and I have in my pants.ini:

[jvm.run.jvm]
options: ['-Xmx2g']

searching for

-Xmx256

across my .pants.d folder, I seem to have lots of java invocations passing that in, but no settings on my side that request it

The help is a bit of as wall of text and confusing, but the summary looks like so:

$ ./pants test.junit --help
test.junit options:
...
cache.test.junit options:
...
jvm.test.junit options: 
...
--jvm-test-junit-options=<option>... (--jvm-test-junit-options=<option>...) ..., --jvm-test-junit-options="[<option>..., <option>..., ...]", --jvm-test-junit-options="+[<option>..., <option>..., ...]" (default: ['-Djava.awt.headless=true','-Xmx1g','-XX:MaxPermSize=256m'])
    Run with these extra JVM options.
...

So @bright-apartment-76721 the 1st heading,

test.junit

are direct junit runner options, the second

cache.test.junit

are general caching options as applied to

test.junit

and the last are general jvm options as applied to

test.junit

. Its that last that you want, so:

[jvm.test.junit]
options: [...]

I see 🙂 thanks

The

[jvm.run.jvm]

you have is the same, but as applied to

./pants run

but what's defaulting to

-Xmx 256m

?

(for the jvm)

is that something inside pants?

So to view provenance, use

./pants options --scope=... --name=

For pantsbuild:

$ ./pants options --scope=jvm.test.junit --name=options
jvm.test.junit.options = ['-Djava.awt.headless=true', '-Xmx1g', '-XX:MaxPermSize=256m'] (from CONFIG)

And with this fix (https://github.com/pantsbuild/pants/pull/4832) soon:

$ ./pants options --scope=jvm.test.junit --name=options
jvm.test.junit.options = ['-Djava.awt.headless=true', '-Xmx1g', '-XX:MaxPermSize=256m'] (from CONFIG in pants.ini)

I expect you'll find (from HARDCODED) which would be the default from within Pants

Lazy survey, does anyone know of any reason not to adopt the python3 subprocess32 subprocess backport (https://github.com/google/python-subprocess32) in Pants? This gives us reliable / maintained subprocess timeouts vs our own hand-rolled.

To target the lazy survey, I'd value @rough-minister-58256's feedback if any for one.

+100

also ftr, I am guiding a newly established language ownership team at work who will be focused on python language improvements targeted at our ML teams

python 3 migration is going to be one of the big focuses

it may make sense to nudge pants in that direction as well, so that we don’t have to keep using backports etc

but for the time being, subprocess32 usage would be great

it solves several issues afaik

Excellent, thanks @rough-minister-58256 , PR coming soon.

Hi all, I am trying to add a javascript lint/fmt goal to pants -- I was wondering if someone has any suggestions/comments to make on https://github.com/pantsbuild/pants/pull/4785?

There was some initial reservations about how to set up javascriptstyle that I tried to address. I know this is not a perfect solution, but am hoping it can be a first attempt?

@rough-minister-58256 - any help on the PR for manylinux1, it will help releasing the bazel rules - https://github.com/pantsbuild/pex/pull/316

if I have a bunch of dependencies in a 3rdparty/jvm/BUILD file

is there a good way to check up on how far behind I am the upstream versions?

like if I have version X in there, and version Y >> X is the latest online, I'd like to know that without having to eyeball the file

basically, I want to be able to programmatically extract/manipulate the contents of that file

I guess I don't really need to be able to programmatically change values in it, but some way of reasoning programmatically about versions I use is useful (and then I can go edit the file by hand)