How can I get the pycharm debugger hooked up to my pytest run (running via

./pants test --debug

)? I followed the tip here: https://www.pantsbuild.org/v2.5/docs/python-test-goal#running-tests-interactively And I see that it connects to pycharm, and stdout is getting forwarded, but I'm unable to step through the code.

Hi all 👋! I'm new to Pants but excited to see what it can do for migrating a multi-repo set up I have into a monorep. I'm hitting a

No module named main

error -- which I can see is due to having two projects with a

src/main.py

. Indicative project structure here:

├── BUILD
├── apple-pie-api
│   ├── BUILD
│   ├── Pipfile
│   ├── Pipfile.lock
│   ├── requirements.txt
│   └── src
│       ├── main.py
│           ...
├── common-requirements.txt
├── pants
├── pants.toml
└── randmoji-api
│   ├── BUILD
│   ├── Pipfile
│   ├── Pipfile.lock
│   ├── requirements.txt
│   └── src
│       ├── main.py
│           ...
├   ...

I hit the issue when I try to execute the output pex for one of the projects; e.g., via

./pants package apple-pie-api:pex_binary && ./dist/apple-pie-api/pex_binary.pex

. Renaming one of the

main.py

(e.g., to

main2.py

) does the trick. But I'm wondering if there's a way to avoid needing to do that? Looking at the source roots https://www.pantsbuild.org/docs/source-roots this isn't in those example set ups, so perhaps I'm off piste here? But this is a set up I'd like to use as part of a transition to pants.

question about how to inform pants of some

dependencies

In the old Makefile, in one of the targets that happened before running tests, it looped over

contrib/runners/*

and ran

python setup.py develop --no-deps

(deps were already installed in another step). How do I do something similar? ie: Whenever running any of the tests under https://github.com/st2sandbox/st2/tree/pants/st2common/tests, I need to make sure that each of the packages under https://github.com/st2sandbox/st2/tree/pants/contrib/runners is installed.

Has anyone else used stevedore extensions in a pants repo? https://pypi.org/project/stevedore/ (including both consuming the extnsions and providing them)

how would i go about debugging why a particular module isn't being included in a library (e,g, not showing up in

filedeps

)

FYI: we revamped how we do changelogs. Instead of having summaries of each release in the docsite, we're going to use blog posts to summarize the release. https://www.pantsbuild.org/docs/changelog links to all the changelogs + highlights of the release (e.g. blog posts)

Hi there, it seems pants has stability issues. For example, when I build a PEX file I sometimes see errors such as " "/Users/mingshiwang/.cache/pants/named_caches/pex_root/unzipped_pexes/478cc1fa371ca40aa3e7dafee735ca438d4a243f/.deps/pex-2.1.42-py2.py3-none-any.whl/pex/pex_info.py", line 10, in <module> from pex import pex_warnings, variables ImportError: cannot import name variables". But when I rerun the command the error goes away. Is this a known issue?

I'm wondering what solution(s) folks have arrived at for needing to run a

pex_binary

against the source code and modify the filesystem in-place. Specifically, I need to be able to run

django

management commands like

./pants run manage.py -- makemigrations

that writes out migrations files. I'm thinking the answer is that if I need to run side-effecting code like that that i need to do it via a pants plugin like other codegen tools?

What's the consensus on monorepoing microrepos into one repo ✖️ Drop the commit history ✔️ Keep the commit history ?

Hi all, I have a question on dependency management. In my repo there is a common requirements.txt at the root folder, and separate requirements.txt files under subfolders of each project. The common requirements.txt contains dependencies that are discoverable to all projects, whereas requirements.txt at the subfolders contain dependencies that may have conflicting versions. I have read (https://www.pantsbuild.org/docs/python-third-party-dependencies) on how to manage multiple requirements files, but it seems the only way to have conflicting versions of dependencies is to write "python_requirement_library()" for each of them. This makes the build script complex. Is there any better way to do this? What is the best practice in this case? Thanks!!

https://twitter.com/pantsbuild/status/1395507448048324608 Introducing Pants 2.5: Shell support, config autodiscovery, and incremental tool adoption. https://blog.pantsbuild.org/introducing-pants-2-5/ Twitter

https://twitter.com/mateornaut/status/1395518229011996678 Pants Build retweeted: There is a shell script at your organization that: a) dates back to the seed round b) understood by <=1 engineer… https://twitter.com/i/web/status/1395518229011996678 Twitter Introducing Pants 2.5: Shell support, config autodiscovery, and incremental tool adoption. https://blog.pantsbuild.org/introducing-pants-2-5/ Twitter

Hi, I’m seeing a weird behavior when running the

package

goal in Pants 2.4.0 🧵

https://twitter.com/mateornaut/status/1395513619148902400 Pants Build retweeted: @EArellanoAZ @pantsbuild the kid and I used Pants 2.4 for a smal python game recently. My first crack at `tailor`… https://twitter.com/i/web/status/1395513619148902400 Twitter

On pants 1.27, I noticed that

jar_library

has a

managed_dependencies

field, but

scala_library

does not? Would it still work if I passed it in via the payload?

class AiqScalaLibrary(ScalaLibrary):
  def __init__(self, java_sources=None, payload=None, managed_dependencies=None, **kwargs):
    if managed_dependencies is not None:
        payload = payload or Payload()
        payload.add_fields({
            'managed_dependencies': PrimitiveField(managed_dependencies),
        })
    super(AiqScalaLibrary, self).__init__(java_sources=java_sources,payload=payload,**kwargs)

I'm trying to use

managed_dependencies

to build a top-level jvm binary, which deps on a scala library, which (deep down) has another scala dep where I want to enforce v1 or v2 of it

https://twitter.com/ErbStephan/status/1396552662456414214 Pants Build retweeted: WOW, the new @pantsbuild is awesome. After 3 years of absence I have just upgraded my example repository and totall… https://twitter.com/i/web/status/1396552662456414214 Twitter Building and packaging Python with @pantsbuild and PEX - an annotated example https://github.com/StephanErb/pexample #python #pantsbuild #pex #monorepo Twitter

https://twitter.com/ErbStephan/status/1396552666591989761 Pants Build retweeted: If you modify a function or bump a requirement, pants will just re-run tests that are directly or indirectly affect… https://twitter.com/i/web/status/1396552666591989761 Twitter

https://twitter.com/ErbStephan/status/1396552668382892033 Pants Build retweeted: Thanks to auto-inference of dependencies only very little boilerplate is needed go get going. And that boilerplate… https://twitter.com/i/web/status/1396552668382892033 Twitter

https://twitter.com/ErbStephan/status/1396552664830332933 Pants Build retweeted: Pants feels a bit like http://pre-commit.com on steroids. With pre-commits you can easily run commands like isort… https://twitter.com/i/web/status/1396552664830332933 Twitter

Is anyone familiar with this error:

ValueError: Failed to begin watching the filesystem: Bad file descriptor (os error 9)

Does Pants support M1 natively or is Rosetta needed? Didn't find anything in the docs besides that arm isn't supported on Linux, but when I try to bootstrap Pants I get the following with both Python 3.9.5 and 3.8.10 - presumably because only x86_64 wheels are available on PyPI.

Hello, I’m trying to run pants helloworld example from https://github.com/pantsbuild/example-python. $ ./pants test helloworld/util/lang_test.py --print-stacktrace Getting: ProcessExecutionFailure: Process ‘Building pytest.pex with 2 requirements: pytest-cov>=2.10.1,<2.12, pytest>=6.0.1,<6.3’ failed with exit code 1

For the first time, I added an archive target and immediately discovered our build server was missing

zip

. Is there a way to get a list of all the system dependencies pants and its plugins have? I’d like to use that as an input foe configuring our build servers.

I am confused by https://www.pantsbuild.org/docs/reference-python-setup#section-resolve-all-constraints .... true/false is clear enough (I think....) but the other options are not clear, is

always

the same as

true

and is

never

the same as

false

what about

nondeployables

, what does this option value mean?

Got a bit stuck again with trying to get django's

manage.py

running under the

run

without chroot solution

Build Meetup conference is next month, and currently seeking talk proposals. It's a fairly small event, probably 60 or so attendees. So if you'd like a low-pressure opportunity to talk about how you manage your build and the interesting challenges you've encountered, could be a great opportunity! I'm on the program committee, so am happy to answer questions if I can. Details in thread.

Trying to generate virtual env for IDE integration (using example-python). I’ve set a version of pytest in pants.toml:

[pytest]
version = "pytest==6.2.4"

And I can see correct version (6.2.4) is used during the test:

./pants test ::
11:08:01.21 [WARN] The constraints file constraints.txt does not contain entries for the following requirements: pytest
11:08:01.21 [WARN] Ignoring `[python_setup].resolve_all_constraints` option because constraints file does not cover all requirements.
11:08:01.22 [INFO] Completed: test - helloworld/greet/greeting_test.py:tests succeeded.
============================= test session starts ==============================
platform darwin -- Python 3.7.7, pytest-6.2.4, py-1.10.0, pluggy-0.13.1
rootdir: /private/var/folders/35/0694cnm970qfhkpw87g3xwdw0000gp/T/process-executionSa6d5C
plugins: cov-2.11.1
collected 1 item

But when I try to inspect all dependencies it shows different version (6.2.3, the one I used before):

./pants --no-pantsd --tag=-lockfile_ignore dependencies --type=3rdparty ::
ansicolors>=1.0.2
protobuf>=3.11.3
pytest==6.2.3
setuptools<54.0,>=50.3.0
translate>=3.2.1

In case anyone was following the workaround for running django

manage.py

, I have figured out a workaround for getting

runserver

reloading to work under a

pex_binary

here: https://pantsbuild.slack.com/archives/C046T6T9U/p1622139116141200?thread_ts=1622124247.126600&amp;cid=C046T6T9U

FYI a couple changes made to our CI docs. Feedback and questions welcome! 🧵 https://www.pantsbuild.org/docs/using-pants-in-ci

Am I correct in understanding/thinking that I can't generate `setup_py`'s

name

argument from a plugin since

setup_py

appears to be a direct alias for

PythonArtifact.__init__

, which has assertions that argument is being provided?