one of the most confusing (as a python beginner too) aspect of trying to grok pants is how the mixin/dependency injection works, and knowing what i can “request” in the conrext of writing a goal/rule — are there any (heh…), uhm — rules around that behaviour? how do i set myself up for success as a first time pants plugin developer?

I'm getting the following warning on every invocation with 2.14.0rc5. Any tips?

13:39:41.24 [WARN] /home/nathanael/.cache/pants/setup/bootstrap-Linux-x86_64/2.14.0rc5_py39/lib/python3.9/site-packages/pkg_resources/_vendor/packaging/specifiers.py:255: DeprecationWarning: Creating a LegacyVersion has been deprecated and will be removed in the next major release
  warnings.warn(

I needed to do something like https://stackoverflow.com/questions/50174130/how-do-i-pytest-a-project-using-pep-420-namespace-packages/50175552#50175552 in my

conftest.py

. Pants doesn't know that

_pytest

comes from the

pytest

package. Where's the most appropriate place to inform that? I could get by with a

# pants: no-infer-dep

, but curious about a more proper fix.

I'm thinking this may be a simple error, but nothing I've tried has worked. I'm trying to get my CI workflow on Github working, but I'm having trouble with passing in my build arguments. My

docker_image

target is very simple:

docker_image(
    name="app",
    image_tags=["{build_args.GIT_COMMIT}", "latest"],
    registries=["{build_args.REGISTRY}"],
)

then I have Github run it using

./pants --docker-build-args="[GIT_COMMIT=${GITHUB_SHA::7}, REGISTRY=${{ steps.login-ecr.outputs.registry }}]" package src/services/app/Dockerfile

. The error is in the 🧵, my guess is I have the syntax wrong in my package statement but I can't figure it out.

Recently, I've been plagued with Pants crashing more often. When Pants crashes, the below error message appears:

native_engine.PantsdClientException: The pantsd process was killed during the run.

If this was not intentionally done by you, Pants may have been killed by the operating system due to memory overconsumption (i.e. OOM-killed). You can set the global option `--pantsd-max-memory-usage` to reduce Pantsd's memory consumption by retaining less in its in-memory cache (run `./pants help-advanced global`). You can also disable pantsd with the global option `--no-pantsd` to avoid persisting memory across Pants runs, although you will miss out on additional caching.

If neither of those help, please consider filing a GitHub issue or reaching out on Slack so that we can investigate the possible memory overconsumption (<https://www.pantsbuild.org/docs/getting-help>).

I tried following the two pieces of advice in the error message, but to no avail. What did help alleviate the crashing was closing all other running applications (I also cleared the Pants cache for good measure). I understand that Pants is being OOM-killed in this type of situation. However, the advice in the error message did not help me at all. Only after I closed all other applications did the crashing stop. Can this error message please be revised to include advice that is easier to understand and follow, i.e. closing all other running applications to reduce memory consumption?

Can I generate

[flake8].extra_requirements

from a

requirements.txt

file ? Instead of hardcoding it in the pants.toml ?

Not with

generate-lockfiles

, but you can if you just use

pex

to lock directly, and set

[python].invalid_lockfile_behavior = "ignore"

If that's something you're interested in LMK.

But also, can I ask why? 🙂

I just have a lot of requirements files I have right now (

requirements-flake8.txt

,

requirements-pytest.txt

, Etc) And was wondering if I can reuse them instead of rewriting them in a config file which is pants specific

So, this is totally possible (I think, haven't tried). Mind making a GitHub Discussion and I'll answer over there (for posterity)

Replied, let me know if that works so I can mark it as the answer 😛

And if you wanna know how to sandbox

pex

for anyone to use let me know as well

Q re: Typescript/Javascript support (recalled wanting to ask about this on the monthly meeting last week). Will probably raise this as a GH discussion too. What would Pants's "purpose" be in TS/JS support? Would the idea be to be: A) a replacement for stuff like

webpack

,

vite

,

esbuild

(yay Go),

swc

(yay Rust), etc (to which I'm a giant 👎 ) - or, B) would it be to use those as plugins and facilitate their usage in a consistent matter, and maybe throw some dep inference + caching in for good measure, or C) Alternative/middleground

prettier

is currently a plugin,

eslint

should be added in (and can be added in, if not already there) - so we've started on the

fmt

and

lint

(and

fix

stories) and then we're stuck with the question of transpiling/bundling (

check

,

package

). For instance, my internal tooling moved from

experimental_shell

to a ViteJS plugin which uses all of my existing configuration (and, more importantly, plugins for Svelte/Kit) and basically calls the correct vite args at the command line. The prospect of using Pants and migrating config from

vite

to yet another transpiler/builder (since the JS ecosystem is the wild west), is a non-starter for me.

Does anyone know if there is a way to turn on unmatched_build_file_globs=warn per-target? I want this because I am trying to normalize our use of python_test_utils (to use it in every test directory), but many directories do not (yet) have a file matched by one of the file globs.

I have some packages that use

importlib.metadata.version

to get the installed version of my packages at runtime from the installation metadata. Unfortunately, pants doesn’t install packages into its hermetic environment when the package is under test, so importing the package fails because

importlib.metadata

cannot find the package. Is there any way around this other than

try/except

in the

__init__.py

for the package?

I'm trying to add a 3rd party dependency that depends on

tables

on mac m1. With vanilla

pip

, I install with extra brew packages and environment vars:

brew install hdf5 c-blosc lzo bzip2

export GRPC_PYTHON_BUILD_SYSTEM_OPENSSL=1                                                     
export GRPC_PYTHON_BUILD_SYSTEM_ZLIB=1
export HDF5_DIR=/opt/homebrew/opt/hdf5                                                        
export BLOSC_DIR=/opt/homebrew/opt/c-blosc

Is there anyway I can propagate these environment vars to

pants

when its building requirement

pex

?

Hi all, bit of a dumb question, but how do I have pants pick up tests within a folder? My current directory is like this and works fine:

tests/
   - test_a.py
   - test_b.py
   - BUILD

I want to have a structure like this:

tests/
   - test_a.py
   - test_b.py
   - BUILD
   - extra/
      - test_c.py
      - test_d.py

What do I need to do to have my BUILD pick up the tests in that directory as well?

I started getting the "too many open files" error after upgrading to 2.14.0 (but had 2.14.0rc5 before). My "ulimit -n" is 1048576, so it should be more than enough. I removed the whole ~/.cache/pants and .pants.d and all that - no help. It's also kind of bad that I can't even run "./pants --version" which should not try to start pantsd. What can I do? I will try to restart, but it shouldn't be needed (I'm thinking about what to do if this happens in CI)

Is there a way to restrict which in-repo modules are allowed to be imported to different in-repo modules with pants?

I want to add some custom checking on a project's dependencies (look for illegal and unused ones, for example). I can do it with a script that calls pants repeatedly and processes the results, but it is slow to do that many times. Is a linter plugin the right thing to do instead?

pex_binary()

takes a

platforms

argument,

python_awslambda()

does not. Is there a reason for this? 🤔

is there an easy way to speed up developing with django + pants? atm the even the most minimal change takes ~8-10 seconds to reload the django development server without pants this is nearly instant

is there an “idiomatic” way to represent targets that are runnable via pants and have stdout / output pipeable to other programs while also having said output on stdout available as an input to another target (via file, or other mechanism) ? that’s a poor explanation, but hopefully you get it 😂

okay, so i’m just mucking around, and trying out a few things — is it a correct assumption that certain fields are not parameterize’able ? for instance — i don’t seem to be successful with having the

command

field be wrapped by

parametrize

while

extra_env_vars

seem to work out … is there some expectancy around parameterization of input fields?

Hi community, I’m wondering has anyone tried to use pants with airflow? Airflow as a shared environment (running multiple data pipelines) would benefit from better dependency management. But there might be architectural changes needed in Airflow in order to execute tasks using Pex binaries.

Is there a way to run

./pants export

against only a specific python tool?