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

    curved-manchester-66006

    04/29/2016, 5:40 PM
    It didn't make it into jenkins 2.0 but there was some discussion about actually deleting bad/clunky/broken api methods used by plugins to instead of leaving them deprecated forever
  • w

    witty-crayon-22786

    04/29/2016, 5:50 PM
    Copy code
    $ ./pants repro
...
               Unknown goal(s): repro
  • w

    witty-crayon-22786

    04/29/2016, 5:50 PM
    Copy code
    $ ./pants help repro

repro options:

--repro-capture=<repro_path>
  • h

    happy-kitchen-89482

    04/29/2016, 6:03 PM
    It’s a subsystem
  • h

    happy-kitchen-89482

    04/29/2016, 6:03 PM
    not a task
  • w

    witty-crayon-22786

    04/29/2016, 6:15 PM
    hah. so i fell for exactly the thing i was sure wasn't dangerous, and everyone else was warning against a few summits ago, eh?
  • e

    enough-analyst-54434

    04/29/2016, 6:17 PM
    So you guessed this was a goal not from 
    ./pants goals
    but 
    ./pants --help-all
    or 
    ./pants options
    or something?
  • e

    enough-analyst-54434

    04/29/2016, 6:18 PM
    Thanks @bored-art-40741 - that does seem to have done the trick
  • w

    witty-crayon-22786

    04/29/2016, 6:18 PM
    my memory was that it was a goal, yea
  • b

    bored-art-40741

    04/29/2016, 6:18 PM
  • b

    bored-art-40741

    04/29/2016, 6:19 PM
    Thanks for cutting into the Jenkins failures
  • e

    enough-analyst-54434

    04/29/2016, 6:20 PM
    NP. I can push further or back over to you, whichever works
  • b

    bored-art-40741

    04/29/2016, 6:20 PM
    I thought of a quick and easy model for getting the benefit of restarting individual shards, let me know what you think
  • b

    bored-art-40741

    04/29/2016, 6:20 PM
    Basically the idea is that for a given (SHA, shard_id) pair, you cache success and short-circuit if you've seen that it has already succeeded
  • b

    bored-art-40741

    04/29/2016, 6:21 PM
    So you always restart the whole build, but you only actually trigger sub-shards that haven't passed for your current SHA
  • e

    enough-analyst-54434

    04/29/2016, 6:21 PM
    ok - the store would be what?
  • b

    bored-art-40741

    04/29/2016, 6:21 PM
    I think that's feature parity with Travis even if it isn't the most obvious route
  • b

    bored-art-40741

    04/29/2016, 6:21 PM
    No storage, it's stateless. Or rather, the storage is inspecting Jenkins' job history
  • e

    enough-analyst-54434

    04/29/2016, 6:21 PM
    oh - master
  • e

    enough-analyst-54434

    04/29/2016, 6:21 PM
    yeah
  • e

    enough-analyst-54434

    04/29/2016, 6:22 PM
    That would work
  • e

    enough-analyst-54434

    04/29/2016, 6:22 PM
    That sounds like a tweak though
  • e

    enough-analyst-54434

    04/29/2016, 6:22 PM
    We could cut over before it exists
  • b

    bored-art-40741

    04/29/2016, 6:22 PM
    Agreed
  • b

    bored-art-40741

    04/29/2016, 6:23 PM
    I think having restarts is important, but I think high stability builds on Jenkins is more important
  • e

    enough-analyst-54434

    04/29/2016, 6:23 PM
    The biggest nasty right now is the mac
  • e

    enough-analyst-54434

    04/29/2016, 6:23 PM
    I'd like to only use it for master commits
  • e

    enough-analyst-54434

    04/29/2016, 6:23 PM
    not PRs
  • e

    enough-analyst-54434

    04/29/2016, 6:23 PM
    Its - I'm not sure - 90% of the time
  • b

    bored-art-40741

    04/29/2016, 6:23 PM
    It's what 90% of the time?
1...122212231224...2420Latest