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

    happy-kitchen-89482

    06/03/2015, 9:02 PM
    @mammoth-autumn-2551: subsystem options are just like task options: they have a scope (in this case python-setup) and you can set values on the cmd line, in pants.ini etc.
  • h

    happy-kitchen-89482

    06/03/2015, 9:02 PM
    There's nothing magical going on.
  • m

    mammoth-autumn-2551

    06/03/2015, 9:02 PM
    thanks @happy-kitchen-89482, I got it working and tested. The review board is up here: https://rbcommons.com/s/twitter/r/2320/
  • h

    happy-kitchen-89482

    06/03/2015, 9:02 PM
    And yes, underscores replace dashes
  • m

    mammoth-autumn-2551

    06/03/2015, 9:03 PM
    unfortunately, this patch does not apply cleanly on 0.0.32 so I’ll wait for the next release after this gets merged in
  • h

    happy-kitchen-89482

    06/03/2015, 9:03 PM
    Yep, I'll review at the other end of this 10 hour flight I'm about to take off on... Unless I have in-flight wifi, which seems unlikely
  • m

    mammoth-autumn-2551

    06/03/2015, 9:04 PM
    no rush from my end @happy-kitchen-89482, I have the hacked together version for us internally
  • m

    mammoth-autumn-2551

    06/03/2015, 9:04 PM
    enjoy your trip
  • a

    ancient-stone-50795

    06/04/2015, 7:49 PM
    Scala question. I'm looking at 
    ForkedJava.scala
    in sbt and it has this line:
    Copy code
    try {
        exitCode = Process(exe +: forkArgs, cwd) ! javacLogger
      } finally {
        javacLogger.flush(exitCode)
      }
  • a

    ancient-stone-50795

    06/04/2015, 7:51 PM
    What is the function of the ! operator in that statement? I'm looking at this because it seems that the stderr is not fully drained and passed into javacLogger.
  • a

    ancient-stone-50795

    06/04/2015, 7:51 PM
    At least, that is my theory.
  • p

    proud-alligator-68508

    06/04/2015, 10:08 PM
    !
    is often used (by convention) as the method name for message passing eg sending to an Actor
  • p

    proud-alligator-68508

    06/04/2015, 10:10 PM
    Looks like in this case, that is a method on ProcessBuilder:
    Copy code
    /**
   * Starts the process represented by this builder, blocks until it exits, and returns the exit code.  Standard output and error are
   * sent to the given ProcessLogger.
   */
  def !(log: ProcessLogger): Int
  • l

    limited-country-20626

    06/05/2015, 4:16 AM
    SO question about PEX if anyone wants to take a crack at it: http://stackoverflow.com/questions/30654854/python-pex-loading
  • a

    ancient-stone-50795

    06/05/2015, 1:35 PM
    @proud-alligator-68508: thanks for the answer. That all seems to be wrapped up in the Scala runtime. I've just tried rebuilding zinc with scala 2.11.6 and there is no output at all from the compiler from zinc after that change.
  • a

    ancient-stone-50795

    06/05/2015, 1:36 PM
    I was under the impression that Twitter had already moved over to zinc for compiling Java. I'm wondering if there is some other difference in their environment where they aren't seeing these problems.
  • e

    enough-analyst-54434

    06/06/2015, 2:54 AM
    @limited-country-20626: Thanks, gave a response a whirl
  • a

    ancient-stone-50795

    06/08/2015, 7:20 PM
    Where are we at for a public release? @bored-art-40741: You were interested and I am too. Do you have time to do it this week? If we want to do it this week I can do it.
  • b

    bored-art-40741

    06/08/2015, 7:32 PM
    I can take a shot at it this week as well, probably not until tomorrow or Wednesday though
  • m

    mammoth-autumn-2551

    06/08/2015, 7:36 PM
    Hey folks, does anyone have any general recommendations for dealing with general ivy frustrations? We frequently encounter (1.) really slow ivy resolutions and (2.) outright resolution failures even when the jars are definitely available
  • m

    mammoth-autumn-2551

    06/08/2015, 7:37 PM
    I should add that IMO our ivy configuration is pretty straightforward, its basically configured to point to a an artifactory instance that is acting as a proxy + cache to a number of different 3rdparty maven repos
  • b

    bored-art-40741

    06/08/2015, 7:37 PM
    Do you have local artifact caching turned on for ivy resolution?
  • b

    bored-art-40741

    06/08/2015, 7:37 PM
    Avoiding resolution even between cleans is generally helpful
  • m

    mammoth-autumn-2551

    06/08/2015, 7:38 PM
    I don’t know...
  • m

    mammoth-autumn-2551

    06/08/2015, 7:38 PM
    -.txt
    -.txt
  • m

    mammoth-autumn-2551

    06/08/2015, 7:38 PM
    is that a pants setting?
  • b

    bored-art-40741

    06/08/2015, 7:38 PM
    Local artifact caching is a pants thing, not an ivy thing
  • b

    bored-art-40741

    06/08/2015, 7:39 PM
    Copy code
    [resolve.ivy]
read_artifact_caches = %(local_artifact_caches)s
write_artifact_caches = %(local_artifact_caches)s
  • m

    mammoth-autumn-2551

    06/08/2015, 7:39 PM
    oh this sounds exciting!
  • b

    bored-art-40741

    06/08/2015, 7:39 PM
    Copy code
    local_artifact_caches: ['%(pants_local_artifact_cachedir)s']
1...456...2420Latest