https://pantsbuild.org/ logo
Powered by
#general
  • b

    bored-art-40741

    01/28/2016, 6:27 PM
    I guess we really don't care about ivy caching, but when we did use ivy we certainly had caching turned on and it worked fine
  • w

    witty-crayon-22786

    01/28/2016, 6:28 PM
    it works fine locally, but not remotely
  • w

    witty-crayon-22786

    01/28/2016, 6:28 PM
    "works" is strange though
  • w

    witty-crayon-22786

    01/28/2016, 6:28 PM
    it doesn't actually do anything useful
  • b

    bored-art-40741

    01/28/2016, 6:28 PM
    Are you saying we can get local shading caching to work?
  • w

    witty-crayon-22786

    01/28/2016, 6:28 PM
    yep
  • b

    bored-art-40741

    01/28/2016, 6:28 PM
    That would be a significant improvement over no caching
  • w

    witty-crayon-22786

    01/28/2016, 6:28 PM
    it's perfectly relocatable
  • w

    witty-crayon-22786

    01/28/2016, 6:28 PM
    see explanation in the comment
  • w

    witty-crayon-22786

    01/28/2016, 6:29 PM
    the mixin is the thing that does the unsafe caching
  • b

    bored-art-40741

    01/28/2016, 6:29 PM
    Oh I missed the last part
  • w

    witty-crayon-22786

    01/28/2016, 6:29 PM
    the core task is fine
  • b

    bored-art-40741

    01/28/2016, 6:30 PM
    Okay, yeah, I'm fine with that. We don't care about Ivy caching at all, but we really care about tool bootstrapping being cached
  • w

    witty-crayon-22786

    01/28/2016, 6:30 PM
    yep
  • a

    ancient-stone-50795

    01/28/2016, 6:46 PM
    @witty-crayon-22786: am I going to get backlash for adding a --no-publish-jar-prompt to get rid of piping 'yes' into my pants command?
  • w

    witty-crayon-22786

    01/28/2016, 6:46 PM
    nope
  • a

    ancient-stone-50795

    01/28/2016, 6:47 PM
    yes | pants is very hard to debug with the interactive debugger...
  • w

    wide-energy-11069

    01/28/2016, 6:48 PM
    @ancient-stone-50795: i’ve done it before with intellij (something messed up recently, so haven’t got the chance to fix it), but its totally doable
  • w

    wide-energy-11069

    01/28/2016, 6:49 PM
    i can send out a doc once i figured that out
  • a

    ancient-stone-50795

    01/28/2016, 7:00 PM
  • a

    ancient-stone-50795

    01/28/2016, 7:00 PM
    just adds an option to turn off the prompt.
  • w

    witty-crayon-22786

    01/28/2016, 7:22 PM
  • j

    jolly-chef-92794

    01/28/2016, 7:27 PM
    Is this if-statement block necessary? Can I delete it? https://github.com/pantsbuild/pants/blob/master/src/python/pants/backend/jvm/ivy_utils.py#L559 We’re having a weird problem where ivy is resolving a 
    test-jar
    instead of a 
    jar
    for a dependency that contains both, with a jar dependency that looks like
    Copy code
    jar('org.apache.hadoop', 'hadoop-common', rev='2.0.0-cdh4.6.0', ext='jar', classifier=None)
    (the version is in our internal nexus) Commenting out that block correctly resolves the 
    jar
    instead of the 
    test-jar
    .
  • j

    jolly-chef-92794

    01/28/2016, 7:28 PM
    ( @enough-analyst-54434 @fancy-queen-20734 @ancient-stone-50795 )
  • b

    bored-art-40741

    01/28/2016, 7:37 PM
    So I'm working on this ivy resolve dir isolation problem, and after struggling with the ivy integration tests a bit I realized that if we proceed with my change, then we can't use the default ivysettings.xml, it will always have to be a custom one that we set.
  • b

    bored-art-40741

    01/28/2016, 7:37 PM
    How do people feel about this?
  • b

    bored-art-40741

    01/28/2016, 7:37 PM
    It seems like something we kind of sort of want to do anyway
  • w

    witty-crayon-22786

    01/28/2016, 7:37 PM
    we have a lot of customization in there.
  • w

    witty-crayon-22786

    01/28/2016, 7:37 PM
    (in ours)
  • b

    bored-art-40741

    01/28/2016, 7:37 PM
    More specifically
1...933934935...2420Latest