pantsbuild #general

Channels

welcome

github-notifications

announce

witty-crayon-22786

09/29/2015, 11:01 PM

(now, at least)

freezing-motorcycle-14222

09/30/2015, 12:04 AM

Oh wow, using the mutex trick that @witty-crayon-22786 mentioned fixed all of the issues that I observed

witty-crayon-22786

09/30/2015, 12:04 AM

neat

witty-crayon-22786

09/30/2015, 12:05 AM

are you planning to upstream this stuff soon?

freezing-motorcycle-14222

09/30/2015, 12:05 AM

Yes

witty-crayon-22786

09/30/2015, 12:05 AM

great

freezing-motorcycle-14222

09/30/2015, 12:05 AM

The diff is quite small. I'll have a reviewboard up later tonight

witty-crayon-22786

09/30/2015, 12:05 AM

excellent

rough-minister-58256

09/30/2015, 1:00 AM

https://code.google.com/p/ayb/source/browse/tools/plink/plink.py <— wonder how related this is to googles .par file implementation? via https://github.com/mzhaom/bazel/commit/efebe83e9fb3d6110aaa463cd3af49ff9b11f1c5

freezing-motorcycle-14222

09/30/2015, 1:16 AM

There's one thing I don't understand about the REPL mutex mixin code, which is why it calls

self.context.release_lock()

witty-crayon-22786

09/30/2015, 1:18 AM

@freezing-motorcycle-14222: will want to git-blame that to the review that added the mutex stuff

freezing-motorcycle-14222

09/30/2015, 1:18 AM

I already did

freezing-motorcycle-14222

09/30/2015, 1:18 AM

There's no explanation in the original review board

freezing-motorcycle-14222

09/30/2015, 1:19 AM

I think it's not necessary for what I'm doing, but I just wanted to make sure there wasn't something obvious that I'm missing

freezing-motorcycle-14222

09/30/2015, 1:19 AM

Oh, never mind

freezing-motorcycle-14222

09/30/2015, 1:19 AM

I see why it's there. John was just preserving the behavior from the original task from before the mutex refactor

freezing-motorcycle-14222

09/30/2015, 1:19 AM

So it's not relevant to my code. Never mind

limited-terabyte-34878

09/30/2015, 3:12 PM

Thanks @witty-crayon-22786 , I think I may call this enough functionality for the superclass for now, add some tests and submit for full review

limited-terabyte-34878

09/30/2015, 3:13 PM

what’s the pants community view on inheritance vs composition? I could reorganize what I’ve done to be more composition oriented per justin’s suggestion

ancient-stone-50795

09/30/2015, 3:53 PM

@gorgeous-fish-11040: FYI, I know you’ve gotten a few issues from our developers, but today one of them came up to me spontaneously and said, “I’m loving the Pants IntelliJ plugin!”

gorgeous-fish-11040

09/30/2015, 3:57 PM

Glad to hear! BTW sorry for the delay for IJ 15. Feeling not to well last few days to get some extra time to fix it.

ancient-stone-50795

09/30/2015, 4:01 PM

Hope you feel better

bored-art-40741

09/30/2015, 5:54 PM

Hey @witty-crayon-22786 @fancy-queen-20734 are you around?

fancy-queen-20734

09/30/2015, 5:54 PM

I am

bored-art-40741

09/30/2015, 5:55 PM

We need to work out what's going on here: https://rbcommons.com/s/twitter/r/2901/

bored-art-40741

09/30/2015, 5:56 PM

There is definitely a bug, but it's difficult to repro and therefore test

bored-art-40741

09/30/2015, 5:56 PM

I think the original authors need to dig into this more

bored-art-40741

09/30/2015, 5:57 PM

And in particular there probably needs to be more unit testing of the status table code and usage

bored-art-40741

09/30/2015, 5:57 PM

The failure mode is so bad that we can't ignore this, and we're uncomfortable with the more or less arbitrary patch that "fixes" it in that RB

witty-crayon-22786

09/30/2015, 6:07 PM

@limited-terabyte-34878: mixins are more like composition... particularly if they never override concrete members