https://pantsbuild.org/ logo
Powered by
#development
  • w

    witty-crayon-22786

    06/16/2020, 7:33 PM
    We should probably bake that thing in as a custom rule
  • a

    aloof-angle-91616

    06/16/2020, 7:37 PM
    @witty-crayon-22786 i would love to use a 
    __run.sh
    file in all of our remote process executions as well, now that i think about it: https://github.com/pantsbuild/pants/pull/10030#discussion_r441095942
  • a

    aloof-angle-91616

    06/16/2020, 7:38 PM
    it seems like the one solution that avoids having two code paths
  • w

    witty-crayon-22786

    06/16/2020, 7:58 PM
    yea, maybe that’s an eventual followup… let’s see how it goes for now!
  • w

    witty-crayon-22786

    06/16/2020, 7:58 PM
    thanks
  • a

    aloof-angle-91616

    06/16/2020, 8:10 PM
    that sounds great!
  • w

    witty-crayon-22786

    06/16/2020, 8:48 PM
    @hundreds-breakfast-49010: mind stamping https://github.com/pantsbuild/pants/pull/10044 ?
  • h

    hundreds-breakfast-49010

    06/17/2020, 5:27 PM
    interested in people's opinions about the naming of the Python class that allows rules to modify their workunits in https://github.com/pantsbuild/pants/pull/10053 - @polite-garden-50641 @hundreds-father-404
    w
    h
    • 3
    • 116
  • w

    witty-crayon-22786

    06/17/2020, 7:53 PM
    bad news good news time re: rust pantsd client!
    h
    h
    • 3
    • 23
  • f

    fast-nail-55400

    06/17/2020, 8:49 PM
    are there good examples in the pants source of wrapping some code in a workunit that someone could point me at?
    w
    h
    • 3
    • 39
  • h

    hundreds-father-404

    06/18/2020, 8:56 PM
    Benjy had a good insight. Since v1, we’ve been dealing with source roots by stripping them. At least for Python, we can instead work around it by setting 
    PYTHONPATH
    . This solves a couple issues, primarily around the output of tools like Pytest and linters. Naturally, those are outputting stripped files, which is confusing. So, in V1, we wrote some plugins like the coverage plugin to restore the full name. These plugins are expensive to maintain The only blocker is that you must tell Pex 
    --inherit-path=fallback
    to pick up 
    PYTHONPATH
    , but this leaks site packages. Benjy is going to work on https://github.com/pantsbuild/pex/issues/764 to fix this
    🎉 1
    w
    • 2
    • 10
  • a

    aloof-angle-91616

    06/18/2020, 8:59 PM
    if you do that you’ll have 90% of what we need to create a virtualenv for an IDE (just need to set the .pth file). not sure what people are doing now to set up python dev from pants projects but at twitter i worked on a venv builder during hackweek. we’re going to be productionizing that soon. just a thought
    🎉 1
  • w

    witty-crayon-22786

    06/18/2020, 10:12 PM
    yea, venv export is a thing.
  • h

    hundreds-father-404

    06/19/2020, 1:05 AM
    Hey @hundreds-breakfast-49010, what is the proper way for me to put a log warning in 
    pytest_runner.py
    ? I want to warn when coverage data could not be generated
  • p

    polite-garden-50641

    06/19/2020, 1:07 AM
    When running tests in the pants repo (in order to debug stuff), how do I disable the creation of 
    pants_local_binary.pex
    so things will run faster?
  • p

    polite-garden-50641

    06/19/2020, 1:07 AM
    I remember there was some flag/env variable I can set, but I don't recall the name
  • h

    hundreds-breakfast-49010

    06/19/2020, 1:08 AM
    @hundreds-father-404 the stdlib python logger functionality gets forwarded to rust and printed inthe same way as rust logs themselves do
    👍 1
    h
    • 2
    • 1
  • h

    hundreds-father-404

    06/19/2020, 1:08 AM
    export NO_REGEN_PEX=true
    👍 1
  • h

    hundreds-breakfast-49010

    06/19/2020, 1:08 AM
    but we're trying to move away from this model
  • h

    hundreds-breakfast-49010

    06/19/2020, 1:09 AM
    what we probably want is for the rule that would generate the error to use the 
    EngineAware
    class I just merged
  • h

    hundreds-breakfast-49010

    06/19/2020, 1:09 AM
    to set the workunit output to warning
  • h

    hundreds-breakfast-49010

    06/19/2020, 1:09 AM
    and then print some kind of additional message, that we don't have an abstraction for yet
  • h

    hundreds-breakfast-49010

    06/19/2020, 1:10 AM
    this is basically teh same problem as that 
    TODO
    on the test runner
  • h

    hundreds-father-404

    06/19/2020, 1:10 AM
    Okay, I’ll go ahead with this current approach for now.
  • w

    witty-crayon-22786

    06/19/2020, 2:28 AM
    i think that we will continue to actually render logs from rules. they just shouldn’t be necessary in the common case. so warnings are a good example of something that is worth rendering there.
    👍 2
  • w

    witty-crayon-22786

    06/19/2020, 2:30 AM
    i think that the difference with the test runner is that it’s the common/default case.
  • h

    happy-kitchen-89482

    06/19/2020, 6:06 AM
    @witty-crayon-22786 I'm getting the 
    DEPRECATED: --enable-pantsd defaulting to False will be removed in version 1.30.0.dev0.
    warning in a repo that doesn't set --enable-pantsd (you should be able to repro this in the pantsbuild/setup repo by running 
    ./pants
    in it).
    h
    • 2
    • 2
  • w

    witty-crayon-22786

    06/19/2020, 9:26 PM
    @hundreds-father-404: still thinking about landing https://github.com/pantsbuild/pants/pull/9825 ?
    h
    • 2
    • 6
  • a

    aloof-angle-91616

    06/20/2020, 2:32 AM
    going to rebase and respond to comments on SnapshotSubset() perf improvement today => https://github.com/pantsbuild/pants/pull/9779
  • a

    aloof-angle-91616

    06/20/2020, 7:45 AM
    https://github.com/pantsbuild/pants/pull/9779 <= SnapshotSubset() perf improvement is now vastly cleaned up and better documented
1...745746747...906Latest