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

    hundreds-father-404

    10/01/2019, 1:00 AM
    We use 
    Collection.of()
    , right? Maybe we could replace that with 
    Tuple[X, ...]
  • a

    aloof-angle-91616

    10/01/2019, 1:00 AM
    and we generally just like using an 
    @dataclass
    wrapper -- there's 
    Collection.of
    in 
    pants.engine.objects
    which we use to wrap collections of objects
  • a

    aloof-angle-91616

    10/01/2019, 1:00 AM
    the 
    @dataclass
    PR fixes that too!
    h
    • 2
    • 2
  • h

    hundreds-breakfast-49010

    10/01/2019, 1:02 AM
    so could I write an 
    @rule
    that has a signature 
    Tuple[X]
    , and then use that?
    a
    h
    w
    • 4
    • 41
  • h

    hundreds-breakfast-49010

    10/01/2019, 7:00 PM
    so apparently this is the error message you get if you forget to specify a 
    name
    attr on the 
    Goal
    subclass
  • h

    hundreds-breakfast-49010

    10/01/2019, 7:00 PM
    which is not very helpful
  • a

    average-vr-56795

    10/01/2019, 7:20 PM
    That feels both really annoying, and really easy to both short-term and long-term fix 😄
  • h

    hundreds-breakfast-49010

    10/01/2019, 10:52 PM
    I'm trying to fix https://github.com/pantsbuild/pants/pull/8329#discussion_r328735829 and I'm not how to go about constructing a rust FFI function that avoids calling 
    scheduler.core.executor.block_on
    that can still return a 
    PyResult
    synchronously
    a
    w
    • 3
    • 62
  • h

    hundreds-breakfast-49010

    10/02/2019, 8:26 PM
  • h

    happy-kitchen-89482

    10/02/2019, 8:27 PM
    These OSX Rust CI timeouts, known? Are we ignoring them? https://travis-ci.org/pantsbuild/pants/jobs/592314229
    w
    • 2
    • 4
  • h

    happy-kitchen-89482

    10/02/2019, 8:28 PM
    Seems to be entirely in setting up xcode etc?
  • w

    witty-crayon-22786

    10/02/2019, 9:49 PM
    @red-balloon-89377, @average-vr-56795: interested in your input on https://docs.google.com/document/d/1Hn73YlhTPROlULTMa_3A-Fdv7hAPiHFII8rvXri5l7E/edit#bookmark=id.swk5himqweo4 if you have time to give it tomorrow
    h
    • 2
    • 70
  • w

    witty-crayon-22786

    10/03/2019, 4:55 AM
  • w

    witty-crayon-22786

    10/03/2019, 6:26 PM
    @hundreds-breakfast-49010: depends what kind of test. 
    run_rule
    vs 
    TestBase
    vs 
    ConsoleRuleTestBase
    vs 
    PantsRunIntegrationTest
  • w

    witty-crayon-22786

    10/03/2019, 6:27 PM
    in the middle two, you need to add to 
    def rules(self): ...
  • w

    witty-crayon-22786

    10/03/2019, 6:27 PM
    for the last one... it needs to actually be in the pex before you start.
  • h

    hundreds-breakfast-49010

    10/03/2019, 6:27 PM
    @witty-crayon-22786 in https://github.com/gshuflin/pants/commit/ae3479cd386b46ef787acf3d7bd8d059eb4b7886 I'm creating a 
    console_rule
    that outputs a type 
    TestWorkspaceGoal
  • w

    witty-crayon-22786

    10/03/2019, 6:27 PM
    so you can't integration test "fake" stuff
  • h

    hundreds-breakfast-49010

    10/03/2019, 6:28 PM
    and as soon as I add 
    register_options
    on that class, I get 
    pants.option.config.Config.ConfigValidationError: No such options scope: test-workspace-goal
  • a

    aloof-angle-91616

    10/03/2019, 6:29 PM
    oh hm
  • w

    witty-crayon-22786

    10/03/2019, 6:29 PM
    might be a bug in 
    ConsoleRuleTestBase
    ... it might not be adding all of the rules required by the Goal
  • h

    hundreds-breakfast-49010

    10/03/2019, 6:30 PM
    do I need to add 
    optionable_rule(TestWorkspaceGoal)
    somewhere?
  • a

    aloof-angle-91616

    10/03/2019, 6:31 PM
    that's not done for 
    List
    , so i don't think so
  • a

    aloof-angle-91616

    10/03/2019, 6:31 PM
    i might try renaming the goal to something that doesn't have another goal as a prefix (i.e. remove the 
    test
    part from the 
    name
    ), not sure if there's a bug there
  • w

    witty-crayon-22786

    10/03/2019, 6:34 PM
    @hundreds-breakfast-49010: so, i think that 
    def rules(self)
    on testbase is not doing the thing that our actual registrar does
    h
    • 2
    • 28
  • h

    hundreds-breakfast-49010

    10/03/2019, 6:35 PM
    hm ok
  • h

    hundreds-breakfast-49010

    10/03/2019, 9:19 PM
    general question: when we talk about hermeticity as it pertains to pex, what specific guarantees does that entail?
  • a

    aloof-angle-91616

    10/03/2019, 9:20 PM
    i have a really interesting answer to that question which slightly revises the model that pex uses to resolve sources and requirements in a cacheable way that i plan to push tonight
  • a

    aloof-angle-91616

    10/03/2019, 9:20 PM
    but my general impression of "hermeticity" is that "the same command line produces a bit-for-bit reproducible output file"
  • a

    aloof-angle-91616

    10/03/2019, 9:22 PM
    i personally don't actually think "reproducibility" is necessary in a lot of cases because i don't think it's necessary to have coherent caching, but i'm not familiar with the reasons why that's necessary, and that's a minority opinion
1...657658659...906Latest