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

    enough-analyst-54434

    10/22/2015, 10:14 PM
    jarring wants self, java wants direct, scala needs transitive
  • e

    enough-analyst-54434

    10/22/2015, 10:14 PM
    like that?
  • w

    witty-crayon-22786

    10/22/2015, 10:14 PM
    @enough-analyst-54434: ok, next question... should i just kill this parameter, and move the definition of how to walk out to build_graph methods?
  • w

    witty-crayon-22786

    10/22/2015, 10:15 PM
    yep, like that
  • w

    witty-crayon-22786

    10/22/2015, 10:15 PM
    (runtime also transitive)
  • e

    enough-analyst-54434

    10/22/2015, 10:15 PM
    I think runtime is a maven swear word
  • e

    enough-analyst-54434

    10/22/2015, 10:15 PM
    not really a valid idea sortof!
  • w

    witty-crayon-22786

    10/22/2015, 10:15 PM
    er
  • w

    witty-crayon-22786

    10/22/2015, 10:16 PM
    you're probably joking?
  • e

    enough-analyst-54434

    10/22/2015, 10:16 PM
    no
  • w

    witty-crayon-22786

    10/22/2015, 10:16 PM
    heh, ok. in any case, probably off topic
  • e

    enough-analyst-54434

    10/22/2015, 10:17 PM
    runtime may not be needed to compile, but they should not interfere
  • e

    enough-analyst-54434

    10/22/2015, 10:17 PM
    for things where they do those things are broken
  • w

    witty-crayon-22786

    10/22/2015, 10:17 PM
    yea, absolutely. but if i want to run a binary, i need my transitive classpath
  • w

    witty-crayon-22786

    10/22/2015, 10:17 PM
    would it be feasible to move the walk mode into build_graph
  • e

    enough-analyst-54434

    10/22/2015, 10:17 PM
    In terms of build graph - not sure, self is self, direct is *.dependencies
  • e

    enough-analyst-54434

    10/22/2015, 10:18 PM
    so not like you need a method for unifying those with walk unless?...
  • e

    enough-analyst-54434

    10/22/2015, 10:18 PM
    Can you expand on that?
  • w

    witty-crayon-22786

    10/22/2015, 10:18 PM
    transitive is .closure
  • w

    witty-crayon-22786

    10/22/2015, 10:18 PM
    i see
  • w

    witty-crayon-22786

    10/22/2015, 10:18 PM
    yea, so this just never should have been a parameter here.
  • w

    witty-crayon-22786

    10/22/2015, 10:20 PM
    the only catch is that you need bread-first search to get the right classpath behaviour
  • w

    witty-crayon-22786

    10/22/2015, 10:20 PM
    and closure() does dfs-ish
  • e

    enough-analyst-54434

    10/22/2015, 10:20 PM
    eh
  • e

    enough-analyst-54434

    10/22/2015, 10:20 PM
    s/right/maven/
  • e

    enough-analyst-54434

    10/22/2015, 10:20 PM
    classpath order should never matter
  • e

    enough-analyst-54434

    10/22/2015, 10:20 PM
    this is a crutch
  • e

    enough-analyst-54434

    10/22/2015, 10:20 PM
    though a real need
  • w

    witty-crayon-22786

    10/22/2015, 10:20 PM
    yes.
  • e

    enough-analyst-54434

    10/22/2015, 10:21 PM
    I think cp randomization as a service would be a good feature
1...583584585...2420Latest