#docker-dev

I get TestExitCode failing consistently

I also get seemly occational failures on TestBuildFails with "Error should be utils.JSONError"

[docker] vieux closed pull request #3814: add env to engine, remove all srv.runtime from the api (master...add_env_to_engine) http://git.io/cVMzZw

shykes: Re the planned trello board, https://huboard.com/ and http://www.gitban.com/ provide a trello-style interface, and are backed by github issues. Not sure if that'd be preferable.

I've never used either, but their premise is interesting, and might be worth considering.

tianon: Yes, those tests are flaky now. There's no point in fixing them now because the extraction of the execution engine is still being worked on.

anyone played around with fig? http://orchardup.github.io/fig/?

deni: you should probably ask questions about userspace tools in the #docker channel, I expect the people there to have more experience with stuff like that

Tranquility: true

Tranquility: Docker's code is changing a bit now, some chunks of code are moving into plugins. If there are some kind of performance problems, they'll be addressed after this extraction of code into plugins

Please don't worry too much about it right now.

ok, thanks

backjlack: ok thanks :)

Are 0.8 changes locked in or is whatever is in master on 2/1?

strictly time-based, so whatever is in master at the time the button gets pressed

then one month later, 0.9 in the same fashion

fish_: will the https api make it into 0.8?

deni: 0.8 is scheduled for next tuesday. Whatever is merged in master at that time will be in the release

seems doable

shykes: yeah that's why i'm asking

:)

that is going to be so awesome

shykes: is there still something left to discuss re https api?

alexlarsson: ping

sorry I'm in a meeting will be back

shykes: np, just let me know if there is something I can work on and I'll pick it up as soon as I find time

would be awesome to have https soon :)

joffrey: Is it just me or is ubuntu:12.10 tagged as latest in the ubuntu repository?

backjlack: possibly. tianon said the same thing. I can change it back if it's a big deal for people

(although if you rely on the "latest" tag being immutable you're going to have a bad time regardless)

joffrey: IMO, the "latest" tag should be assigned to the newest LTS version available.

joffrey, backjlack: I agree; that's why I set it that way in the library file :) I'll be regenerating those images today though, since backjlack reviewed my script changes \o/

tianon: backjlack: fixed it in the meanwhile

thanks joffrey :)

oh, I also have a PR that needs to be fixed to reference ubuntu:* now :D

joffrey: That's great. Thank you!

hypothetical scenario: i have an image on the index that was manually built and has alreay a couple of commits. if i rebuild the image using a dockerfile and try and push that new image with the same name what will happen?

will the system figure out that it's a new image and just nuke the old history?

or am i better of deleting the images from the index first and then recreating them.

but even then, what if someone already pulled down those images. when he does the pull after the switch what will happen?

deni: it will overwrite

pulling after the switch will overwrite also

on the target system

shykes: awesome

tnx

ps for usage questions could you use #docker instead of #docker-devm that way other users can see the answer too

#docker-dev is more about the development process - sometimes a blurry line but still :)

shykes: i did, before, nobody knew :) still, sorry for the noise

deni: gotta be patient, or try it out. #docker is most active in the US daylight hours

keeb: i figured

tianon: that makes sense, because 1.9 isn't valid in the test I ran.

[docker] tianon pushed 2 new commits to master: http://git.io/2ygx8Q

docker/master 5a3fa3c Tianon Gravi: Add a few small tweaks to mkimage-debootstrap, especially to help with lucid...

docker/master e1b210b Tianon Gravi: Merge pull request #3588 from tianon/debootstrap-tweaks...