#general
- w
witty-crayon-22786
03/14/2017, 5:06 PM(if you wanted to use something other than
)1.6 - p
purple-leather-29803
03/14/2017, 5:07 PMi’d like to publish a SNAPSHOT-esque version to my local maven repo
, for some iteration with the sbt-assembly plugin’s use of jarjar. is there a quick incantation for that or should i read through these docs to figure it out?~/.m2 - w
witty-crayon-22786
03/14/2017, 5:08 PM
should do it./pants publish.jar --no-dryrun --local=~/.m2/repository src/main/java/org/pantsbuild/jarjar:👍 1 - p
purple-leather-29803
03/14/2017, 5:28 PMcool, thanks! - h
happy-kitchen-89482
03/15/2017, 5:07 AMDoes anyone have any novel strategies for invalidating products of in-repo custom tasks when the task code itself changes? Beyond bumping the
?implementation_version() - h
happy-kitchen-89482
03/15/2017, 5:07 AMTwo options I’ve been oscillating between: - h
happy-kitchen-89482
03/15/2017, 5:07 AM1) A custom fingerprint strategy that mixes in the fprints of the relevant tasks - h
happy-kitchen-89482
03/15/2017, 5:08 AM2) Having compile-scope dependencies from the targets being acted on to the targets of the custom tasks - h
happy-kitchen-89482
03/15/2017, 5:08 AMBut both are somewhat clunky - b
billions-airport-86951
03/15/2017, 1:50 PM@purple-oil-45048 I tried running Pants under the Windows 10 Subsystem for Linux thing and it worked quite well (I didn't use it for long mind; other things caused me to switch to an Ubuntu install)✨ 1 - p
purple-oil-45048
03/15/2017, 3:40 PMThanks @billions-airport-86951 ! My colleagues are stuck with Windows 7 .. will have to think a bit more how to best approach this - r
rough-minister-58256
03/15/2017, 8:54 PMwelp, looks like travis is fubar - f
full-kilobyte-93429
03/15/2017, 8:55 PMi am fixing a checkstyle issue from my commit - r
rough-minister-58256
03/15/2017, 8:56 PMlooks like they’re having network issues - seeing things like unable to connect to host for things like the container boostrap - f
full-kilobyte-93429
03/15/2017, 8:57 PMok, correction on mine: not checkstyle but real import issue from merging two consecutive commits that are in race - w
witty-crayon-22786
03/15/2017, 8:58 PM@full-kilobyte-93429 : the warning? - w
witty-crayon-22786
03/15/2017, 8:58 PMwe can fix that in a followup... i've fixed it in my branch - f
full-kilobyte-93429
03/15/2017, 8:58 PMno error - w
witty-crayon-22786
03/15/2017, 8:58 PMoh - f
full-kilobyte-93429
03/15/2017, 8:58 PMCopy codeF811:ERROR tests/python/pants_test/backend/python/tasks2/test_select_interpreter.py:016 redefinition of unused 'PythonSetup' from line 12 |from pants.python.python_setup import PythonSetup - w
witty-crayon-22786
03/15/2017, 8:58 PMhm. that looks more like benjy's change? - f
full-kilobyte-93429
03/15/2017, 8:59 PMyea, mine switched the order due to isort - f
full-kilobyte-93429
03/15/2017, 8:59 PMand benji's change removed the line i moved - f
full-kilobyte-93429
03/15/2017, 8:59 PMmerge was clean - w
witty-crayon-22786
03/15/2017, 9:00 PMfunky - f
full-kilobyte-93429
03/15/2017, 9:00 PMi didn't get his change in my last merge so we both committed - f
full-kilobyte-93429
03/15/2017, 9:00 PMgithub consider the merge is fine - f
full-kilobyte-93429
03/15/2017, 9:04 PM - w
witty-crayon-22786
03/15/2017, 9:06 PMshipit - h
happy-kitchen-89482
03/16/2017, 3:27 AMSay a task T can produce two products, A and B. And there is an upstream task TA that requires product A and an upstream task TB that requires product B. Is there a way for T to know which of the two products is actually required in a given run and only produce that one?