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

    enough-analyst-54434

    05/13/2016, 5:17 PM
    yup, has an api for it
  • f

    full-kilobyte-93429

    05/13/2016, 5:18 PM
    cool, will take a look
  • e

    enough-analyst-54434

    05/13/2016, 5:19 PM
  • e

    enough-analyst-54434

    05/13/2016, 5:19 PM
    SHould be handled already
  • e

    enough-analyst-54434

    05/13/2016, 5:20 PM
    Aha - may be bug in _add_entry being handed dir
  • f

    full-kilobyte-93429

    05/13/2016, 5:20 PM
    like this pretty standard...
    Copy code
    Manifest-Version: 1.0
Main-Class: com.twitter.scalding.Tool
  • e

    enough-analyst-54434

    05/13/2016, 5:20 PM
    it does not look inside to check each file
  • e

    enough-analyst-54434

    05/13/2016, 5:20 PM
    By standard I mean path and not contents
  • e

    enough-analyst-54434

    05/13/2016, 5:20 PM
    The python side does not look at contents
  • e

    enough-analyst-54434

    05/13/2016, 5:21 PM
    So - that's definitely what the bug is - the manifest detection only works when adding files, not dirs
  • c

    cuddly-garage-87447

    05/13/2016, 5:24 PM
    @witty-crayon-22786: BTW what would you think of that: https://github.com/pantsbuild/intellij-pants-plugin/issues/147
  • c

    cuddly-garage-87447

    05/13/2016, 5:24 PM
    it could help with long path
  • w

    witty-crayon-22786

    05/13/2016, 5:25 PM
    could be neat... i don't use ij though
  • c

    cuddly-garage-87447

    05/13/2016, 5:26 PM
    ok, well if people using intellij are ok to integrate that I'll give it a try
  • e

    enough-analyst-54434

    05/13/2016, 5:31 PM
    @full-kilobyte-93429: I can hop on this fix unless you are ready to run with it
  • e

    enough-analyst-54434

    05/13/2016, 5:31 PM
    ... that said - why is a custom manifest used for that? why not have a 
    jvm_binary
    target?
  • e

    enough-analyst-54434

    05/13/2016, 5:32 PM
    Its not even a fancy manifest
  • e

    enough-analyst-54434

    05/13/2016, 5:33 PM
    I can make 0 sense of the example case
  • e

    enough-analyst-54434

    05/13/2016, 5:36 PM
    Commented on the issue, grabbed it
  • f

    full-kilobyte-93429

    05/13/2016, 5:49 PM
    @enough-analyst-54434: thanks! the team somehow needs two entry points one from 
    jvm_binary
    , the other from this, i can clarify with them if alternative is possible
  • f

    full-kilobyte-93429

    05/13/2016, 5:50 PM
    so maybe no need working on until i hear from them
  • e

    enough-analyst-54434

    05/13/2016, 6:03 PM
    well, it is in fact a bug
  • e

    enough-analyst-54434

    05/13/2016, 6:03 PM
    They should just have never been affected with that particular manifest unless I'm missing something
  • l

    limited-country-20626

    05/13/2016, 6:14 PM
    as an update on buildgen since I saw it mentioned - I will be releasing an update this weekend which should make it fully consumable for outside repos - with the caveat that there is likely some work to be done to play nicely with codegen that we don't use internally. I will email the list when it is out, with some sort of accompanying docs.
  • b

    bored-art-40741

    05/13/2016, 6:21 PM
    Wait, @witty-crayon-22786, what do you mean release validation thread?
  • b

    bored-art-40741

    05/13/2016, 6:21 PM
    My understanding is that we're only "validating" RCs
  • b

    bored-art-40741

    05/13/2016, 6:21 PM
    Hence the "pre" syntax
  • w

    witty-crayon-22786

    05/13/2016, 6:22 PM
    @bored-art-40741: ... rc == pre from my perspective
  • w

    witty-crayon-22786

    05/13/2016, 6:22 PM
    maybe i missed something.
  • b

    bored-art-40741

    05/13/2016, 6:22 PM
    I believe we specifically looked for something "not rc" because this wasn't considered a release candidate
1...129312941295...2420Latest