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

    enough-analyst-54434

    07/31/2015, 5:49 PM
    large is safe
  • w

    witty-crayon-22786

    07/31/2015, 5:49 PM
    and if you compile scala, consider modeling your build after 
    pants.ini.isolated
    in the repo
  • m

    mammoth-autumn-2551

    07/31/2015, 5:49 PM
    we do not
  • e

    enough-analyst-54434

    07/31/2015, 5:49 PM
    This is Ugo country
  • e

    enough-analyst-54434

    07/31/2015, 5:49 PM
    they would never
  • w

    witty-crayon-22786

    07/31/2015, 5:49 PM
    well, java too 😃
  • e

    enough-analyst-54434

    07/31/2015, 5:50 PM
    Somehow Twitter whipped him into that, not sure how
  • w

    witty-crayon-22786

    07/31/2015, 5:51 PM
    as an fyi: we use zinc-java to compile ~2mm lines of java. we've found two issues, and they're both patched in upstream zinc... waiting on a release
  • w

    witty-crayon-22786

    07/31/2015, 5:51 PM
    advantages: isolation, parallelism
  • w

    witty-crayon-22786

    07/31/2015, 5:52 PM
    still nailing down whether it's faster... it was mostly just necessary in our case due to the language mix
  • e

    enough-analyst-54434

    07/31/2015, 5:54 PM
    @bored-art-40741: review is up here fwiw on the OSX CI failure thing @ancient-stone-50795 found: https://rbcommons.com/s/twitter/r/2551/
  • a

    ancient-stone-50795

    07/31/2015, 5:59 PM
    I can’t reproduce the issue any more in the currently committed master (I haven’t patched in 2551) I’m running the idea_integration test a couple more times.
  • e

    enough-analyst-54434

    07/31/2015, 5:59 PM
    k
  • e

    enough-analyst-54434

    07/31/2015, 6:00 PM
    There was a bug, diff 2 just pushed
  • e

    enough-analyst-54434

    07/31/2015, 6:00 PM
    You could still get a NSP exception from the 1st diff
  • e

    enough-analyst-54434

    07/31/2015, 6:02 PM
    ...now the _as_process call is guarded b its only 2 internal callers
  • r

    rough-minister-58256

    07/31/2015, 6:03 PM
    apologies for the breakage gents, clearly need to get some more tests in place around this stuff - thanks for the quick fixes John
  • e

    enough-analyst-54434

    07/31/2015, 6:03 PM
    no worrries and no problem
  • e

    enough-analyst-54434

    07/31/2015, 6:03 PM
    The biggest issue is the tests are mock I think
  • e

    enough-analyst-54434

    07/31/2015, 6:03 PM
    But its hard to run a controlled non-mock test against pids and the like
  • w

    witty-crayon-22786

    07/31/2015, 6:04 PM
  • w

    witty-crayon-22786

    07/31/2015, 6:04 PM
    (python moreso)
  • a

    ancient-stone-50795

    07/31/2015, 6:04 PM
    @enough-analyst-54434: RE:https://rbcommons.com/s/twitter/r/2551/ just a doc comment, LGTM.
  • e

    enough-analyst-54434

    07/31/2015, 6:07 PM
    @witty-crayon-22786: that approach doesn't help much here since the only injectable interface would be a psutil.Process factory, and either that would need to return a hand-made mock or else, just use mocks
  • r

    rough-minister-58256

    07/31/2015, 6:07 PM
    yeah, will have to think that through some more
  • w

    witty-crayon-22786

    07/31/2015, 6:08 PM
    yea, doesn't help when your libraries aren't designed for testing
  • e

    enough-analyst-54434

    07/31/2015, 6:08 PM
    Yeah and a native bridge sortof never will be
  • e

    enough-analyst-54434

    07/31/2015, 6:09 PM
    This is a pretty special case of actually hard to test for real
  • m

    mammoth-autumn-2551

    07/31/2015, 6:16 PM
    Well thanks to your help today @enough-analyst-54434 and @ancient-stone-50795 , we’ve shaved off 6.5 minutes from our deploy. Where do I send the beer?
  • m

    mammoth-autumn-2551

    07/31/2015, 6:17 PM
    Any chance anyone has the appetite to give me an overview on the pants artifact cache, and if / how I should use it?
1...117118119...2420Latest