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

    wide-energy-11069

    12/08/2016, 10:58 PM
    @happy-kitchen-89482 did you explicitly removed (#issue number) from commit message? https://github.com/pantsbuild/pants/commits/master
  • h

    happy-kitchen-89482

    12/08/2016, 10:59 PM
    Yes
  • w

    wide-energy-11069

    12/08/2016, 11:00 PM
    ok. i’m doing the release so it was a couple of more turns to make to find it. if you don’t mind, we can just leave it there by default
  • h

    happy-kitchen-89482

    12/08/2016, 11:00 PM
    sure
  • h

    happy-kitchen-89482

    12/08/2016, 11:00 PM
    If we want to codify that it should probably be in the workflow documentation
    👍 1
  • e

    enough-analyst-54434

    12/09/2016, 3:47 AM
    Apropos of almost nothing - pushd/popd behave differently on osx (old bash :/) and linux - osx pushd/popd's silently, modern bash echos the dirstack
  • e

    enough-analyst-54434

    12/09/2016, 3:50 AM
    I think twitter guidance on using pushd/popd vs a subshell with a single cd is misguided unless performance is critical - but then why bash? - a (cd $DIR; commands...) is good for all the reasons context managers, RAII, etc are good - plus no platform difference
  • r

    rough-minister-58256

    12/09/2016, 8:48 AM
    I think the general idea behind pushd/popd is mainly readability. it reads easier because you can indent subcommands in the context of pushd’d dirs and nest, e.g.
  • r

    rough-minister-58256

    12/09/2016, 8:48 AM
    Copy code
    TMP=$(mktmp)
pushd $TMP
  ..do some stuff..
  pushd $TMP/xyz
    ..do some stuff..
  popd
popd
  • r

    rough-minister-58256

    12/09/2016, 8:48 AM
    reads much like a context mgr etc and the nesting makes it “clearer” you’re doing something in the context of another dir
  • r

    rough-minister-58256

    12/09/2016, 8:48 AM
    but you raise perfectly valid points and afaik there’s no formal guidance around that - just tribal
  • r

    rough-minister-58256

    12/09/2016, 8:49 AM
    iirc a wickman thing
  • h

    happy-kitchen-89482

    12/09/2016, 3:03 PM
    @wide-energy-11069 You've already cut the release?
  • b

    billions-airport-86951

    12/09/2016, 3:07 PM
    Is it possible to get the publish goal to additionally publish jvm_apps? I've tried:
    Copy code
    publish_extras: {
    'deployable_archives': {
      'extension': 'tar.gz',
    },
  }
  • w

    wide-energy-11069

    12/09/2016, 5:33 PM
    @happy-kitchen-89482: yeah. waiting on the doc for the announcement. Will send a note next time if cut early.
  • f

    flaky-twilight-29457

    12/09/2016, 5:41 PM
    is there an option to run the test suite that the travis ci runs on my box to validate my changes?
  • f

    flaky-twilight-29457

    12/09/2016, 5:42 PM
    I lack sufficiently complex builds to trigger the malfunction I currently see
  • f

    flaky-twilight-29457

    12/09/2016, 6:03 PM
    nice, found them. I guess they were too obviously located xD
  • w

    witty-crayon-22786

    12/09/2016, 6:05 PM
    @billions-airport-86951 : you might be able to write something to do that... but deployment/bundling is generally not really compatible with publishing
  • w

    witty-crayon-22786

    12/09/2016, 6:05 PM
    in particular, publishing is broken up into as many jars as you have targets, which are then fetched using ivy
  • w

    witty-crayon-22786

    12/09/2016, 6:05 PM
    whereas bundling gives you everything in one archive
  • w

    witty-crayon-22786

    12/09/2016, 6:06 PM
    so the output of 
    ./pants bundle $jvm_app
    is generally what you'd deploy
  • b

    billions-airport-86951

    12/09/2016, 6:06 PM
    @witty-crayon-22786 : we thought that might be the case; we're building tar.gz archives containing a dockerfile which we would like to push to jfrog along with the artifact releases
  • w

    witty-crayon-22786

    12/09/2016, 6:08 PM
    ok
  • w

    witty-crayon-22786

    12/09/2016, 6:08 PM
    should be possible to do with the publish extras i guess
  • w

    witty-crayon-22786

    12/09/2016, 6:08 PM
  • w

    witty-crayon-22786

    12/09/2016, 6:09 PM
    ...oh, i see what you were trying to do... hm.
  • w

    witty-crayon-22786

    12/09/2016, 6:10 PM
    so you have a 
    provides
    clause on the 
    jvm_app
    ? never seen that before
  • w

    witty-crayon-22786

    12/09/2016, 6:11 PM
    ...but yea, either way. that's going to be a little bit strange. it will start by trying to publish all of the transitive dependencies of the jvm_app
  • b

    billions-airport-86951

    12/09/2016, 6:12 PM
    no... will try adding one
1...156915701571...2420Latest