| *** zoli__ has joined #baserock | 04:23 | |
| *** tpollard_ has joined #baserock | 05:14 | |
| *** doffm has joined #baserock | 05:14 | |
| *** perryl_ has joined #baserock | 05:18 | |
| *** tiagogomes_ has quit IRC | 05:26 | |
| *** CTtpollard has quit IRC | 05:26 | |
| *** perryl has quit IRC | 05:26 | |
| *** doffm_ has quit IRC | 05:26 | |
| *** zoli__ has quit IRC | 05:30 | |
| *** zoli__ has joined #baserock | 05:30 | |
| *** tiagogomes_ has joined #baserock | 05:30 | |
| *** zoli__ has quit IRC | 05:35 | |
| *** paulw has joined #baserock | 06:59 | |
| *** sherm_ has joined #baserock | 07:00 | |
| *** sherm__ has joined #baserock | 07:00 | |
| *** a1exhughe5 has joined #baserock | 07:00 | |
| *** sherm__ has quit IRC | 07:01 | |
| *** sherm_ has quit IRC | 07:01 | |
| *** sherm_ has joined #baserock | 07:01 | |
| *** petefoth_ has joined #baserock | 07:09 | |
| *** petefoth has quit IRC | 07:10 | |
| *** petefoth has joined #baserock | 07:12 | |
| *** Albert has joined #baserock | 07:23 | |
| *** Albert has quit IRC | 07:24 | |
| *** Albert has joined #baserock | 07:24 | |
| *** rdale has joined #baserock | 07:53 | |
| *** Albert has quit IRC | 07:54 | |
| *** mariaderidder has joined #baserock | 08:19 | |
| *** Albert has joined #baserock | 08:22 | |
| *** gary_perkins has joined #baserock | 08:38 | |
| *** zoli__ has joined #baserock | 08:43 | |
| *** edcragg has joined #baserock | 08:49 | |
| *** a1exhughe5 has quit IRC | 08:58 | |
| *** ssam2 has joined #baserock | 08:59 | |
| *** ChanServ sets mode: +v ssam2 | 08:59 | |
| *** Albert has quit IRC | 09:03 | |
| *** fay_ has quit IRC | 09:08 | |
| *** fay__ has joined #baserock | 09:08 | |
| *** Albert has joined #baserock | 09:10 | |
| *** a1exhughe5 has joined #baserock | 09:12 | |
| jmacs | The last error in the log from my Jetson YBD build is: [m4-tarball] ERROR: command failed in directory /src/tmp/staging/tmpLnsuXt | 09:13 |
|---|---|---|
| straycat | franred, yep not supported right now, but not too difficult to add | 09:18 |
| Kinnison | jmacs: Perhaps there'll be something slightly more useful *before* that error | 09:19 |
| jmacs | I couldn't say what is useful and what isn't | 09:20 |
| Kinnison | I suggest pastebinning as much of the log as you think you can to paste.baserock.org for paulsherwood then | 09:22 |
| Kinnison | Or uploading it somewhere if it's huge | 09:22 |
| jmacs | http://paste.baserock.org/uqulezuget is all I have in scrollback | 09:23 |
| Kinnison | Wow, that's confusing as all get out | 09:26 |
| * Kinnison thinks ybd's simulated distributed build needs some marshalling of results | 09:26 | |
| Kinnison | jmacs: Could you pull out and pastebin /src/cache/ybd-artifacts/m4-tarball@44b06679f0c7b52dc7e82745c2aabfb93719895b9215a55d34b77fb21974f4ea.build-log please? | 09:27 |
| jmacs | I'll try and figure out how to do that after this meeting | 09:30 |
| Kinnison | Cool | 09:30 |
| Kinnison | I think that'll be more use | 09:31 |
| nowster | Anyone else seen this gdb build failure? https://admin.codethink.co.uk/pnopaste/?2797 | 09:31 |
| nowster | (or similar0 | 09:31 |
| nowster | ) | 09:31 |
| Kinnison | looks like some kind of inlining failure | 09:32 |
| Kinnison | Or perhaps a linker hiccough | 09:32 |
| rdale | a system i've just built won't start because /systems/default/kernel is missing - is that a known problem? | 09:32 |
| nowster | yeah, that's what I thought. | 09:32 |
| nowster | inlining bug of some sort. | 09:32 |
| Kinnison | rdale: sounds like you built/deployed a system which doesn't have a kernel | 09:33 |
| Kinnison | rdale: In that the write extension should put any kernel available into place | 09:33 |
| Kinnison | nowster: In future, please don't use Codethink internal pastebins on a public channel though :) | 09:33 |
| nowster | sorry. | 09:33 |
| Kinnison | nowster: paste.baserock.org exists :) | 09:33 |
| nowster | I know. | 09:33 |
| SotK | rdale: what system did you build? | 09:34 |
| rdale | an openwrt system | 09:34 |
| rjek | paste.baserock.org annoys me because it randomly selects a highlighter and horizontally-wide things are obscured | 09:34 |
| rdale | maybe i've removed the kernel by mistake | 09:35 |
| jmacs | I can't figure out how to get a file into paste.baserock.org, but I suspect the last few lines will be all you need: http://paste.baserock.org/qemimutido | 09:37 |
| Kinnison | yep, ./config.status: line 1: awk: not found is pretty much it | 09:38 |
| Kinnison | implies that somehow the staging area is missing some of the build-deps | 09:38 |
| DavePage | Sounds awkward ;) | 09:38 |
| nowster | wtf do non development strata require strata/tools ? | 09:39 |
| pedroalvarez | rjek: if you know a better pastebin engine, I can give it a try | 09:45 |
| tiagogomes_ | nowster which non development strata? | 09:45 |
| nowster | lots of things import statrata/tool.morph | 09:46 |
| nowster | strata/tools.morph | 09:46 |
| SotK | I assume there is something in tools which they need | 09:47 |
| rjek | pedroalvarez: pastebin.com seems pretty good. | 09:47 |
| pedroalvarez | rjek: opensource? | 09:48 |
| rjek | pedroalvarez: Is there a need for it to be? | 09:48 |
| rjek | (Why do we need to run our own when there are hundreds of others already?) | 09:48 |
| pedroalvarez | I see | 09:49 |
| nowster | tiagogomes_: strata/apache-httpd-server.morph | 09:49 |
| pedroalvarez | then, if you don't like, use another :) | 09:49 |
| nowster | ...databases.morph | 09:49 |
| nowster | .... lighttpd-server.morph | 09:49 |
| nowster | ... lua.morph | 09:50 |
| nowster | ... erlang.morph | 09:50 |
| rdale | i noticed that with the erlang.morph - it builds against build-essential | 09:53 |
| franred | nowster, they may require some packages included in tools.morph to build | 09:53 |
| rjek | pedroalvarez: It's more that other people keep using it :) | 09:54 |
| pedroalvarez | rjek: hahah true :) | 09:54 |
| nowster | perhaps tools.morph should be split? Certainly gdb doesn't need to be made for those? | 09:54 |
| tiagogomes_ | looking at tools, the only one that may be needed is `file` | 09:54 |
| tiagogomes_ | I also think that `file` is essential and shouldn't be on tools | 09:55 |
| Kinnison | 'file' is an odd duck -- nominally it ought to not be build-depended on | 09:55 |
| Kinnison | But for some reason some things do | 09:56 |
| tiagogomes_ | some of chunks on those strata might need vala as well | 09:56 |
| franred | I wouldn't say that vala is a tool :S | 09:56 |
| nowster | vala more like what c++ was like 30 years ago | 10:07 |
| rjek | Dare anybody use the six-letter c-word? | 10:09 |
| * rjek shall: cfront! | 10:09 | |
| Kinnison | naughty | 10:10 |
| nowster | rjek: where are the fish and chip shops? :) | 10:11 |
| straycat | 2 things i'd like gerrit to do, 1) enforce dependency ordering in the 'related changes' list, 2) the ability to mark that you're reviewing something | 10:12 |
| *** Albert has quit IRC | 10:12 | |
| SotK | I thought it already did the 1st? | 10:12 |
| straycat | almost but not quite, apparently | 10:13 |
| * straycat invokes richard_maw | 10:13 | |
| nowster | straycat: where did you get the 4cc of mouse blood? | 10:13 |
| * richard_maw understood that it did it in change number, which is not the same if you insert a new change between two you previously had | 10:13 | |
| straycat | nowster, you ask a cat where he finds mouse blood? :) | 10:14 |
| SotK | I've seen it get really confused when stuff is updated without updating the whole topic, but its always been in dependency order for me I think | 10:14 |
| pdar | straycat: I think you can add yourself as a reviewer to show that you are reviewing something too? | 10:14 |
| SotK | perhaps I just ahven't noticed it | 10:15 |
| straycat | pdar, that's a good point | 10:15 |
| * straycat wonders whether we might adopt a convention that we *try* to wait for approval from all members of the review list | 10:17 | |
| straycat | that would avoid some random -1ing i ended up doing when I was testing things a while back | 10:17 |
| pedroalvarez | how would that work? | 10:19 |
| *** Albert has joined #baserock | 10:19 | |
| pedroalvarez | so, whenever you want to review something, you add yourself, and nothing can't be merged until all reviewers have a +1? | 10:19 |
| pedroalvarez | s/have/vote/ | 10:19 |
| straycat | generally, i don't think we could unconditionally adopt that policy because sometimes merging is urgent and people aren't always around | 10:19 |
| straycat | but it would avoid the problem of having something merged mid-review | 10:20 |
| * SotK wonders what happens when someone gives a quick review on v1 of something, then doesn't review v2, v3, v4 but is still in the reviewers list | 10:20 | |
| pedroalvarez | SotK: is not a bad Idea that someone that has -1ed a previous verison has to check if the reasons of the -1 were removed | 10:21 |
| pedroalvarez | but as straycat said, this will slow down the process | 10:22 |
| petefoth | I don't think we want to add any more friction in the review process | 10:22 |
| straycat | SotK, good point, i'm not sure, you could prod them for a vote but a lack of a response could actually result from them being mid-review | 10:22 |
| SotK | pedroalvarez: I agree with that, but I'm more concerned about someone giving a +1 at a glance before someone giving a deeper review finds a reason for a -1 | 10:22 |
| petefoth | SotK: is that something that has happened often enough to make it worth worrying about? | 10:23 |
| straycat | oh well, i guess this is why i wanted a way to indicate that i'm actively reviewing a change | 10:24 |
| * SotK has seen it a few times | 10:24 | |
| SotK | straycat: a "I am reviewing" button would be really useful | 10:25 |
| ssam2 | anyone know what '/var/lib/machines' is for? | 10:26 |
| petefoth | So we should discuss whether the benefits of fixing it outweigh the possible problems it would cause, make a decision based on the discussion, and add something appropriate to 'Policies' | 10:26 |
| ssam2 | and why I get 'operation not permitted' when I try to remove it? | 10:26 |
| ssam2 | this has broken `system-version-manager remove` | 10:26 |
| richard_maw | ssam2: it's a subvolume | 10:27 |
| ssam2 | ah | 10:27 |
| richard_maw | ssam2: hmm, `system-version-manager remove` will need to be recursive subvolume aware | 10:27 |
| *** radiofree has quit IRC | 10:27 | |
| straycat | SotK, yes it would be nice | 10:28 |
| ssam2 | richard_maw: seems that since it's in /var it shouldn't be being removed by system-version-manager at all | 10:29 |
| pedroalvarez | who creates that subvolume? :/ | 10:29 |
| ssam2 | this may be a symptom of the deeper problem we have that we mount the /var subvolume after systemd has already done a bunch of stuff in /var | 10:29 |
| ssam2 | (i'm guessing systemd creates it) | 10:29 |
| *** radiofree has joined #baserock | 10:30 | |
| richard_maw | pedroalvarez: systemd does | 10:31 |
| richard_maw | ssam2: hmm, good point, that would be a better fix | 10:32 |
| *** mariaderidder has quit IRC | 10:32 | |
| *** radiofree has quit IRC | 10:35 | |
| *** radiofree has joined #baserock | 10:37 | |
| franred | umm, what do we have in git.git ? http://git.baserock.org/cgi-bin/cgit.cgi/delta/git.git/tree/README.rdoc | 10:41 |
| richard_maw | franred: do you recall when we had a duplicate lorry file for a component, and the source repository key was accidentally left empty? | 10:42 |
| richard_maw | franred: well, the result was that the target repository expanded to delta/.git | 10:42 |
| richard_maw | franred: which the git server interpreted as delta/git.git | 10:42 |
| franred | richard_maw, I vaguely remember it, yes. so, we are still in process to fix this | 10:44 |
| pedroalvarez | we could add a check for empty keys | 10:45 |
| pedroalvarez | so that this doesn't happen again | 10:45 |
| richard_maw | we've gotten rid of the duplicate unit that did the push, but I'm not sure of the best way to clean it up | 10:45 |
| *** mariaderidder has joined #baserock | 10:45 | |
| franred | do we have a back up of the git stagin area of days/weeks before this issue? - even if we have a previous trove we can roll back to that point | 10:47 |
| pedroalvarez | also, to fix the master branch we could either change the lorry file to allow force pushing of that branch, or just remove it and wait for lorry-controller to lorry it again | 10:49 |
| franred | pedroalvarez, will that add the 4/5 baserock branches that we made in the past? | 10:52 |
| franred | - if we want to keep them for backwards compatibility/ reproducibility - | 10:52 |
| pedroalvarez | franred: sorry, I meant remove the master branch | 10:52 |
| pedroalvarez | I was aware of the baserock/ branches, although only one doesn't look a personal branch | 10:53 |
| franred | pedroalvarez, yep, you are right | 10:54 |
| petefoth | Seems you don't need an OpenID to edit the wiki. As described in http://wiki.baserock.org/editing-the-wiki/ you can use a 'local' login where you register with a username and password :) | 10:55 |
| pedroalvarez | I assume that removing the master branch is not a good idea though | 10:55 |
| pedroalvarez | petefoth: I'll try to remember that page whenever someone asks how to edit the wiki | 10:56 |
| *** Krin has quit IRC | 10:56 | |
| pedroalvarez | petefoth: and thanks! | 10:57 |
| petefoth | pedroalvarez: I'm gopign to tweak mny recent changes to the 'COntributing' page to link there | 10:57 |
| petefoth | s/going to/have just/ | 10:59 |
| * petefoth wishes we had a spec for what we want w.b.o to contain, instead of it being a container for 'whetever we think of at the time' | 11:03 | |
| Kinnison | Wikis always contain "whatever we thought of at the time" | 11:03 |
| Kinnison | that's part of their charm | 11:03 |
| DavePage | petefoth: Surely that's the point of a wiki :) | 11:03 |
| petefoth | DavePage:if that is the case then the main BAserock website should *not* be a wiki. | 11:04 |
| petefoth | It's like having a major suite of opensource tools and resources without any specifications. Oh wait.... | 11:05 |
| jmacs | The main page of w.b.o. is fairly well laid out, I think | 11:05 |
| jmacs | You could make a spec for some specific pages | 11:06 |
| petefoth | jmacs: agreed. Thank you for the suggestion. | 11:07 |
| paulsherwood | jmacs: thanks! | 12:08 |
| paulsherwood | Kinnison: i expect this is the make-depends-explicit problem, jmac is using an old set of definitions | 12:11 |
| paulsherwood | ie i think his checkout does not include http://git.baserock.org/cgi-bin/cgit.cgi/baserock/baserock/definitions.git/commit/?h=baserock/ps/explicit-build-depends-in-build-essential&id=6ec4900e35b1cfdeec28d37f652fd171abd42e78 | 12:12 |
| Kinnison | paulsherwood: aah | 12:17 |
| Kinnison | paulsherwood: I figured he'd have used a recent checkout after you told him he didn't need a workspace | 12:17 |
| * Kinnison spanks himself for assuming things | 12:17 | |
| paulsherwood | Kinnison: what did you mean by marshalling? | 12:18 |
| Kinnison | paulsherwood: the logs for the builds are interspersed | 12:18 |
| Kinnison | paulsherwood: they could do with being grouped up | 12:18 |
| paulsherwood | they all end up in the artifacts dir? | 12:18 |
| Kinnison | the console output | 12:19 |
| Kinnison | if you look at what jmacs pastebinned -- there's lots of confusing statements in there | 12:19 |
| Kinnison | it's very hard to discover what relates to what and how | 12:19 |
| jjardon | pedroalvarez: about the issue that bootstrap system can not deploy locally (as mkfs.btrfs is not on them). Maybe we should move btrfs-progs to core so it will be present in the cross- systems? | 12:20 |
| paulsherwood | i prefer it that way (i'm learning what the *-commands are by osmosis) but i take your point. anyway the actual log is stated to be in the artifact cache, ie...2015-05-12 19:07:36 [m4-tarball] ERROR: log file is at /src/cache/ybd-artifacts/m4-tarball@44b06679f0c7b52dc7e82745c2aabfb93719895b9215a55d34b77fb21974f4ea.build-log | 12:21 |
| pedroalvarez | jjardon: hm.. do we want to support all kind of deployments from the cross-bootstrap image? | 12:21 |
| Kinnison | yes, but scanning through the console output to decide if there's > 1 log worth examining was hard | 12:21 |
| paulsherwood | i need to make the context managers silent on error | 12:21 |
| Kinnison | paulsherwood: ^^ | 12:21 |
| paulsherwood | Kinnison: i'm aiming for | grep ERROR | 12:22 |
| pedroalvarez | jjardon: I think that we shouldn support that, and just suggest to build a build-system with that image and use that instead | 12:22 |
| paulsherwood | but again i take your point | 12:22 |
| Kinnison | paulsherwood: Grep for ERROR is only handy if you *expect* a failure | 12:22 |
| paulsherwood | i do :) | 12:22 |
| pedroalvarez | jjardon: note that the cross-boostrap image that you are using hasn't been built very cleanly | 12:23 |
| paulsherwood | anyway, i think i'm getting close to actually releasing ybd. | 12:24 |
| paulsherwood | i wish deploy was less of a can of worms though | 12:24 |
| pedroalvarez | jjardon: alternatively you could have cross-bootstrap-ed a build system | 12:24 |
| jjardon | pedroalvarez : ok, but that would require another step in the http://wiki.baserock.org/guides/how-to-cross-bootstrap/. And build systems are around ~170 components, that you will have to compile again | 12:26 |
| jjardon | ah! yeah, that would work I think. But whats the point of having cross-* systems then? | 12:27 |
| pedroalvarez | they have the minimal ammount of things to have morph to do builds | 12:27 |
| pedroalvarez | so that you face less problems with some chunks when building in some architectures, and you wait less the cross-bootstrap process to complete | 12:30 |
| jjardon | yeah, but for sure you want to deploy those system at some point; and seems that you have to build a build- system at that point anyway | 12:33 |
| paulsherwood | we need to simplify this | 12:34 |
| pedroalvarez | moving to a new architecture is not easy | 12:34 |
| paulsherwood | maybe all build-systems should do cross-bootstrap? :) | 12:34 |
| pedroalvarez | You can cross-bootstrap even a opentack system | 12:35 |
| Kinnison | the result of the second stage of cross-bootstrap is not "clean" | 12:35 |
| pedroalvarez | but it's going to take a lot of time, to end up with a snowflake | 12:35 |
| Kinnison | it was splodged together unpleasantly by cpio and badgers, stoats *AND* weasels | 12:35 |
| Kinnison | Use it to then *nicely* build a proper build-system | 12:35 |
| Kinnison | no more | 12:35 |
| Kinnison | FEAR THE WEASELS | 12:35 |
| * pedroalvarez concurs | 12:38 | |
| * jjardon only have 1 failure building a build system for armv5, better than what I was expecting! | 12:39 | |
| pedroalvarez | :) | 12:40 |
| pedroalvarez | once fixed that, deploy as a tarball, and enjoy your clean environment :) | 12:41 |
| pedroalvarez | and mkfs.btrfs, etc | 12:41 |
| pedroalvarez | wow, I've just started investigating why suds is not lorrying, and it doesn't appear in the lc-status page!! | 12:47 |
| * pedroalvarez is confused, but he will have to figure out himself | 12:47 | |
| Kinnison | bad .lorry file? | 12:48 |
| pedroalvarez | Kinnison: that should be it | 12:48 |
| pedroalvarez | python-packages.lorry with 712 lines :) | 12:49 |
| Kinnison | bleh | 12:51 |
| Kinnison | time to jsoncheck that sucker | 12:51 |
| * pedroalvarez checks the logs to see if there are more problems | 12:53 | |
| pedroalvarez | only that one, JSON error confirmed | 12:53 |
| pedroalvarez | found the comma | 12:58 |
| Kinnison | heh | 12:59 |
| pedroalvarez | https://gerrit.baserock.org/#/c/635/1/open-source-lorries/python-packages.lorry | 13:01 |
| pedroalvarez | maybe tiagogomes_ wants to review it ^ | 13:02 |
| franred | pedroalvarez, +2 | 13:03 |
| tiagogomes_ | pedroalvarez -1 | 13:03 |
| pedroalvarez | :/ | 13:03 |
| tiagogomes_ | I had put the comma there intentionally | 13:03 |
| pedroalvarez | :/ | 13:03 |
| * SotK once again feels the urge to finish his "yaml lorries" patch | 13:04 | |
| straycat | would be nice for sure | 13:04 |
| franred | SotK, +1 | 13:04 |
| tiagogomes_ | pedroalvarez is it breaking the lorries? I may have confused json with python | 13:04 |
| pedroalvarez | tiagogomes_: indeed, json is broken with that comma | 13:05 |
| tiagogomes_ | pedroalvarez, ops :) | 13:05 |
| tiagogomes_ | yeah, yaml for lorries please | 13:06 |
| * pedroalvarez submits | 13:06 | |
| pedroalvarez | thanks! :) | 13:06 |
| tiagogomes_ | pedroalvarez no, thank you | 13:07 |
| *** zoli__ has quit IRC | 13:16 | |
| *** zoli__ has joined #baserock | 13:17 | |
| * pedroalvarez continues investigating suds | 13:17 | |
| *** zoli__ has quit IRC | 13:21 | |
| * pedroalvarez joins #suds and finds only one person | 13:23 | |
| pedroalvarez | I hope he is one of the maintainers, otherwise I think I can't do anything else than lorry the tarball | 13:23 |
| rjek | Is it you? | 13:23 |
| pedroalvarez | rjek: hehe, no, we are 2 now | 13:23 |
| franred | pedroalvarez, have you seen https://bitbucket.org/jurko/suds ? | 13:27 |
| pedroalvarez | I'm wondering how did I miss it | 13:28 |
| pedroalvarez | because in their project page they point to another bitbucket repo | 13:29 |
| pedroalvarez | franred: thanks! having a look at it | 13:29 |
| franred | pedroalvarez, no probs :) | 13:29 |
| pedroalvarez | franred: looks like that's one of the 7ish forks of suds present in pypi | 13:32 |
| franred | pedroalvarez, pfff :S :/ have you check that Openstack does not have its own one? | 13:33 |
| pedroalvarez | It has to be this one (6166 downloads in the last day) https://pypi.python.org/pypi/suds/0.4 | 13:34 |
| pedroalvarez | "Uploaded on 2010-09-15" That makes me sad | 13:35 |
| pedroalvarez | ah! looks like the version 4.0 is in the repo, but not tagged :) | 13:40 |
| franred | pedroalvarez, have you checked if that version is in the svn repository? | 13:40 |
| franred | pedroalvarez, true, so maybe worth to point to the HEAD master? | 13:41 |
| *** edcragg has quit IRC | 13:47 | |
| *** zoli__ has joined #baserock | 13:48 | |
| *** sherm_ has quit IRC | 14:12 | |
| *** sherm_ has joined #baserock | 14:12 | |
| *** sherm_ has quit IRC | 14:13 | |
| pedroalvarez | hm.. does anybody know why we don't have urllib3 in baserock? | 14:18 |
| pedroalvarez | looks like we lorried it for openstack, but we don't have a definition for it? | 14:18 |
| ssam2 | I think it's part of 'requests' | 14:24 |
| pedroalvarez | ssam2: thanks! | 14:36 |
| pedroalvarez | ssam2: then, should i be able to import urllib3 in a system that has requests? | 14:37 |
| ssam2 | yes | 14:37 |
| pedroalvarez | urllib3 source it's indeed inside requests, but that import didn't work | 14:39 |
| ssam2 | hmm | 14:40 |
| ssam2 | 'import requests.packages.urllib3' | 14:41 |
| ssam2 | may be using that embedded copy isn't appropriate for what you are doing though | 14:41 |
| pedroalvarez | it's a runtime dependency (python requirement) for something else | 14:42 |
| *** petefoth_ has joined #baserock | 14:43 | |
| * pedroalvarez adds it | 14:43 | |
| pedroalvarez | thanks Sam! | 14:43 |
| *** petefoth has quit IRC | 14:43 | |
| *** petefoth_ is now known as petefoth | 14:43 | |
| richard_maw | ooh, systemd added support for overlayfs in nspawn: http://cgit.freedesktop.org/systemd/systemd/commit/?id=5a8af538aee99741ab492506b4284fdd26b597d1 | 14:45 |
| *** perryl_ is now known as perryl | 14:45 | |
| richard_maw | if we don't need to support chroots, we could replace most of our containerising code with calls to nspawn | 14:46 |
| pedroalvarez | but we do :/ | 14:49 |
| Kinnison | richard_maw: coo | 14:50 |
| * SotK grumbles as he forgets that Gerrit doesn't save half finished replies | 14:54 | |
| * Kinnison wonders how hard it'd be to write a tool which uses git review, git format-patch etc to send to the ML a patchset from a gerrit linked changeset | 14:56 | |
| SotK | Kinnison: probably really easy | 14:57 |
| Kinnison | I may do that then | 14:57 |
| Kinnison | because I find it really hard to review effectively in gerrit | 14:58 |
| SotK | I thought the rationale behind gerrit was to stop patches from being on the ML? | 14:58 |
| Kinnison | Sadly | 14:58 |
| Kinnison | well, part of it | 14:58 |
| SotK | ssam2: is there any chance you could answer richard_maw's comment on https://gerrit.baserock.org/#/c/630/ ? | 15:01 |
| pedroalvarez | The more I use gertty, the more I like it | 15:01 |
| pedroalvarez | everything is less messy | 15:02 |
| ssam2 | SotK: a slim chance, if I can figure out the answer :) | 15:03 |
| SotK | heh, thanks | 15:04 |
| pedroalvarez | meh, we are using python-prettytable from master | 15:32 |
| pedroalvarez | and I have this requirement PrettyTable>=0.7,<0.8 | 15:33 |
| pedroalvarez | franred: any reason to use master instead of a tag? | 15:34 |
| franred | pedroalvarez, none, just that 0.7 is 2 years old and master 10 months, so I assume that using master would be good - also I can not see 0.8 in the ctgit for pretty-table | 15:36 |
| pedroalvarez | nope, but I assume that master is going for a version greater than 0.7.2, that could be 0.8 | 15:37 |
| franred | pedroalvarez, did that requirement change from previous versions of the Openstack package you are updating? | 15:38 |
| pedroalvarez | franred: no, same requirement in the version of keystoneclient that we are using | 15:39 |
| franred | pedroalvarez, then I suggest you to keep the version we are using... I haven't seen any error with pretty-tables using master | 15:40 |
| pedroalvarez | well, that's relative | 15:41 |
| pedroalvarez | I've seen hundreds of errors | 15:41 |
| richard_maw | ssam2: biff re morph-distbuild script | 15:41 |
| pedroalvarez | franred: they may not be caused by this chunk, but I think doing that is not a good decision | 15:42 |
| pedroalvarez | others may disagree | 15:42 |
| *** a1exhughe5 has quit IRC | 15:44 | |
| franred | pedroalvarez, feel free to change it if you don't feel happy with master in that chunk - IMHO, if tags are pretty old and master has some commits which are bug fixes we should think to use master instead of the latest tag, but it is just my opinion | 15:49 |
| pedroalvarez | but, seriously, how do you know that the master is compatible? | 15:50 |
| * pedroalvarez points out again the requirement "PrettyTable>=0.7,<0.8" | 15:51 | |
| richard_maw | jjardon: please run the test suite before submitting changes | 15:51 |
| pdar | ssam2: i was looking at ypu patches for workspaceless baserocks. I saw your comment here https://gerrit.baserock.org/#/c/579/1 and wondered what subsytem deployments are. I dont know morph too well, are there any docs on it? | 15:52 |
| franred | pedroalvarez, prettyTable, does not have 0.8 release, only forks do - so you are safe using master | 15:52 |
| pedroalvarez | franred: how do we know? | 15:52 |
| franred | pedroalvarez, we have intensively tested | 15:53 |
| richard_maw | It looks like I'll have to review ssam2's https://gerrit.baserock.org/#/q/status:open+project:baserock/baserock/morph+branch:master+topic:sam/workspaces-optional-3 branch some other day | 15:53 |
| ssam2 | that's ok, i'm still finding issues with it myself ;) | 15:53 |
| pedroalvarez | franred: what kind of tests, and all of them passed? | 15:54 |
| *** petefoth has quit IRC | 15:54 | |
| ssam2 | pdar: I've checked `morph help deploy` and the wiki, and it doesn't seem there is any documentation on what the 'subsystem deployment' feature is | 15:54 |
| franred | pedroalvarez, if you want to use 0.7.2 change it - nothing else to comment about this | 15:54 |
| pedroalvarez | I take that as it wasn't really well tested | 15:55 |
| ssam2 | pdar: here's an example though: http://git.baserock.org/cgi-bin/cgit.cgi/baserock/baserock/definitions.git/tree/clusters/release.morph#n26 | 15:55 |
| *** petefoth has joined #baserock | 15:55 | |
| ssam2 | pdar: in that system the initramfs is effectively a 'subsystem': basically it's a blob of data created by Morph that goes inside another system | 15:55 |
| *** petefoth has quit IRC | 15:56 | |
| franred | pedroalvarez, in 9 months I haven't seen any error in prettytables non anyone else who reports error when visualizing tables in Openstack using master... so as I said before, it was intensively tested - after that you can use/think whatever you want | 15:58 |
| pdar | cool, thanks sam | 16:01 |
| *** Albert has quit IRC | 16:02 | |
| ssam2 | wow, I have a nasty feeling `morph deploy` is broken in master of morph, and also in the 15.19.1 release | 16:36 |
| SotK | broken in what way? :/ | 16:36 |
| Zara | :S | 16:36 |
| ssam2 | http://paste.baserock.org/ekemalegow (this is list-artifacts, but it's the same error I saw for deploy) | 16:36 |
| SotK | oh dear :( | 16:38 |
| ssam2 | I'm confused though, I would have thought ./check would catch this | 16:38 |
| SotK | it should do if its broken in deploy I'd imagine | 16:38 |
| SotK | afaik there are no tests for list-artifacts though | 16:38 |
| ssam2 | ah, I see, deploy does work | 16:39 |
| ssam2 | but in my branch deploy was broken because I'd changed how it worked | 16:39 |
| ssam2 | I think the problem is that if you call create_source_pool(filenames=[a,b]) then it's fine | 16:39 |
| * SotK sees the bug I think | 16:39 | |
| ssam2 | but if you call create_source_pool(filename=xx) then it's not | 16:39 |
| SotK | morphlib/sourceresolver.py, line 619 | 16:40 |
| SotK | "if filenames is None and filenames is not None" | 16:40 |
| Kinnison | Eww | 16:40 |
| SotK | s/filenames is not/filename is not/ | 16:40 |
| Kinnison | eww more | 16:40 |
| Kinnison | a function with 'filename' and 'filenames' as variable names :-( | 16:41 |
| SotK | (the substitution being the fix) | 16:41 |
| ssam2 | yes, it's not nice | 16:41 |
| Kinnison | a better fix would thoroughly rename one of those variables | 16:41 |
| ssam2 | a better fix would remove one of those variables, I think | 16:41 |
| ssam2 | in fact, I know for sure | 16:42 |
| SotK | +1 | 16:42 |
| *** mariaderidder has quit IRC | 16:42 | |
| ssam2 | Sotk: are you going to do one, shall I? | 16:42 |
| SotK | ssam2: I can send a quick one, but I was planning to leave in about 2 minutes | 16:44 |
| ssam2 | i'll do the proper fix then | 16:44 |
| SotK | thanks | 16:44 |
| Zara | Hi, I'm getting an error I'm not sure about: http://paste.baserock.org/zujawaniji . This is the system I'm trying to build: https://github.com/Zarathecat/baserock-openbmc/blob/crosstest/systems/openbmc-system.morph (we thought adding coreutils would fix the error, but it didn't) | 16:45 |
| pedroalvarez | Zara: groupadd is being built by shadow | 16:47 |
| pedroalvarez | adding core should solve the issue, but you might want to avoid that | 16:48 |
| pedroalvarez | Zara: also you can just remove ntpd stratum from your system | 16:49 |
| ssam2 | this error does seem to show up in ./check, weird | 16:49 |
| ssam2 | I'm sure I ran that before making the 15.19.1 tag | 16:49 |
| ssam2 | this is with commit eda5cb29d0200ee55c0ed50d2afe1a17dbd7577a of morph | 16:50 |
| ssam2 | I am quite dumb so probably ran ./check in a totally different checkout of Morph or something | 16:50 |
| Zara | pedroalvarez: adding core worked, thanks :) (we think we need ntpd in the system atm) | 16:50 |
| Zara | ssam2: It looks a bit like a very strange error me and mwilliams_ct were getting the other week, which was something to do with git HEAD (we never did work out exactly what was going wrong)-- but could be totally unrelated. | 16:52 |
| ssam2 | Zara: that error is quite common in Python, where some code expects a list but gets None instead | 16:53 |
| Zara | yeah, that just reminded me because it was very strange (a set of definitions that worked fine locally gave a build error that only occurred if we got the remote repo with morph checkout, not morph branch.) | 16:57 |
| *** jonathanmaw has quit IRC | 16:57 | |
| ssam2 | ouch | 17:01 |
| ssam2 | Fix for the issue mentioned above: https://gerrit.baserock.org/#/c/637/ | 17:01 |
| *** ssam2 has quit IRC | 17:02 | |
| *** radiofree has quit IRC | 17:05 | |
| *** radiofree has joined #baserock | 17:05 | |
| *** bashrc has quit IRC | 17:05 | |
| franred | ssam2, I've run the test for eda5cb29d0200ee55c0ed50d2afe1a17dbd7577a in morph and they passed :/ | 17:08 |
| *** rdale has quit IRC | 17:14 | |
| *** rdale has joined #baserock | 17:15 | |
| *** tiagogomes_ has quit IRC | 17:28 | |
| *** sambishop has quit IRC | 17:28 | |
| *** zoli__ has quit IRC | 17:36 | |
| *** sambishop has joined #baserock | 17:59 | |
| *** gary_perkins has quit IRC | 18:48 | |
| richard_maw | damn, sorry I introduced the filenames bug, I'm perplexed how it could have slipped through given I also ran ./check before submission | 19:08 |
| *** petefoth has joined #baserock | 19:26 | |
| *** petefoth has quit IRC | 19:36 | |
| *** rdale has quit IRC | 19:48 | |
| SotK | I guess the test suite doesn't actually test anything that would cause the error | 19:52 |
| SotK | yep, most stuff uses BuildCommand.create_source_pool, which avoided the bug | 19:53 |
| SotK | but certify, list-artifacts and generate-manifest-csv use sourceresolver.create_source_pool, which doesn't avoid the bug | 19:55 |
| SotK | and none of those three even have yarns | 19:55 |
| SotK | s/even // | 19:56 |
| * SotK disappears | 19:56 | |
| *** brlogger` has joined #baserock | 21:04 | |
| *** Kinnison_ has joined #baserock | 21:05 | |
| *** De|ta_ has joined #baserock | 21:05 | |
| *** brlogger has quit IRC | 21:09 | |
| *** Kinnison has quit IRC | 21:09 | |
| *** ratmice__ has quit IRC | 21:09 | |
| *** De|ta has quit IRC | 21:09 | |
| *** bjdooks has quit IRC | 21:09 | |
| *** ratmice__ has joined #baserock | 21:09 | |
| *** Kinnison_ has quit IRC | 21:09 | |
| *** Kinnison_ has joined #baserock | 21:09 | |
| *** bjdooks has joined #baserock | 21:12 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!