Bump futures - Githubissues

twitter-archive / commons

Twitter common libraries for python and the JVM (deprecated)

http://twitter.github.com/commons

Other

2.1k stars 565 forks source link

Bump futures #471

Closed OniOni closed 4 years ago

OniOni commented 5 years ago

Problem

We have a dependency that pulls in a conflicting version of futures.

Solution

Loosen constraint on futures as to yield a resolvable set of dependencies.

Result

This should have no effect on end users. futures releases seem to be just bug fixes .
I ran, ./pants test tests/python/twitter/common/concurrent::, successfully.

Notes

Let me know if I should run any additional tests
I'm also happy to go with a pinned version of futures, let me know what makes the most sense.

CLAassistant commented 5 years ago

Thank you for your submission, we really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

OniOni commented 5 years ago

I see build failed in CI. I don't seem to be able to rerun things in Travis. But I ran ./build-support/bin/ci.sh -s -j locally.

I'm getting some failures, but as far as I can tell, they're all linked to not being able to start a zookeeper cluster.

I get 54 failures. Greping the pytest output, I get 54 instances of the NotStarted: Could not start Zookeeper cluster! exception.

So I'm not sure if my failing CI build is flakiness or a legitimate failure.

mateor commented 5 years ago

@OniOni works with me at Foursquare. @stuhood or @kwlzn can you rerun the CI for us?

OniOni commented 5 years ago

Thanks to whomever reran the build. Looks like failure is same a last time. Will dig into this, thanks!

OniOni commented 5 years ago

Wouldn't mind a rerun. I got a green on travis build for my fork: https://travis-ci.org/OniOni/commons/builds/476936976

kwlzn commented 5 years ago

reran

OniOni commented 5 years ago

@kwlzn thank you so much. Looks like a different failure. Could I bother you for another rerun?

I'll keep track of failures, hopefully I can try and fix those flakes (if they're flakes, that is)

test_sampler_base

mateor commented 5 years ago

If the CI runs are the most straightforward way to debug, you can turn on travisci.org builds for your forked copy.

This is still free and while it won't matriculate status back here, it gives you access to a rebuild button and a single-node iteration loop

OniOni commented 5 years ago

@mateor: I did indeed do this. It failed on the first run, but I did get a green on the second run (https://travis-ci.org/OniOni/commons/builds/476936976). This is why I asked for that rerun 8 days ago as I am fairly confident these are flakes.

To be fair, tests are flaking pretty badly and that is an issue in itself. I'm not convinced that my changes accentuated flakes, but it's hard to tell without historical data.

grimreaper commented 4 years ago

Thank you for your contribution. Unfortunately, we’re not going to continue maintaining twitter-commons and are archiving all pull requests.