#general
- b
billowy-minister-38128
10/29/2016, 6:25 PMMore specifically, this happens when trying to build the pex. Seems we are missing a wheel, but lowercase socketserver is a Python 3 convention. - b
billowy-minister-38128
10/29/2016, 6:26 PMWill create issue on Monday - e
enough-analyst-54434
10/29/2016, 6:39 PMSounds good. The 'Running from sources' in http://pantsbuild.github.io/howto_develop.html is very much not about building a pex, so the details in the issue will help - b
billowy-minister-38128
10/29/2016, 7:03 PMSorry, that is build with pex. I'll have repro steps for you - l
limited-country-20626
10/29/2016, 11:59 PMI once had some really gnarly socket errors when setting up a new Pants repo from scratch on OSX. Feel free to tag @limited-country-20626 on the issue as well, and I can see if your problem looks familiar. - b
brainy-pilot-84216
10/31/2016, 6:48 AMSo I got the running from sources build working, but it doesn't seem to be using the java code I modified. Do I need to add something the to PYTHON_PATH or BACKEND_PACKAGES env variable that isn't in the doc? - b
brainy-pilot-84216
10/31/2016, 6:48 AMThe command line I'm using is: - b
brainy-pilot-84216
10/31/2016, 6:49 AMExecuted from my repo, and ../../pants is the pants repo - w
wide-energy-11069
10/31/2016, 6:52 AM@brainy-pilot-84216 it would be better to execute pants from the build root otherwise the behavior isn’t too well defined - b
brainy-pilot-84216
10/31/2016, 6:52 AMOh -- the instructions are the opposite - b
brainy-pilot-84216
10/31/2016, 6:53 AMWait -- you mean the repo I'm trying to build or the pants repo? - w
wide-energy-11069
10/31/2016, 6:54 AMoh you got two repos i see. in that case it would be good to use absolute path as PYTHONPATH - b
brainy-pilot-84216
10/31/2016, 6:54 AMI can do that. It seems to be using the correct paths but I think it's picking up the wrong backend packages. - b
brainy-pilot-84216
10/31/2016, 6:54 AMLet me try. One sec - w
wide-energy-11069
10/31/2016, 6:55 AMit's picking up the wrong backend packages.
do you mean there are conflicting names for the backend packages between yours and the ones in pants repo? - b
brainy-pilot-84216
10/31/2016, 6:56 AMNo. I've modified the builtin java junit package and am trying to run that in my repo - b
brainy-pilot-84216
10/31/2016, 6:56 AMI'm trying to get native scala tests working without the @RunWith - b
brainy-pilot-84216
10/31/2016, 6:56 AMI have the code done and now want to test it - w
wide-energy-11069
10/31/2016, 7:00 AMhmm I wasn’t too sure about the workflow on the tooling side such as
. Might need to see if other folks know about it. I am trying to do some modification on that too, but haven’t got to that yetsrc/java/org/pantsbuild/tools/runner - w
wide-energy-11069
10/31/2016, 7:05 AMbased on https://rbcommons.com/s/twitter/r/4324/diff/1#0, the runner needs to be published locally for
to consumejunit.py - b
brainy-pilot-84216
10/31/2016, 7:11 AMI'm trying to bump the version number and see if that does it - b
brainy-pilot-84216
10/31/2016, 7:11 AMrelease.sh should publish locally with -n, no? - w
wide-energy-11069
10/31/2016, 7:13 AMnot sure. but i guess you can also temporarily hardcode the jar location intojunit.py - e
enough-analyst-54434
10/31/2016, 2:04 PMI have to run and won't be online til afternoon, but dead end alert - e
enough-analyst-54434
10/31/2016, 2:04 PMRunning from sources means running from Python sources - e
enough-analyst-54434
10/31/2016, 2:04 PMJava tool code must be compiled and packaged - e
enough-analyst-54434
10/31/2016, 2:05 PMThere are separate docced instructions for that you should be able to find - b
billowy-minister-38128
10/31/2016, 5:22 PManyone know where the docs are for publishing the java? it would be great to have end-to-end instructions in one place. - w
witty-crayon-22786
10/31/2016, 5:51 PM@billowy-minister-38128 : http://www.pantsbuild.org/3rdparty_jvm.html#using-a-snapshot-jvm-dependency is one example