IRC logs for #baserock for Tuesday, 2014-08-19

*** flatmush [] has quit [Ping timeout: 240 seconds]00:00
*** flatmush [] has joined #baserock00:02
*** flatmush [] has quit [Ping timeout: 260 seconds]00:08
*** flatmush [] has joined #baserock00:08
*** flatmush [] has quit [Ping timeout: 245 seconds]00:12
*** flatmush [] has joined #baserock00:16
*** flatmush [] has quit [Ping timeout: 245 seconds]00:25
*** flatmush [] has joined #baserock00:26
*** flatmush [] has quit [Ping timeout: 246 seconds]00:31
*** flatmush [] has joined #baserock00:31
*** flatmush [] has quit [Ping timeout: 244 seconds]00:59
*** flatmush [] has joined #baserock01:02
*** flatmush [] has quit [Ping timeout: 240 seconds]01:28
*** flatmush [] has joined #baserock01:29
*** flatmush [] has quit [Ping timeout: 240 seconds]01:37
*** flatmush [] has joined #baserock01:41
*** flatmush [] has quit [Ping timeout: 245 seconds]01:57
*** flatmush [] has joined #baserock01:58
*** flatmush [] has quit [Ping timeout: 240 seconds]02:12
*** flatmush [] has joined #baserock02:16
*** flatmush [] has quit [Ping timeout: 260 seconds]02:25
*** flatmush [] has joined #baserock02:26
*** flatmush [] has quit [Ping timeout: 250 seconds]02:38
*** flatmush [] has joined #baserock02:41
*** flatmush [] has quit [Ping timeout: 272 seconds]03:01
*** flatmush [] has joined #baserock03:12
*** flatmush [] has quit [Ping timeout: 250 seconds]03:18
*** flatmush [] has joined #baserock04:34
*** flatmush [] has quit [Ping timeout: 240 seconds]04:39
*** flatmush [] has joined #baserock04:47
*** flatmush [] has quit [Ping timeout: 250 seconds]04:55
*** flatmush [] has joined #baserock04:57
*** flatmush [] has quit [Ping timeout: 264 seconds]05:08
*** flatmush [] has joined #baserock05:24
*** flatmush [] has quit [Ping timeout: 244 seconds]05:38
*** flatmush [] has joined #baserock05:40
*** flatmush [] has quit [Ping timeout: 250 seconds]05:47
*** flatmush [] has joined #baserock05:52
*** flatmush [] has quit [Ping timeout: 272 seconds]06:21
*** flatmush [] has joined #baserock06:22
*** flatmush [] has quit [Ping timeout: 240 seconds]06:33
*** flatmush [] has joined #baserock06:34
*** flatmush [] has quit [Ping timeout: 245 seconds]06:39
*** flatmush [] has joined #baserock06:40
*** flatmush [] has quit [Ping timeout: 240 seconds]06:46
*** flatmush [] has joined #baserock06:47
*** flatmush [] has quit [Ping timeout: 255 seconds]06:52
*** flatmush [] has joined #baserock06:52
*** fay_ [] has joined #baserock07:04
*** franred [] has joined #baserock07:49
*** jonathanmaw [] has joined #baserock08:22
*** ssam2 [] has joined #baserock08:54
jjardonHi!, I was trying to deploy the other day and I kept getting this message error (deploy used to work before): Any idea where should I put VERSION_LABEL and why is needed now and not before?09:04
straycatYou should put it in the gnome-deploy cluster morph, though I don't know why it's needed now.09:06
pedroalvarezBaserock Intrepid I-0.1 has been released!09:06
rjekIntrepid?  What does this mena?09:11
pedroalvarezrjek: the previous release was called Horizon H-1.1, so they are just using the alphabet09:12
pedroalvarezand random words09:12
rjekOh, this is Genivi?09:12
straycatThat the release can be likened to a fearless adventurer?09:13
rjekIntrepid was not a good choice.  That was a release of Ubuntu.09:13
pedroalvarezhm... did get confused with the name?09:13
pedroalvarezno, it's called Intrepid09:14
rjek(This is another problem with codenames, IMO: they're fine for internal use, but for releases they're a bad idea.  People naturally expect them to be unique in a given field.  People don't expect that of boring numbers.)09:15
straycatCodenames based on a particular scheme rather than random words might work better.09:16
rjekYou still run the risk of namespace clash :)09:17
* straycat nods09:17
petefothstraycat: nope! Numbers - done properly - is what you need (IMHO)09:17
rjekStep one: Pick a name and Google it.  If it has more than a couple of thousand hits, return to step one.09:21
straycatYou could write a crawler that uses machine learning to classify codenames as "software" or "other" Then remove any codenames in the "software" category and pick from those :p09:21
rjekStep two: Use numbers.09:21
jonathanmawpedroalvarez, not quite random words for GENIVI releases, they're space-themed09:21
Kinnisonjjardon: Looks like it's doing an upgrade (raw disk image in place from previous deploy?) so it needs VERSION_LABEL since otherwise it can't know what to call the upgraded system09:21
ssam2jjardon: yeah, the change is that your .img file now already exists09:22
ssam2it should probably make that clearer09:22
richard_mawIMO it should only attempt an upgrade if --upgrade is set09:29
ssam2richard_maw: of course. I think we never updated rawdisk.write to actually check that09:42
KinnisonAnyone know where all these: 2014-08-19 09:47:14 DEBUG run external command: [['git', 'config', '-z', 'core.bare']]09:47
Kinnisoncome from?09:47
Kinnisonwhen running morph09:47
richard_mawmorphlib.gitdir.GitDirectory IIRC09:47
richard_mawthere's a property that calls it, and fails to cache it09:48
KinnisonIt consumes many seconds of runtime on my rns of morph09:48
Kinnison(this is negligible, but irritating)09:48
richard_mawI wish we were on python3, there's all kinds of useful decorators for this kind of thing10:01
straycatWhat stops us from using python3?10:17
KinnisonVarious of our deps are still python2 only10:19
KinnisonWe'd need to update them10:19
KinnisonIn some cases, update == recode10:19
richard_mawwe use cliapp + all its depends, pyyaml and pyfilesystem10:19
richard_mawthough I think the latter 2 have python3 support10:19
persiaCan six help in that direction, or is it only for avoiding upgrades?10:20
richard_mawAIUI six is supposed to help if you still need to support older versions of python10:21
richard_mawso perhaps useful for porting cliapp to python3, but not so useful for morph itself10:21
richard_mawsince we could just opt to use python310:21
paulsherwoodthe GENIVI names are 'space-themed'10:29
* rjek tries to think what Intrepid might be, other than the class of vessel featured in Star Trek: Voyager10:30
richard_mawwell, you will grasp at straws if you need every letter of the alphabet10:30
persiaDepends on the naming scheme.  For something like "flower names", or "fruits", it's usually safe.  For something like "Planets", one has trouble getting started.10:32
* paulsherwood hopes to see GENIVI Klingon in his liefetime10:32
rjekWords in the dictionary, in alphabetical order.  Also makes versions comparable.10:32
* richard_maw would prefer first thing you see that starts with that letter10:33
paulsherwoodwin 4010:35
richard_mawpaulsherwood: currently tidying up a patch to fix that weirdness you had with a build that failed to contact the trove, so decided to use the default commands10:35
pedroalvarezrichard_maw: will it fail instead of using the default commands?10:36
jjardonKinnison: ssam2 that was it, Ive removed the previous image and it works.10:40
Kinnisongood to hear10:40
paulsherwoodany idea why "tar xvf ../baserock-e424935744d-devel-system-jetson.tar.gz" from baserock-jetson instructions would give me "tar: no gzip/bzip2/xz magic"10:49
KinnisonIt might not be compressed, yet is named .gz10:50
richard_mawdo you have `file`, and if you do, what does it say?10:50
paulsherwood../baserock-e424935744d-devel-system-jetson.tar.gz: HTML document, ASCII text10:51
paulsherwoodmaybe it was uploaded strangely?10:51
KinnisonWhat URL did you download it from and I'll take a looksee10:51
paulsherwoodcurl -k -O
De|ta_ is now known as De|ta10:52
paulsherwoodi believe this is a new image radiofree posted overnight10:52
richard_mawpaulsherwood: use http not https10:54
Kinnisonhebetude% curl -k -I                                                             ~10:54
KinnisonHTTP/1.1 401 Authorization Required10:54
Kinnisonhebetude% curl -k -I                                                              ~10:54
KinnisonHTTP/1.1 200 OK10:54
* Kinnison concurs. http not https10:54
paulsherwoodah. i was cutting and pasting instructions from wik. i'll fix them10:54
paulsherwoodwhoa.. that's going to take 45 minutes, i may need a different network10:56
Kinnisonit's a 400 meg file10:57
Kinnisonca. 150kB/s is believable for our internal network right now10:57
SotKpaulsherwood: copy/pasting those instructions worked for me10:58
paulsherwoodSotK: when did you last run them? radiofree tweaked things overnight...10:59
SotKpaulsherwood: this morning11:00
paulsherwoodweird, then11:00
richard_mawit could be our internal network vs external network issues11:01
richard_mawSotK: were you using the internal network, guest network or a wifi dongle?11:02
paulsherwoodrichard_maw: are you saying it's me, again? :)11:02
richard_mawpaulsherwood: not unless there's a manner in which you _are_ the internal network11:02
richard_mawsince I would have gotten the same result as you did, had I followed the instructions on the wiki page11:02
SotKrichard_maw: I was using a wifi dongle when I did it, fwiw11:02
richard_mawyep, it's some issue with our internal network, so https would normally be appropriate instructions11:03
persiaY'all need to find a better network :)  I usually complain about my networks, but they don't usually cause http vs. https sorts of issues.11:04
Kinnisonhttp vs. https is that the server is on doesn't have (a) a certificate or (b) SNI11:04
Kinnisonor if it does have SNI, it's not doing it quite right11:04
persiaSo, I think it's not your network.  I can replicate Kinnison's results from my DC, where https doesn't work.11:07
* Kinnison is quite confident that he's right11:08
persiaHmm.  c0b2381ed8c9fe2af96a98aaab60383757e9caf9 doesn't come with a comment, but I think it's best reverted (as done in 97cdd4f8f6f9e91b5aeff76e3c04f24331c98809)11:11
Kinnisonwhich repo is that?11:11
Kinnisonrichard_maw: I'm pondering your cat-file change and will reply after I've had some lunch11:14
doffmHas anyone tried kvm (libvirtd) on baserock recently?11:25
richard_mawdoffm: not recently, it's not part of the devel systems any more because we didn't need it for day-to-day development, and needing to build it slowed down our day-to-day development12:25
doffmrichard_maw: Ok, thanks. Not working on my devel machine. I'll do another devel build with it in. Check it hasn't bit rotted.12:45
Kinnisonrichard_maw: What's your opinion of ?  (Note, untested beyond running firehose with it)14:10
Kinnisonrichard_maw: there's definite whitespace cockups :-)14:12
richard_mawKinnison: there looks to be something a bit odd with the whitespace, and I'd prefer the dict to be _config, so people don't try to use that instead of get_config14:12
KinnisonI might turn it into a proper patch at somepoint then14:12
Kinnisonrichard_maw: newer version:
Kinnisonrichard_maw: Also I'd like to put into definitions master to mean I can build newer cpython which is needed for our CA/CI/CD pipe14:51
richard_maw+2 to the config thingy14:51
richard_mawfor the python2 symlink, I've have gone with `test -e`, and `condition || command` isn't my favourite idiom, as it feels the wrong way aroud to me. I'd prefer `if [ ! -x "$..." ]; then ln -s ...; fi`14:55
* rjek keeps thinking there's something on his screen because of the ~ at the end of Kinnison's pastes at 1154.14:56
rjekYou should all talk more so it scrolls off the top.14:56
richard_mawin this case I think `ln -sf` would replace it correctly, but we're better off with what the original link target is if it exists14:56
richard_mawI'm on the fence as to whether we ought to reverse the condition, and use a proper flow control structure, so +1 rather than +214:57
Kinnisonrjek: use a bigger font14:57
KinnisonAnd yeah, I should find a way to more easily not copy the RPS1 when doing pastes14:58
KinnisonAnyone else want to consider if they'd accept ?14:59
Kinnisondoffm: ^^ (since it was your change to python's morphology which I'm updating)14:59
paulsherwoodKinnison: +115:04
Kinnisonpaulsherwood: to the python morph change?15:04
paulsherwoodas a ruby user, i'm ok with the idiom :)15:05
* richard_maw adds another thing about ruby to the list15:05
Kinnisonrichard_maw: You can confirm that your +2 was correctly applied here:
richard_mawKinnison: yep, looks good15:08
pedroalvarezKinnison: is that fix related to this error:?15:10
pedroalvarezERROR: Command failed: git config -z morph.repository15:10
richard_mawit just caches the result15:10
Kinnisonit's related to the tonnes of 'git config -z core.bare' you tend to see in logs15:10
richard_mawpedroalvarez: where do you see that error?15:11
pedroalvarezrichard_maw: only happened once, not sure if I can reproduce it15:11
Kinnisonrichard_maw: You can see the result of my definitions change here:
Kinnisonpaulsherwood: ^^^15:11
richard_mawpedroalvarez: you will see non-fatal messages about that in the logs of builds if you have git repositories in your system branch that didn't come from a morph edit15:12
pedroalvarezrichard_maw: this is what happened in my case
pedroalvarezrichard_maw: hm... maybe casued by an incomplete `morph edit`15:14
pedroalvarezrichard_maw: And I fixed it removing the upstream folder of my workspace, so you are probably right15:15
*** jonathanmaw [] has quit [Quit: Leaving]16:23
*** fay_ [] has quit [Remote host closed the connection]17:01
*** franred [] has quit [Quit: Leaving]17:32
*** flatmush [] has quit [Quit: Leaving.]17:32
*** ssam2 [] has quit [Quit: Leaving]18:16
jjardonhi, after upgrading to latest morph, I get this when building Any idea if there is any recent change that can cause this?20:14
KinnisonYou're not using latest morph entirely  think20:41
KinnisonAnd if you are, then you have some bad morphologies somewhere20:42
Kinnisonbecause we're being more strict about morphology content20:42
KinnisonI suggest you run with higher debug (--verbose --log-level=debug --log=/dev/stdout) and then see if you can see which morphology is causing the error20:42
KinnisonAlso I suggest you raise on the ML that the error is unhelpful in diagnosing what's up20:43
jjardonKinnison: ok, thanks for the hints. I asked because Ive build succesfully just before switch to use morph master instead the one included in the chroot image20:45
KinnisonYeah, we've been cleaning up morphologies.  If you can integrate the work that has been done in master of definitions then you'll be in a cleaner place overall too20:45
KinnisonEsp. since then you could play with firehose for tracking all the gnome repos20:45
Kinnisonwhich might automate your work for you20:46
jjardonKinnison: yep, I was about to rebase my gnome system with the current definitions ;) Let me try if I found something with the more verbose logs and if not i will ask in the list20:49
pedroalvarezjjardon: rebasing to master of definitions is not going to be easy :/20:58
pedroalvarezjjardon: I'm just thinking that you can use the same script we have run to reorganize your definitions repository and fix your chunk morphologies20:59
pedroalvarezjjardon: and it will move your chunk morphologies to definitions.git, and I know you are interested in that20:59
jjardonpedroalvarez: yeah, that was the idea. But Id like to check before if the new morph still build my old definitions20:59
Kinnisonjjardon: Pedro and Fran pushed a fix to definitions *before* we did the move21:00
Kinnisonjjardon: If you rebase on top of that, you'll only have your morphologies at risk of not being supported21:01
pedroalvarezI was going to suggest that21:02
pedroalvarezjjardon: rebasing on top of e424935744d9faf27327ee40987620412c41716f should be safe21:02
jjardonok, thanks for the help21:05
jjardononly to be completely clear: is it normal that my current system/definitions, that build fine with the morph included in the baserock image from 18 of july, stop building when swicthing to morph master?21:13
pedroalvarezjjardon: is a situation that we are aware of21:14
jjardonpedroalvarez: ok, then I do not bother you anymore ;)21:15
pedroalvarezjjardon: not bothering at all21:15
pedroalvarezall the questions are welcome :)21:15
jjardoncan I also make a suggestion? I think it would be a good idea to separate definitions from any code/script. We have this problem in GNOME where the modulesets are in the same repo as jhbuild its difficult to follow the development of the modulesets definitions or the tool (jhbuild)21:19
jjardonits not the same as morph is in a separate repo but I saw the number of script in the definitions repo is growing21:20
pedroalvarezYeah, I see your point, now the log in definitions.git is a mix of commits of some scripts and commits of changes in the morphologies21:23
jjardonexactly. It happened to us that some people are more interested in the tool (jhbuild) and other in the modulesets.21:26
pedroalvarezjjardon: what decison have you taken?21:27
pedroalvarezto fix this issue, I mean21:27
jjardonseparate the modulesets in a different repo ( But a lot of infrastructure changes are needed so is on hold at the moment21:29
jjardon"[Build 1/1860]  [stage1-binutils-bins] Fetching to local cache: artifact stage1-binutils-bins" seems to be working again :)21:41
pedroalvarezjjardon: and reusing fresh cache from :)21:42
pedroalvarezwow, 186021:43
jjardonpedroalvarez: :) and this is only up to gnome-shell. A more complete system would be quite bigger21:51
* jjardon owns a beer to the organize-morphologies script22:03
jjardonbut still I get an error:
pedroalvarezjjardon: is this running organize-morphologies script?22:10
jjardonpedroalvarez: yes22:10
jjardonnote I moved wayland-ivi-extension to a new weston-genivi stratum22:11
pedroalvarezwhich sha1 are you using of the wayland-ivi-extension repo?22:12
jjardonthis is how it looks like:
jjardonoh, let me check22:12
jjardonpedroalvarez: 8b59529cf3b279a948f253bc79263d00d3f051a522:15
pedroalvarezjjardon: found it!22:15
pedroalvarezs/moprh/morph/ in that stratum22:16
jjardonpedroalvarez: good catch! ;)22:18
jjardonmmm, seems I hit another:
jjardonalso, it would be nice if the script could ignore the repos with uri "file://", will take a look tomorow23:07

Generated by 2.15.3 by Marius Gedminas - find it at!