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

    happy-kitchen-89482

    08/24/2015, 9:02 PM
    And I agree with John's analysis of why this is different from the JVM subsystem, which models a running JVM, not the state of JDK installs on a machine.
  • e

    enough-analyst-54434

    08/24/2015, 9:03 PM
    Part of the help output confusion is a cake and eat it too thing
  • e

    enough-analyst-54434

    08/24/2015, 9:03 PM
    If the output did not include fully scoped flag names, it would be perhaps more clear the section header was the [scope]
  • e

    enough-analyst-54434

    08/24/2015, 9:04 PM
    And the flag was the exact option name
  • e

    enough-analyst-54434

    08/24/2015, 9:05 PM
    But I think you were an advocate for fully-scoped flag names Eric
  • e

    enough-analyst-54434

    08/24/2015, 9:05 PM
    There is certainly a tension there - useful for CLI, not useful for how do I pants.ini
  • a

    ancient-stone-50795

    08/24/2015, 9:09 PM
    Indeed I am the fully-scoped flag name champion. Well, I think I have made my peace with it. I’m not sure about how to glue together a common subsystem for both 
    Distribution
    and 
    JvmPlatform
    and making it obvious which subsystem to depend on. I’m going to put a little PR in to make the internal message we have when we forget to do that a bit clearer though.
  • w

    witty-crayon-22786

    08/24/2015, 9:13 PM
    usually namespaces are nested
  • w

    witty-crayon-22786

    08/24/2015, 9:13 PM
    in the case of subsystems, they are not
  • w

    witty-crayon-22786

    08/24/2015, 9:14 PM
    (ie, the dash versus dot issue above.)
  • h

    happy-kitchen-89482

    08/24/2015, 9:15 PM
    BTW I have that change that shows subsystem help inline with the scope help ready to go, just tweaking it a bit.
  • w

    witty-crayon-22786

    08/24/2015, 9:17 PM
    great, thanks.
  • a

    ancient-stone-50795

    08/24/2015, 9:17 PM
    @happy-kitchen-89482: were you aware that the doc generation needs to incorporate the list of subsystems? http://pantsbuild.github.io/options_reference.html
  • a

    ancient-stone-50795

    08/24/2015, 9:17 PM
  • e

    enough-analyst-54434

    08/24/2015, 9:20 PM
    It would be so awesome if @billowy-alarm-80641 took that one and drove it home!
  • a

    ancient-stone-50795

    08/24/2015, 9:20 PM
    Yeah, there’s already an example of what to do in the help printing
  • b

    billowy-alarm-80641

    08/24/2015, 9:21 PM
    Sorry I've been too busy trying to get on the same page w/ you in rb2650. 😐 (I'm gonna update that one again today, after I finish triaging this bug w/ ruby thrift gen...)
  • a

    ancient-stone-50795

    08/24/2015, 9:23 PM
    @billowy-alarm-80641: I encountered a similar issue in https://rbcommons.com/s/twitter/r/2687/ and just decided to remove these functions. I think 
    pants()
    pointers have been deprecated even longer than 
    with_artifact()
    and 
    with_sources()
  • b

    billowy-alarm-80641

    08/24/2015, 9:23 PM
    It's been a long, non-standard deprecation cycle for 
    pants()
  • j

    jolly-chef-92794

    08/24/2015, 9:23 PM
    So is the consensus that I should move the 
    —jdk-paths
    option into the 
    [distribution-locator]
    subsystem, and otherwise keep the patch as-is? (Also, should the namespace in that case be 
    distribution-locator
    , or 
    jvm-distribution
    , or 
    jvm-distribution-locator
    …)
  • b

    billowy-alarm-80641

    08/24/2015, 9:24 PM
    I think I'm going to announce an end-date via the standard 
    @deprecated()
    operator
  • b

    billowy-alarm-80641

    08/24/2015, 9:24 PM
    and fine an issue for how to design something that warns folks that do infrequent upgrades
  • b

    billowy-alarm-80641

    08/24/2015, 9:24 PM
    about this stuff
  • a

    ancient-stone-50795

    08/24/2015, 9:25 PM
    Which users still use 
    pants()
    pointers?
  • e

    enough-analyst-54434

    08/24/2015, 9:25 PM
    No one
  • b

    billowy-alarm-80641

    08/24/2015, 9:25 PM
    there are still some inside of Twitter
  • b

    billowy-alarm-80641

    08/24/2015, 9:25 PM
    in legacy repos that aren't our main monorepo
  • b

    billowy-alarm-80641

    08/24/2015, 9:25 PM
    😞
  • w

    witty-crayon-22786

    08/24/2015, 9:26 PM
    unsupported!
  • w

    witty-crayon-22786

    08/24/2015, 9:26 PM
    =P
1...212213214...2420Latest