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

    rough-minister-58256

    06/23/2015, 8:11 PM
    maybe not git in all cases tho. hrmph.
  • r

    rough-minister-58256

    06/23/2015, 8:12 PM
  • w

    witty-crayon-22786

    06/23/2015, 8:13 PM
    oh, you're right.
  • w

    witty-crayon-22786

    06/23/2015, 8:18 PM
    @rough-minister-58256: http://stackoverflow.com/a/943537 =(
  • a

    ancient-stone-50795

    06/23/2015, 8:19 PM
    @witty-crayon-22786: I think that could happen, but the main point is to get a 1:1 mapping between a JarDependency and an actual jar.
  • a

    ancient-stone-50795

    06/23/2015, 8:19 PM
    (Regarding synthetic targets to represent the ivy graph)
  • w

    witty-crayon-22786

    06/23/2015, 8:20 PM
    @ancient-stone-50795: yea, makes sense. just a thought while we're in there.
  • r

    rough-minister-58256

    06/23/2015, 8:20 PM
    lovely.
  • w

    witty-crayon-22786

    06/23/2015, 8:21 PM
    @ancient-stone-50795: one reason to think about it now perhaps is that the resolved graph would be a native concept
  • w

    witty-crayon-22786

    06/23/2015, 8:21 PM
    so if you wanted to completely embed the resolved graph in your build files, you could just do that
  • w

    witty-crayon-22786

    06/23/2015, 8:22 PM
    er, native is the wrong word.
  • w

    witty-crayon-22786

    06/23/2015, 8:22 PM
    but my point is that jar_dependency could become a target type, perhaps.
  • w

    witty-crayon-22786

    06/23/2015, 8:23 PM
    or something other equivalent concept.
  • a

    ancient-stone-50795

    06/23/2015, 8:23 PM
    ok
  • w

    witty-crayon-22786

    06/23/2015, 8:25 PM
    foursquare wrote a replacement ivy resolver. so what if all we used ivy to do was turn 
    jar_library
    targets into a fully resolved graph of synthetic targets, but didn't use it for fetching
  • w

    witty-crayon-22786

    06/23/2015, 8:25 PM
    then we'd have a synthetic graph of explicit intransitive stuff
  • w

    witty-crayon-22786

    06/23/2015, 8:25 PM
    and the pure foursquare's fetcher (sorry, resolver is the wrong word) would consume that graph
  • w

    witty-crayon-22786

    06/23/2015, 8:25 PM
    s/pure//
  • p

    proud-alligator-68508

    06/23/2015, 8:48 PM
    pom resolve does do some resolving, fwiw
  • p

    proud-alligator-68508

    06/23/2015, 8:48 PM
    finds intersections of transitive reqs
  • w

    witty-crayon-22786

    06/23/2015, 8:49 PM
    oh. i thought you were declaring everything explicitly?
  • p

    proud-alligator-68508

    06/23/2015, 8:49 PM
    it just throws its hands up on conflicts
  • p

    proud-alligator-68508

    06/23/2015, 8:49 PM
    and requires explicit help
  • w

    witty-crayon-22786

    06/23/2015, 8:49 PM
    hm. but you can ask ivy to do that?
  • p

    proud-alligator-68508

    06/23/2015, 8:49 PM
    we’re also requiring everything to be globally compatible
  • p

    proud-alligator-68508

    06/23/2015, 8:49 PM
    so we can union subsets for free
  • p

    proud-alligator-68508

    06/23/2015, 8:50 PM
    eg we never pick a different version of foo depending on what we resolving for
  • w

    witty-crayon-22786

    06/23/2015, 8:51 PM
  • w

    witty-crayon-22786

    06/23/2015, 8:51 PM
    strict should do the same thing
  • p

    proud-alligator-68508

    06/23/2015, 8:53 PM
    IIRC it is broken in some way
1...192021...2420Latest