pantsbuild #announce

Channels

welcome

github-notifications

development

witty-crayon-22786

07/08/2016, 5:30 PM

@happy-kitchen-89482: yea, fine with me

happy-kitchen-89482

07/08/2016, 5:31 PM

thx

happy-kitchen-89482

07/08/2016, 5:34 PM

Done

witty-crayon-22786

07/15/2016, 9:02 PM

prepped 1.2.0dev0: https://rbcommons.com/s/twitter/r/4084/

witty-crayon-22786

07/15/2016, 9:38 PM

...argh. need to update version parsing before then apparently.

mammoth-pharmacist-54215

07/15/2016, 9:39 PM

need to update version parsing?

mammoth-pharmacist-54215

07/15/2016, 9:52 PM

anyone else have a minute to take a look at: https://rbcommons.com/s/twitter/r/4083/ (1.1.0rc8 release)

witty-crayon-22786

07/15/2016, 10:23 PM

@mammoth-pharmacist-54215:

1.2.0dev0

is not legal semver, so

src/python/pants/version.py

will not accept it.

mammoth-pharmacist-54215

07/15/2016, 11:06 PM

I see you mean the code in pants won’t accept it as opposed to pypi or pip

witty-crayon-22786

07/18/2016, 5:27 PM

@mammoth-pharmacist-54215: an alternative might be to just say that stable branches are always created from a release tag

witty-crayon-22786

07/18/2016, 5:28 PM

...ie, a

dev

tag in master

witty-crayon-22786

07/18/2016, 5:28 PM

and the tag you create the stable branch from is the first rc

mammoth-pharmacist-54215

07/18/2016, 5:28 PM

so always do the first release from master and then push that change to the stable branch… than after that we do rc releases from the stable branch?

witty-crayon-22786

07/18/2016, 5:29 PM

yea

witty-crayon-22786

07/18/2016, 5:29 PM

either way, at that point it's just a question of naming.

mammoth-pharmacist-54215

07/18/2016, 5:29 PM

yeah, would be good to nail down though so we do it consisently.

rough-minister-58256

07/18/2016, 5:30 PM

we have a ton of problems internally with resolving at build time

mammoth-pharmacist-54215

07/18/2016, 5:30 PM

Plus it keeps the release names a little more sane

witty-crayon-22786

07/18/2016, 5:31 PM

mammoth-pharmacist-54215

07/18/2016, 5:31 PM

the first RC that we send out an email for feedback will be a dev release instead of rc which is a little weird but not awful

mammoth-pharmacist-54215

07/18/2016, 9:22 PM

Last week Stu and I ran into a problem because x.y.zrcN is not a valid semver. Would people prefer we update the code in pants used to check the version or update the docs and switch to using the dash?

mammoth-pharmacist-54215

07/18/2016, 9:23 PM

to unblock last week we released with a dash in the version numbers.

witty-crayon-22786

07/18/2016, 9:24 PM

@mammoth-pharmacist-54215: we still have the ticket open about that one

witty-crayon-22786

07/18/2016, 9:25 PM

https://github.com/pantsbuild/pants/issues/3382

mammoth-pharmacist-54215

07/18/2016, 9:25 PM

thanks

mammoth-pharmacist-54215

07/18/2016, 9:26 PM

this actually goes over the pre vs rc which is a slightly different issue.

mammoth-pharmacist-54215

07/18/2016, 9:27 PM

happy to just make a decision if people are okay with that.

witty-crayon-22786

07/18/2016, 9:28 PM

@mammoth-pharmacist-54215: see the ticket title though

mammoth-pharmacist-54215

07/18/2016, 9:29 PM

I see, not sure if that will address the issue without the dash but I’ll test

mammoth-pharmacist-54215

07/21/2016, 9:15 PM

So we have a use case at twitter that broke from a recent bundle change I made. I am unable to find a way to fix internally. Basically we have a target that gets constructed with TargetMacro.Factory.wrap. Inside the bundle patch there were some guards added that didn’t exist before one of them gets confused in this use case. if vt.target in self.context.target_roots: