| *** zoli__ has joined #baserock | 00:02 | |
| *** zoli___ has joined #baserock | 00:02 | |
| *** zoli__ has quit IRC | 00:02 | |
| *** zoli___ has quit IRC | 00:02 | |
| *** zoli__ has joined #baserock | 00:03 | |
| *** zoli__ has quit IRC | 00:08 | |
| *** zoli__ has joined #baserock | 00:10 | |
| *** zoli__ has quit IRC | 00:10 | |
| *** zoli__ has joined #baserock | 00:11 | |
| *** zoli__ has quit IRC | 00:15 | |
| *** zoli__ has joined #baserock | 00:32 | |
| *** zoli__ has joined #baserock | 00:32 | |
| *** zoli__ has joined #baserock | 00:33 | |
| *** zoli__ has joined #baserock | 00:51 | |
| *** zoli__ has quit IRC | 00:55 | |
| *** zoli__ has joined #baserock | 01:06 | |
| *** radiofree has quit IRC | 01:16 | |
| *** radiofree has joined #baserock | 01:20 | |
| *** jmacs_ has joined #baserock | 01:24 | |
| *** perryl_ has joined #baserock | 01:28 | |
| *** perryl has quit IRC | 01:28 | |
| *** zoli__ has quit IRC | 01:31 | |
| *** jmacs has quit IRC | 01:35 | |
| *** tiagogomes_ has quit IRC | 01:35 | |
| *** sambishop has quit IRC | 01:35 | |
| *** doffm has quit IRC | 01:35 | |
| *** SotK has quit IRC | 01:35 | |
| *** benbrown_ has quit IRC | 01:35 | |
| *** straycat has quit IRC | 01:35 | |
| *** Kinnison has quit IRC | 01:35 | |
| *** tiagogomes_ has joined #baserock | 01:39 | |
| *** sambishop has joined #baserock | 01:39 | |
| *** benbrown_ has joined #baserock | 01:39 | |
| *** doffm has joined #baserock | 01:39 | |
| *** SotK has joined #baserock | 01:39 | |
| *** straycat has joined #baserock | 01:39 | |
| *** Kinnison has joined #baserock | 01:39 | |
| *** tiagogomes_ has quit IRC | 01:40 | |
| *** sambishop has quit IRC | 01:40 | |
| *** doffm has quit IRC | 01:40 | |
| *** SotK has quit IRC | 01:40 | |
| *** benbrown_ has quit IRC | 01:40 | |
| *** straycat has quit IRC | 01:40 | |
| *** Kinnison has quit IRC | 01:40 | |
| *** tiagogomes_ has joined #baserock | 01:41 | |
| *** sambishop has joined #baserock | 01:41 | |
| *** benbrown_ has joined #baserock | 01:41 | |
| *** doffm has joined #baserock | 01:41 | |
| *** SotK has joined #baserock | 01:41 | |
| *** straycat has joined #baserock | 01:41 | |
| *** Kinnison has joined #baserock | 01:41 | |
| *** persia_ has joined #baserock | 03:36 | |
| *** zoli__ has joined #baserock | 04:39 | |
| *** zoli___ has joined #baserock | 05:13 | |
| *** zoli__ has quit IRC | 05:13 | |
| *** zoli___ has quit IRC | 06:21 | |
| *** zoli__ has joined #baserock | 06:21 | |
| *** sherm_ has joined #baserock | 07:18 | |
| *** a1exhughe5 has joined #baserock | 08:05 | |
| SotK | hmm, so I'm running a sysroot deployment with my OSTree branch and getting loads of complaints from mv ("Invalid argument" and "Directory not empty") | 08:34 |
|---|---|---|
| *** gfinney has joined #baserock | 08:34 | |
| *** gfinney_ has joined #baserock | 08:34 | |
| SotK | Its also slow, but that is probably due to IO issues on the box my VM is running on | 08:34 |
| *** gfinney_ has quit IRC | 08:35 | |
| *** gfinney has quit IRC | 08:35 | |
| *** gfinney has joined #baserock | 08:35 | |
| *** gfinney_ has joined #baserock | 08:35 | |
| *** mdizzle has joined #baserock | 08:42 | |
| richard_maw | directory not empty is probably from calls to rmdir | 09:02 |
| richard_maw | I can't think of any reason why you'd get EINVAL | 09:03 |
| richard_maw | try running the mv commands through strace, or use stuff from shutil? | 09:03 |
| richard_maw | at least then you ought to get better error messages | 09:03 |
| *** bashrc has joined #baserock | 09:06 | |
| *** rdale has joined #baserock | 09:07 | |
| *** jonathanmaw has joined #baserock | 09:10 | |
| *** gary_perkins has joined #baserock | 09:26 | |
| *** ssam2 has joined #baserock | 10:01 | |
| *** ChanServ sets mode: +v ssam2 | 10:01 | |
| *** zoli__ has quit IRC | 10:05 | |
| *** zoli__ has joined #baserock | 10:05 | |
| mauricemoss_ | I was running the native-bootstrap again with the latest Morph on MIPS for verification and got this: http://paste.baserock.org/bopihexawo Should git://git.baserock.org/delta/python-packages/pylru be added to cross-bootstrap.morph? It finished with the old Morph without errors. | 10:12 |
| ssam2 | mauricemoss_: new morph does indeed need pylry | 10:13 |
| ssam2 | pylru | 10:13 |
| *** lachlanmackenzie has joined #baserock | 10:14 | |
| ssam2 | mauricemoss_: however, it seems to me that strata/cross-bootstrap.morph is duplicating a whole lot of strata/morph-units.morph | 10:15 |
| ssam2 | I think that was done in the past when 'morph' was in the 'tools' stratum and we wanted to avoid building the whole of 'tools' during a cross-bootstrap | 10:16 |
| ssam2 | but now there's a morph-utils stratum you could just include it in the system you're bootstrapping, and remove all the duplicated chunks from cross-bootstrap.morph | 10:16 |
| mauricemoss_ | ok ssam2, I'll have a look. | 10:17 |
| tiagogomes_ | that needs to be done for every croosboot-strap system | 10:18 |
| tiagogomes_ | currently probably they will fail due not having 'six' anymore in the cross-bootstrap systems, even if the morph's ref doesn't include pulry | 10:19 |
| mauricemoss_ | I think this should fix the issue. I missing the big picture in order to say how to move stratums, though. | 10:20 |
| mauricemoss_ | http://paste.baserock.org/otobohavas | 10:20 |
| ssam2 | mauricemoss_: that looks like it'll fix the issue indeed | 10:20 |
| ssam2 | to get an idea of the bigger picture, perhaps compare the list of chunks in morph-utils.morph with the chunks in cross-bootstrap.morph and see how many of them overlap | 10:21 |
| franred | tiagogomes_, six is in python-core so you can add that stratum to the croosboot-strap dependencies | 10:22 |
| *** lachlanmackenzie has quit IRC | 10:23 | |
| pedroalvarez | mauricemoss_: morph nesss pylru as a 'build-depends' | 10:23 |
| pedroalvarez | nesss? need | 10:23 |
| pedroalvarez | +s | 10:24 |
| mauricemoss_ | true, ta pedroalvarez | 10:24 |
| ssam2 | adding it as a build dependency is good, it also needs to be added to the list of strata in the actual cross-bootstrap system so that it's present at runtime | 10:27 |
| ssam2 | same with morph-utils | 10:27 |
| *** lachlanmackenzie has joined #baserock | 10:37 | |
| mauricemoss_ | flup, bottle and git-fat are missing in the cross-bootstrap.morph | 10:41 |
| ssam2 | right. so everything from morph-utils is duplicated in cross-bootstrap except for 3 (fairly small) components | 10:42 |
| radiofree | rdale: morning! | 10:48 |
| radiofree | are sinatra, gabba things you can import with the ruby gem importer? | 10:48 |
| rdale | morning | 10:48 |
| rdale | yes if gabba is a gem | 10:49 |
| ssam2 | sinatra you can, there are even instructions | 10:49 |
| radiofree | great, thanks | 10:49 |
| rdale | is that 'gabba, gabba hey!'? | 10:49 |
| radiofree | it must be named after that | 10:50 |
| radiofree | was sinatra developed by frank? | 10:50 |
| DavePage | "Frank Sinatra (chairman of the board) for having so much class he deserves a web-framework named after him." http://www.sinatrarb.com/about.html | 10:51 |
| radiofree | hah | 10:52 |
| radiofree | small question about openstack, do we have a baserock system with Keystone, Nova and Glance? | 11:02 |
| radiofree | and Horizon? | 11:02 |
| pedroalvarez | radiofree: without neutron? | 11:03 |
| pedroalvarez | and why horizon? | 11:04 |
| radiofree | network is optional | 11:04 |
| radiofree | neutron = networking right? | 11:04 |
| straycat | yeah | 11:05 |
| radiofree | i'm looking at cloud foundry, there guides give deployment instructions using an openstack web gui, which i've been told was horizon? | 11:05 |
| radiofree | s/there/their | 11:05 |
| franred | radiofree, yes, the openstack web ui is horizon | 11:07 |
| franred | radiofree, is network optional or they are using the old nova-networking? | 11:07 |
| radiofree | franred: i suppose it's possible to configure stuff without using that though? | 11:07 |
| franred | radiofree, yes, you don't need horizon to configure anything | 11:08 |
| radiofree | it's a bit confusing, "(Optional) OpenStack Networking: Provides network scaling and automated management functions that are useful when deploying complex distributed systems. " | 11:08 |
| radiofree | as an openstack service | 11:08 |
| radiofree | however the next step then says "The following OpenStack networks:" | 11:09 |
| radiofree | 'An external network with a subnet'... | 11:09 |
| mauricemoss_ | ssam2, I think this fix should be enough: http://paste.baserock.org/awuvususuc Why should Morph be present at runtime when running './native-bootstrap'? | 11:09 |
| radiofree | so we have neutron in there anyway? | 11:09 |
| franred | radiofree, so you either need neutron or nova-networking | 11:09 |
| radiofree | which is the one in either Havana, Icehouse or Juno? | 11:09 |
| radiofree | (which release do we have in baserock?) | 11:10 |
| franred | radiofree, we are using Juno, tag 2014.2.1 | 11:10 |
| franred | and we are using neutron to configure the networking | 11:10 |
| radiofree | ok, so i guess we meet all the requirements then | 11:10 |
| franred | we also expect to have horizon soon..ish | 11:10 |
| radiofree | wonderful! | 11:11 |
| radiofree | do we have cinder? | 11:12 |
| franred | yes, cinder is integrated too | 11:12 |
| franred | radiofree, but we are at the moment trying to stabilize openstack system | 11:13 |
| radiofree | ok thanks for the help | 11:14 |
| pedroalvarez | mauricemoss_: Morph doesn't need to be present at runtime AFAIK | 11:16 |
| pedroalvarez | mauricemoss_: adding Morph (and its dependencies) to the cross-bootstrap.morph system, will make ./native-bootstrap to build it, so you end up with a system with Morph installed | 11:16 |
| mauricemoss_ | ok, that's what I thought. I'd rather avoid building everything from scratch because it will take around ~27h for MIPS and it finished without an error with the old Morph. | 11:18 |
| pedroalvarez | mauricemoss_: It shouldn't rebuild everything, but I agree. You can try to install manually pylru, remove everything that has been installed from ./native-bootstrap script, and continue | 11:22 |
| pedroalvarez | s/continue/run again .\/native-boostrap/ | 11:22 |
| pedroalvarez | same for bottle and flup if they are needed | 11:23 |
| pedroalvarez | mauricemoss_: btw, typo in your latest paste: s/bootle/bottle/g | 11:23 |
| mauricemoss_ | ta | 11:28 |
| pedroalvarez | mauricemoss_: let me know if you need more help with this. I've had a lot of fun with cross-bootstrap in the past | 11:33 |
| ssam2 | mauricemoss_: sorry, I think I've not been clear | 11:37 |
| ssam2 | bottle, and flup aren't needed during cross-bootstrap at all | 11:37 |
| ssam2 | my point was that duplication is bad, and the cross-bootstrap stratum almost completely duplicates the morph-utils stratum | 11:38 |
| ssam2 | so my suggestion was that if you had time to clean up the crossbootstrap process, you could remove everything from cross-boostrap that is already in morph-utils, then include morph-utils in your system | 11:39 |
| ssam2 | if not, fine. You don't need botle and flup, anyway, they are just used by morph-cache-server which isn't involved in local building | 11:39 |
| *** fay_ has quit IRC | 11:45 | |
| *** fay_ has joined #baserock | 11:45 | |
| mauricemoss_ | ok, I see. Isn't the idea of the cross-bootstrap.morph to keep the chunks to a minimum? So including morph-utils in the system would bring in chunks that aren't needed at this point? | 11:48 |
| pedroalvarez | mauricemoss_: yes, this was the idea when all the things in morph-utils were in tools.morph. Now we have morph-utils stratum which has the essentials to run morph. | 11:50 |
| *** lachlanmackenzie has quit IRC | 11:54 | |
| *** jonathanmaw has quit IRC | 12:08 | |
| *** lachlanmackenzie has joined #baserock | 12:08 | |
| ssam2 | mauricemoss_: yes, but as you identified, it would only bring in 3 extra chunks | 12:15 |
| ssam2 | and they are all pretty small | 12:15 |
| mauricemoss_ | I'll prepare a patch, but independent from mips64b. | 12:16 |
| nowster | Is there any reason *not* to have /src as ext4? | 12:34 |
| *** jonathanmaw has joined #baserock | 12:35 | |
| pedroalvarez | nowster: no, and we have change the instructions recently: http://wiki.baserock.org/quick-start/#index3h2 | 12:36 |
| nowster | ah, good! | 12:36 |
| ssam2 | seems we hit some weird morph bug on http://mason-x86-64.baserock.org/ but it's building stuff now | 12:58 |
| pedroalvarez | how odd | 13:00 |
| pedroalvarez | :S | 13:00 |
| radiofree | is it possible to view historic logs from mason? or do they get cleaned out | 13:01 |
| perryl_ | ssam2: i've pushed changes for the distbuild protocol versioning if you want to give them a quick once-over | 13:08 |
| pedroalvarez | they don't get cleaned out, but if mason is redeployed we lost them | 13:08 |
| paulsherwood | is w.b.o up for anyone? | 13:11 |
| jmacs_ | Yes | 13:12 |
| perryl_ | paulsherwood: yes | 13:12 |
| paulsherwood | odd | 13:12 |
| paulsherwood | and now it is for me, too | 13:12 |
| *** jmacs_ is now known as jmacs | 13:12 | |
| *** perryl_ is now known as perryl | 13:12 | |
| *** fay_ has quit IRC | 13:13 | |
| ssam2 | perryl_: ta, will have a look and merge it | 13:13 |
| *** zoli__ has quit IRC | 13:14 | |
| *** fay_ has joined #baserock | 13:18 | |
| * paulsherwood has somehow broken network on his br vm.... | 13:21 | |
| paulsherwood | anyone have suggestions on how to troubleshoot? http://paste.baserock.org/puselafihe | 13:23 |
| jmacs | paulsherwood: Are you in Ducie house? | 13:23 |
| paulsherwood | i'm in Germany | 13:24 |
| jmacs | Ah. That ifconfig looks fine then. | 13:25 |
| richard_maw | paulsherwood: you can stop it renaming the devices by the instructions at the end of http://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames/ | 13:25 |
| richard_maw | paulsherwood: but that probably isn't causing your network problems itself | 13:25 |
| richard_maw | paulsherwood: can you paste the output of `ip route`? | 13:25 |
| paulsherwood | http://paste.baserock.org/mesejomafa | 13:26 |
| paulsherwood | afaik i haven't *done* anything to network lately.... but i have upgraded my devel several times | 13:27 |
| * paulsherwood left factory behind a year ago | 13:27 | |
| richard_maw | I can't see anything obviously wrong there, as I don't know the network layout where you are. Can you ping 10.7.255.251 ? | 13:29 |
| paulsherwood | no, it hangs, times out | 13:29 |
| paulsherwood | (from baserock vm) - and http://paste.baserock.org/ciwukeropo from host | 13:30 |
| richard_maw | I've never seen "Communication prohibited by filter" before | 13:31 |
| paulsherwood | hmmm... i did update virtualbox recently. maybe that's it | 13:32 |
| jjardon | Hi, any idea why ca-certificates is needed in the core stratum? the git history doesnt say much | 13:33 |
| richard_maw | paulsherwood: it looks like ping is failing on your host to me, which implies some firewalling to me, but it might just be for ICMP packets | 13:33 |
| petefoth | paulsherwood: are you getting your internet connection via public WiFi, or some other way that might be imposing a filter? | 13:34 |
| paulsherwood | jjardon: i originally needed it for go/docker, and curl | 13:34 |
| paulsherwood | others also came to need it | 13:34 |
| *** jonathanmaw has quit IRC | 13:35 | |
| paulsherwood | petefoth, richard_maw - yes, it may be the specifics of my situation. sadly i fear i won't know for sure til i return to uk | 13:35 |
| * petefoth remembers a wifi network at a customer site which blocked almost eery network operation | 13:35 | |
| petefoth | paulsherwood: can you try connecting vie your phone? | 13:35 |
| petefoth | s/vie/via/ | 13:35 |
| paulsherwood | petefoth: nope. 3 has decided to outlaw tethering abroad | 13:36 |
| petefoth | OK - I'm out of ideas. Sorry | 13:37 |
| * paulsherwood is delighted by that, obviously, especially given how much he spends on his phone every month | 13:37 | |
| jjardon | paulsherwood: doesnt seems to fit the criteria to be on core: "Core components of a Baserock base system that are build tools required for the BSP and Foundation strata, but a step above build-essential" Maybe it would be better to move it to foundation? | 13:37 |
| paulsherwood | petefoth: thanks anyway :) | 13:37 |
| *** jonathanmaw has joined #baserock | 13:37 | |
| paulsherwood | jjardon: would you move curl there too? | 13:37 |
| petefoth | paulsherwood: I have a 3 APYG MObile Braoidaband SIM which specifically does allo tethering. USed it in my (dual SIM) phone while abrad earlier this month. Might be worth considering that (plus a mifi) for future reference | 13:38 |
| paulsherwood | petefoth: i'll do that. and scream at 3 in passing :) | 13:38 |
| jjardon | paulsherwood: dont think so, git depends on curl | 13:39 |
| paulsherwood | jjardon: so curl requires ca-certs, if it's accessing https sites for example | 13:39 |
| jjardon | paulsherwood: is it a build or a runtime dependency? | 13:39 |
| paulsherwood | runtime | 13:39 |
| persia | I believe ca-certificates is core because anything that uses ssl wants it. | 13:39 |
| petefoth | paulsherwood: do they actually *enfore* the prohibition? My T-Mobile SIM doesn't allow tethering in it's T&Cs, but it does actually work, at least vai the protable WiFi hotspot feature in Adnroid, in emergencies | 13:40 |
| persia | So we decided to put it next to ssl | 13:40 |
| * petefoth gets back to work | 13:40 | |
| jjardon | paulsherwood: ah, you solve that in your system definitions morph file, not need to put as a build dependency | 13:40 |
| *** jonathanmaw has quit IRC | 13:42 | |
| paulsherwood | jjardon: yes, true. no need to have it in build-depends: for curl. but i still think it's required in the same stratum as curl | 13:42 |
| *** jonathanmaw has joined #baserock | 13:42 | |
| radiofree | shall we add --disable-tests to the systemd configure options? | 13:46 |
| Kinnison | I'd rather that we run the tests | 13:46 |
| radiofree | do we run them? | 13:46 |
| persia | Wouldn't that mean we stopped testing systemd at build-time? | 13:46 |
| paulsherwood | this is use-case specific | 13:46 |
| Kinnison | But I appreciate that some people don't like that since it causes extra artifact-build time cost | 13:46 |
| persia | paulsherwood: How so? When do you want to label something correctly built when it isn't tested? | 13:46 |
| paulsherwood | if *i'm* building, i want speed. others may want tests | 13:46 |
| radiofree | i've spent 5+ building tests that i assume will never get run? | 13:47 |
| persia | Are they not run during the build? | 13:47 |
| Kinnison | radiofree: If systemd's "make" builds tests then that seems odd | 13:47 |
| Kinnison | radiofree: unless it also runs them | 13:47 |
| paulsherwood | persia: this leads to a debate about what is 'tested' | 13:47 |
| Kinnison | normally tests are only built during 'make check' | 13:47 |
| radiofree | will tell you if they're run after it's finished building | 13:48 |
| persia | paulsherwood: I think the appropriate place for that debate is in each upstream project, individually, and that we should not impose an ideology. | 13:48 |
| paulsherwood | i favour system level tests, rather than component tests | 13:48 |
| paulsherwood | persia: i'm not imposing anything. i'm just saying, that something passes its unit tests does not convince me that it's working as a whole | 13:48 |
| Kinnison | I favour testing at every possible level, but appreciate that for integration work, system level tests are probably of more use. | 13:49 |
| persia | I also like system level tests, but if someone has a test to detect successful miscompile, this is useful feedback to get *before* I go build, deploy, and do my system testing. | 13:49 |
| rjek | Although systemd already ships with a load of tests. | 13:49 |
| rjek | Recreating them to run on a built system is a lot of work | 13:49 |
| persia | The key question, to which I've not seen an answer, is "Do we run them?" | 13:49 |
| radiofree | looks like you have to run make check to actually run the tests | 13:49 |
| Kinnison | radiofree: if it's building tests we're not running, then the question is 'do we add make check or do we --disable-tests' | 13:50 |
| persia | Building them and not running them is an ecologically unsound strategy | 13:50 |
| paulsherwood | heh | 13:50 |
| pedroalvarez | jjardon: curl needs ca-certificates at build time. | 13:50 |
| paulsherwood | pedroalvarez: well it can build without, it just wont work with ssl? | 13:51 |
| persia | Which then breaks git, which breaks lorry, which breaks workflow | 13:51 |
| pedroalvarez | paulsherwood: ye | 13:51 |
| paulsherwood | ok, i propose -1 for moving ca-certificates | 13:51 |
| pedroalvarez | Indeed, installing ca-certificates later doesn't fix curl | 13:52 |
| * SotK tests `morph gc` with his OSTree artifact cache and realises with mild despair that its worked and he needs to redownload all his artifacts now | 13:53 | |
| paulsherwood | i'm +1 for adding make check to systemd if others agree, but note that it would be *better* to support both usecases | 13:53 |
| persia | radiofree: If we aren't running the tests, my vote would be for --disable-tests, along with a shell comment in the YAML indicating that to enable them one has to add `make check` to the build commands. | 13:53 |
| paulsherwood | it would be nice if we could do build --and-test | 13:54 |
| persia | That requires structural changes to definitions. Isn't there a group defining the requirements for the next definitions format? They could probably use that sort of suggestion. | 13:55 |
| Kinnison | I am concerned about ensuring equivalence of artifacts built with and without testing enabled | 13:55 |
| paulsherwood | Kinnison: they would not be equivalent, i believe | 13:55 |
| Kinnison | In that case I think the default should be to build with testing | 13:57 |
| Kinnison | and you should be able to do --no-testing | 13:57 |
| Kinnison | to use non-tested artifacts | 13:57 |
| paulsherwood | ok by me | 13:57 |
| pedroalvarez | And if you do that, you have to rebuild everything? | 13:57 |
| Kinnison | yep | 13:58 |
| Kinnison | unless there's published variants of the artifacts | 13:58 |
| Kinnison | which if the upstreams are sane, and we're using sotk's OSTree cache should be minimal impact for us to publish | 13:58 |
| radiofree | should we be doing this for everything then? | 13:58 |
| radiofree | e.g qtbase has loads of test, but we don't build them | 13:59 |
| Kinnison | persia: since you're aware of a group defining requirements for the next definitions format, perhaps you can suggest to them that enabling/disabling tests should be part of that -- perhaps a "flag" if they're going down that route mentioned at the meetup? | 13:59 |
| Kinnison | radiofree: at the point that we can turn them on and off, I think the answer to your question is 'yes' | 13:59 |
| persia | Kinnison: To make sure I understand, you are suggesting adding `make check` in the systemd case, and generally trying to enable testing everywhere? | 13:59 |
| Kinnison | persia: I would advocate that only when we are able to control the tests, to allow people such as paulsherwood to disable them | 13:59 |
| persia | Kinnison: My only awareness on the subject is based on my memory of Meetup summaries. | 14:00 |
| Kinnison | persia: aah :( | 14:00 |
| * Kinnison has been unwell since the meetup and hasn't been in much contact with the community :( | 14:00 | |
| * Kinnison shall endeavour to find out who is looking at the definitions thing and then pass on the requirements | 14:00 | |
| * paulsherwood was in said group, and wrote up some stuff, to which richard_maw replied. this could be a useful further reply | 14:00 | |
| persia | Aha! So for now, --disable-tests, and with new-shiny-definitions, `make check`? | 14:00 |
| Kinnison | paulsherwood: aah on the ML? I'm afraid I have 1000+ mails on there :( | 14:00 |
| Kinnison | paulsherwood: left to read | 14:00 |
| Kinnison | persia: aye | 14:00 |
| * pedroalvarez wants speed by default | 14:01 | |
| persia | Works for me. | 14:01 |
| persia | pedroalvarez: Have you considered oil-cooled build servers? | 14:01 |
| paulsherwood | Kinnison: http://wiki.baserock.org/meetup/declarative-definitions/ | 14:01 |
| paulsherwood | i'll add this discussion to the bottom of that page | 14:02 |
| Kinnison | paulsherwood: awesomesauce | 14:03 |
| pedroalvarez | I didn't know people was using oil to cool down computeres | 14:03 |
| pedroalvarez | s/res/rs/ | 14:03 |
| jjardon | paulsherwood: pedroalvarez oh, rigth. Nevermind then | 14:03 |
| Kinnison | pedroalvarez: oil-based cooling systems are messier but safer when they explode vs. water-based ones | 14:04 |
| Kinnison | and some oil-based cooling systems just submerge the entire system | 14:04 |
| Kinnison | and then circulate the oil like a huge fish tank | 14:04 |
| pedroalvarez | just found a picture of someone frying chips on its oil-based cooling system | 14:05 |
| radiofree | so systemd didn't run the tests by default | 14:07 |
| Kinnison | I vote +2 on: 14:00 < persia> Aha! So for now, --disable-tests, and with new-shiny-definitions, `make check`? | 14:08 |
| persia | Since that was just a restatement of what I understood of your proposal, I suspect you need another voice (but I'll +1 to be that voice) | 14:09 |
| * paulsherwood notes that his wiki update, will magically link to this conversation :-) | 14:09 | |
| persia | It's not magic: it's the superpowers of the ops team. | 14:09 |
| paulsherwood | +1 | 14:09 |
| pedroalvarez | :D | 14:09 |
| jjardon | paulsherwood: Kinnison persia : FYI, there are patches in review for the first step: add support to morph to detect different versions of the format of definitions (basically check the VERSION file) | 14:11 |
| Kinnison | cool | 14:11 |
| jjardon | -1 to run tests by default: we are not developers, but integrators; we normally use stable versions of any package; they normally have passed the test to be released | 14:14 |
| pedroalvarez | jjardon: if you were a gnome developer and you were using baserock to build a gnome system, would you like to run tests when building? | 14:16 |
| jjardon | pedroalvarez: if Im a GNOME developer and I use baserock to develop the component Im in charge of, sure. If Im using baserock to build the whole system, I dont think so | 14:18 |
| jjardon | in the end all depends if you have the time/knowledge to fix any problem the tests detect | 14:19 |
| petefoth | So both use cases are valid - the decision is which should be the default. Do we think that Baserock will be used more commonly by developers or by integrators? Do we have any evidence, or we just guessing? | 14:22 |
| bwh | jjardon: What makes you think upstream will have run the tests on the same architecture you're using, and with the same versions of dependencies? | 14:22 |
| persia | I think morph should focus on the integrator use-case. | 14:23 |
| Kinnison | I think that once the new definitions stuff is in place, it can be up to the definitions which is default | 14:23 |
| persia | It would be nice to have a tool that could read definitions, and run the commands against a local git workspace for a project to support developers, but I think that is more sensibly a separate tool. | 14:23 |
| Kinnison | and that which we choose to be default for the demo/reference baserock systems can be our choice later | 14:23 |
| persia | As for problems detected by tests: I think we should assume anyone has the means to fix these in our context. If only by changing the ref or playing with git-bisect until they find a case where the tests pass. In most cases, the core developers can be reached by mail, and appreciate reports of test failures. | 14:26 |
| Zara | tried to upgrade (to) devel system, got this error: Too many mounts (maximum of 1024) | 14:35 |
| Zara | (this is running cycle.sh) | 14:36 |
| paulsherwood | Zara: if you do system-version-manager list, how many entries do you see? | 14:36 |
| Zara | just factory | 14:37 |
| Zara | so 1 | 14:37 |
| paulsherwood | hmmm. i'm baffled | 14:37 |
| Zara | I've been getting all sorts of weird errors with this VM; tried clearing src and upgrading to see if that'd help, but now I'm getting weird errors in that process, too. Might just start from scratch. | 14:38 |
| * SotK realises OSTree will make it easy for us to have multiple artifact-cache-server entries if we so desire (istr someone wanting multiple caches in the past but I may be wrong) | 15:42 | |
| pedroalvarez | that was me :) | 15:43 |
| pedroalvarez | i have a cache server here in my local network, but I'd like sometimes to use cache.baserock.org whenever possible | 15:44 |
| persia | I'd also like hierarchical cache servers: trying local, then cache.baserock.org | 15:50 |
| persia | I've also had discussions with an organisation that want workgroup-local caches, then an organisation-wide cache, with expected very rare use of cache.baserock.org | 15:51 |
| *** zoli__ has joined #baserock | 16:13 | |
| *** jonathanmaw has quit IRC | 16:19 | |
| *** jonathanmaw has joined #baserock | 16:26 | |
| tiagogomes_ | is there any machine booting using pxe that I can use to do a test? | 16:34 |
| radiofree | ssam2: http://paste.fedoraproject.org/190381/14248828/ | 16:47 |
| radiofree | that's my upgrade cluster once it's flashed in that way | 16:48 |
| straycat | does anyone know what's going on with the mason? i've ended up needing to build coreutils on master | 16:50 |
| ssam2 | straycat: do you mean mason-x86-64.baserock.org? when I last checked it was building stuff | 16:51 |
| *** zoli__ has quit IRC | 16:51 | |
| *** jonathanmaw has quit IRC | 16:52 | |
| straycat | ssam2, yes | 16:53 |
| ssam2 | it's currently building 'fuse' | 16:55 |
| *** zoli__ has joined #baserock | 16:56 | |
| straycat | ahh :/ | 16:57 |
| *** a1exhughe5 has quit IRC | 16:58 | |
| *** sherm_ has quit IRC | 17:26 | |
| Zara | has anyone else recently (ie: in the last few days) tried building a new baserock vm (using virt-manager) and then upgrading it to a devel vm? (just trying to work out if my odd run of errors over the last couple of days is definitely all my own fault) | 17:30 |
| *** CTtpollard has quit IRC | 17:37 | |
| *** gfinney_ has quit IRC | 17:38 | |
| *** gfinney has quit IRC | 17:38 | |
| *** mdizzle has quit IRC | 17:44 | |
| *** ssam2 has quit IRC | 17:54 | |
| *** zoli__ has quit IRC | 17:58 | |
| *** bashrc has quit IRC | 18:15 | |
| *** CTtpollard has joined #baserock | 18:21 | |
| *** CTtpollard has quit IRC | 18:30 | |
| *** franred has quit IRC | 18:34 | |
| *** SotK has quit IRC | 18:37 | |
| *** bwh has quit IRC | 18:37 | |
| *** perryl has quit IRC | 18:37 | |
| *** jmalk has quit IRC | 18:37 | |
| *** Zara has quit IRC | 18:37 | |
| *** mwilliams_ct has quit IRC | 18:37 | |
| *** wdutch has quit IRC | 18:37 | |
| *** jmacs has quit IRC | 18:37 | |
| *** paulsherwood has quit IRC | 18:37 | |
| *** DavePage has quit IRC | 18:37 | |
| *** petefoth1ringham has quit IRC | 18:37 | |
| *** ridgerunner has quit IRC | 18:37 | |
| *** pdar has quit IRC | 18:37 | |
| *** benbrown_ has quit IRC | 18:37 | |
| *** kejiahu has quit IRC | 18:37 | |
| straycat | o/ | 18:37 |
| *** petefotheringham has joined #baserock | 18:39 | |
| *** bwh has joined #baserock | 18:39 | |
| *** gary_perkins has quit IRC | 18:41 | |
| *** perryl has joined #baserock | 18:41 | |
| *** gary_perkins has joined #baserock | 18:41 | |
| *** SotK has joined #baserock | 18:46 | |
| *** tiagogomes_ has quit IRC | 18:47 | |
| *** CTtpollard has joined #baserock | 18:55 | |
| *** gfinney has joined #baserock | 18:57 | |
| *** gfinney_ has joined #baserock | 18:57 | |
| *** lachlanmackenzie has quit IRC | 19:21 | |
| *** CTtpollard has joined #baserock | 19:37 | |
| *** CTtpollard has quit IRC | 19:37 | |
| *** bfletcher has left #baserock | 19:43 | |
| *** gfinney_ has quit IRC | 20:01 | |
| *** gfinney has quit IRC | 20:01 | |
| *** gary_perkins has quit IRC | 21:21 | |
| *** kejiahu has joined #baserock | 22:34 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!