https://pantsbuild.org/ logo
Powered by
#pex
  • r

    rough-minister-58256

    06/18/2018, 8:21 PM
    at Twitter we usually deploy to Mesos/Aurora - no Docker
  • r

    rough-minister-58256

    06/18/2018, 8:23 PM
    pex provides python environmental isolation in that case and Mesos/Aurora provides resource isolation akin to what most ppl would get w docker alone
  • d

    dazzling-florist-39383

    06/18/2018, 8:24 PM
    yeah, i did figure it was p basic
  • d

    dazzling-florist-39383

    06/18/2018, 8:24 PM
    but figured it wouldn’t hurt to ask
  • d

    dazzling-florist-39383

    06/18/2018, 8:24 PM
    thanks
  • r

    rough-minister-58256

    06/18/2018, 8:28 PM
  • d

    dazzling-florist-39383

    06/18/2018, 8:41 PM
    The Pinterest one is what got me working on our monorepo
  • d

    dazzling-florist-39383

    06/18/2018, 8:41 PM
    Time to reread I guess
  • d

    dazzling-florist-39383

    06/18/2018, 8:42 PM
    :-)
  • h

    happy-kitchen-89482

    06/19/2018, 4:17 PM
    I didn’t even know Pinterest are using Pants
  • d

    dazzling-florist-39383

    06/19/2018, 4:23 PM
    yeah, i think they’re using a mix
  • d

    dazzling-florist-39383

    06/19/2018, 4:23 PM
    but this article was what got me re-evaluating pants as a solution internally, and it’s looking more like we may adopt it. definitely adopting pex one way or naother
    👏 2
  • a

    aloof-angle-91616

    07/05/2018, 7:35 PM
    @enough-analyst-54434 there's a CI failure in my PR https://github.com/pantsbuild/pants/pull/6022 in the osx shard where pex is failing to resolve locally-built dists in 
    PythonRun
    only, with `Failed to execute PEX file, missing macosx_10_12_x86_64-cp-27-cp27m compatible dependencies for: my local dist`: see https://travis-ci.org/pantsbuild/pants/jobs/400569783. this does not repro on my osx laptop. in 
    SetupPyRunner
    we're tagging the locally-built dist with 
    --plat-name
    set to 
    get_local_platform()
    if there are native sources. 
    get_local_platform()
    has a TODO linking to this issue: https://github.com/pantsbuild/pex/issues/511, and it seems reasonable to think these are related, but it's not clear to me what would cause this resolution failure.
    e
    • 2
    • 90
  • r

    rough-minister-58256

    07/06/2018, 1:51 AM
    fasthello-1.0.0%2Bb2746123820ccd47afb2e520476673e3aea8a39a.defaultfingerprintstrategy.da39a3ee5e6b.897870053801-cp27-cp27m-macosx_10_12_x86_64.whl
  • r

    rough-minister-58256

    07/06/2018, 1:52 AM
    that 
    %2B
    looks like an incorrectly urlencoded 
    +
    in whl filename to me, likely breaking resolution of the bdist
  • r

    rough-minister-58256

    07/06/2018, 1:53 AM
    but thats just in the pex debug logging - hard to say what the actual filesystem looks like in a travis run.
  • a

    aloof-angle-91616

    07/06/2018, 2:50 AM
    ok see i had that thought but i thought it was a crazy thought
  • a

    aloof-angle-91616

    07/06/2018, 2:51 AM
    i don't know if i'm seeing that locally (which might explain the disparity), i will check
  • a

    aloof-angle-91616

    07/06/2018, 2:53 AM
    i am but that's just more info
  • a

    aloof-angle-91616

    07/06/2018, 3:09 AM
    PEX_PYTHON_PATH
    overriding 
    --python-setup-interpreter-search-paths
    seems weird to me, it wasn't clear that it was because 
    /etc/pexrc
    was set -- i would assume that the user would want the build tool to set the path when the build tool is invoking pex (as opposed to uses of pex on the command line), and if i didn't have root access to this machine i wouldn't be able to change what python is used whatsoever
  • a

    aloof-angle-91616

    07/06/2018, 3:10 AM
    haha, once i figured that out i immediately found it was indeed a 
    2.7.10
    thing because it reproes without even trying, unbelievable
  • a

    aloof-angle-91616

    07/06/2018, 3:11 AM
    this is all i've ever wanted so now i can prob fix that in a jiffy
  • a

    aloof-angle-91616

    07/06/2018, 3:12 AM
    thanks @rough-minister-58256 and @enough-analyst-54434 for help itc today
  • a

    aloof-angle-91616

    07/06/2018, 3:13 AM
    ...and i spoke too soon. still progress (EDIT: i didn't speak that soon)
  • r

    rough-minister-58256

    07/06/2018, 3:59 AM
    PEX_PYTHON_PATH
    should never get you to a 2.7.10 interpreter?
  • r

    rough-minister-58256

    07/06/2018, 4:00 AM
    but yeah, its like that for runtime interpreter symmetry. once pants creates the pex, we still manage which python it executes against at the system level. so we also use the exact same configuration at build time so that everything is consistent.
  • r

    rough-minister-58256

    07/06/2018, 4:01 AM
    basically, we consider the interpreter as a “managed” runtime component that no1 should ever have to think explicitly about (except for maybe pants devs.. heh)
  • r

    rough-minister-58256

    07/06/2018, 4:57 AM
    ah, I see: it works locally w/ 
    P_P_P
    but not in e.g. travis
  • a

    aloof-angle-91616

    07/06/2018, 6:37 AM
    oh, see i wanted it to go to a 2.7.10 interpreter so i could repro this failure
  • a

    aloof-angle-91616

    07/06/2018, 6:38 AM
    i'm still seeing it checking out master locally which is literally incorrect because master has passed this osx shard with 2.7.10 so keeping that in mind before seeing which cache makes it work again
1...495051...74Latest