 |
| Kinnison | pedroalvarez: I wonder what ate the working area | 16:27 |
| pedroalvarez | I'm not sure if we are carrying these kind of errors since we had /home as btrfs | 16:28 |
| *** jonathanmaw has joined #baserock | 16:28 | |
| *** a1exhughe5 has quit IRC | 16:44 | |
| pedroalvarez | jjardon: emacs lorry fixed. Thanks for pointing that out | 16:50 |
| *** jonathanmaw has quit IRC | 17:03 | |
| jjardon | pedroalvarez: thank you for fix it! ;) | 17:13 |
| *** zoli__ has quit IRC | 17:14 | |
| *** mdunford has joined #baserock | 17:15 | |
| mdunford | Hi Baserock! Can anyone help me understand if Baserock will run given specific memory on the hardware? I've got some details... | 17:18 |
| ssam2 | specific memory... are you talking about a type of memory chip? | 17:20 |
| ssam2 | i don't know too much about hardware | 17:20 |
| mdunford | As in they have 4Gbit DDR2 RAM and 2Gbit Flash | 17:21 |
| jmacs | 4Gbit? Do you mean 4Gbyte? | 17:21 |
| mdunford | I was quoting. If Gbit makes zero sense it was probably a typo :) | 17:22 |
| jmacs | It's unusual, but not unheard of, to quote memory and storage sizes in bits | 17:24 |
| ssam2 | 2GB of flash sounds reasonable to store a Baserock rootfs of a small size | 17:24 |
| paulsherwood | mdunford: is this some kind of tiny device? | 17:24 |
| ssam2 | 2Gbit is about 256MB which is pretty limited (possible, but would need some work to shrink systems) | 17:24 |
| mdunford | paulsherwood: yes it will be a tiny device | 17:24 |
| ssam2 | 4GB RAM is fine to run Baserock, 4Gbit is 512MB I guess which again is small but not impossible, depending on what the device needs to do | 17:25 |
| pedroalvarez | yeah, don't expect that device to run a Trove | 17:27 |
| paulsherwood | heh | 17:27 |
| *** mdizzle has quit IRC | 17:36 | |
| *** franred has quit IRC | 17:42 | |
| mdunford | Thanks for the help. I'll try to confirm bit/byte | 17:46 |
| ssam2 | anyone want to be an early adopter tester of gerrit.baserock.org? | 17:47 |
| ssam2 | i'm still getting mirroring working, but would be good tomorrow to have a few people try to push changes through it and see all the places where the config I've created is broken | 17:48 |
| straycat | o/ | 17:48 |
| jjardon | ssam2: o/ | 17:49 |
| ssam2 | thanks! ok, let's have a temporary #baserock-gerrit channel to avoid spamming this one | 17:50 |
| ssam2 | (and to avoid highlighting the numerous mistakes i'll have made in public :) | 17:50 |
| bashrc_ | I've been mucking about with gerrit.baserock.org | 17:51 |
| jjardon | style question: do you prefer the copyritgth line being update in the commit that change that file or in the end of the patch series? (You can argue that a change in the copyrigth is not related with what the commit is changing) | 17:52 |
| * straycat tends to just squash it into the patch that changes the file | 17:52 | |
| straycat | while we're doing this can we start replacing (C) with © | 17:53 |
| straycat | you may make the second line # -*- coding: utf-8 -*- for python files | 17:53 |
| straycat | assuming the first line is a #! | 17:53 |
| jjardon | straycat: sure, but not sure you can type that in a baserock system ;) (no utf support) | 17:55 |
| straycat | ahh >.> | 17:58 |
| straycat | we should fix that | 17:58 |
| *** zoli__ has joined #baserock | 18:01 | |
| *** bashrc_ has quit IRC | 18:07 | |
| straycat | what was the url for the storyboard i lost a ton of browser stuff | 18:16 |
| jjardon | straycat: yeah, its the bug I noticed when start using baserock, I can not type my name to set up git! ;) unfortunately didnt have time to take a look to it https://storyboard.baserock.org/#!/story/4 | 18:17 |
| straycat | ahh cool thanks | 18:18 |
| *** ssam2 has quit IRC | 18:25 | |
| *** tiagogomes_ has quit IRC | 18:26 | |
| *** edcragg has quit IRC | 18:28 | |
| *** edcragg has joined #baserock | 18:29 | |
| *** lachlanmackenzie has quit IRC | 18:38 | |
| *** mdunford has quit IRC | 18:45 | |
| *** gary_perkins_ has quit IRC | 19:09 | |
| *** gfinney has quit IRC | 19:44 | |
| *** gfinney has joined #baserock | 19:45 | |
| *** rdale has quit IRC | 19:46 | |
| *** zoli__ has quit IRC | 19:55 | |
| *** zoli__ has joined #baserock | 19:57 | |
| *** mdunford has joined #baserock | 20:48 | |
| *** gfinney has quit IRC | 20:48 | |
| *** zoli__ has quit IRC | 20:56 | |
| *** franred has joined #baserock | 21:05 | |
| *** franred has quit IRC | 21:10 | |
| *** mdunford has quit IRC | 21:21 | |
| *** zoli__ has joined #baserock | 21:21 | |
| *** mdunford has joined #baserock | 21:40 | |
| *** mdunford has quit IRC | 22:11 | |
| *** edcragg has quit IRC | 22:18 | |
| richard_maw | \o/ first code in systemd: http://cgit.freedesktop.org/systemd/systemd/commit/?id=d422e52a3523ad0955bec4f9fbed46e234d28590 | 22:28 |
| richard_maw | it got merged 10 minutes after I left the office | 22:30 |
| jjardon | richard_maw: congrats! | 23:03 |
| * jjardon sends his first branch for review using gerrit | 23:38 | |
| *** zoli__ has quit IRC | 23:52 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!