Sorry, my Mac is so broken since yesterday. I have a fix, just waiting to get it committed and pushed.

Take your time, lol We're a ways a way from upgrading anywho

I’ve been forced to use MS Defender, and it’s service has broken down and making my I/O supere sluggish and something eats my disk at a rate of Gb/minutes…

Oh yeah, I remember all sorts of nightmares about antivirus at my last job

I still have 17% free disk left… so I think I hav time to push this before calling it the night. company support tells me they have a fix for me tomorrow…

Well dont try too hard 😛

hehe, oh yes!

just running

./pants version

can take well over a minute… 😉

I want to copy a file into all of my

python_distribution

targets. Can I use

relocated_files

(maybe one per

python_distribution

target) to do that? Or will

relocated_files

move the file no matter where it is used (so only one

relocated_files

target per file)?

@bitter-ability-32190 https://github.com/pantsbuild/pants/pull/15884 😉

Now you can have a registry repository setting like

"ibm-{target_repository}"

take a close look at the test case for image_ref formatting… 🙂

Yup! Makes sense to me! I'll let @hundreds-father-404 review the contents of the change when they are back, but seems like it'd work flawlessly ❤️

Is there a way to have pants add

--builtins

for all the registered targets when linting

pants-plugins/macros.py

?

pants-plugins/macros.py:20:26: F821 undefined name 'python_artifact'
pants-plugins/macros.py:34:5: F821 undefined name 'python_distribution'
pants-plugins/macros.py:45:26: F821 undefined name 'python_artifact'
pants-plugins/macros.py:71:5: F821 undefined name 'python_distribution'
pants-plugins/macros.py:81:5: F821 undefined name 'shell_sources'
pants-plugins/macros.py:86:5: F821 undefined name 'resources'

^^ that's annoying 😛

Been looking through the docs and mucking around for awhile now, but I'm stuck. How do I get the

.whl

file from a

python_distribution

target into the build context for a

docker_image

target? Or is this an anti-pattern? Works great for

pex_binary

, but there is no

output_path

argument for a

python_distribution

. And how do I also enforce a dependency on a

python_distribution

target in my

docker_image

target?

It should be doable, but idk if Pants can handle it today. Let me poke around at the code and find out tonight when I'm back at a desktop

FWIW I think this ties into a proposal im working on which should allow pants to be flexible enough to handle this.

I would have thought this would work today. I think all BuiltPackage instances are placed into the

docker_image

build context. No, @curved-television-6568? The problem might be knowing what exact path it's at.

FWIW I think inference works for anything with an output path field. I remember adding that

I worked around it with an

archive

target that assembled the wheels built from all the

python_distribution

targets I'm interesting in.

archive

targets are available in the build context, so I just copy that into the image, unpack it, and install all the wheels it contains

So something like:

archive(
    name="web-tarball",
    packages=[
        "src/python/org/services/silly_web_backend:silly_web_backend-dist",
        "src/python:org-libs"
    ],
    format="tar.gz"
)

docker_image(
    name="web_backend",
    dependencies=[
        ":msg",
        ":web-tarball",
        ":image_scripts"
    ]
)

Which means you have access to

<http://path.to|path.to>.docker.target/web-tarball.tar.gz

in the build context

My codebase is extremely large—it’s the codebase for my entire company, and we have a monorepo

Hi there, I'm trying to create a pex for raspberry pi, and doesn't seems to succeed.. got confused with all the platform wheel flags, but trying to pex f.e numpy, and it always put inside numpy-1.21.6-cp37-cp37m-manylinux_2_12_x86_64.manylinux2010_x86_64.whl, how can I make sure when I create a pex all the dependencies will be able to run there?

Hi everyone, we are currently adopting pants and really enjoy using it even though there are plenty of questions left. I am looking forward to discuss interesting topics here. The first question I have is if there is some kind of "watch" mode for running tests. When unsing

./pants run

the command is watching the filesystem and restarts. Is there something similiar for

./pants test

? I know this from https://nx.dev/packages/jest/executors/jest#watch where you can enter watch mode with

yarn test myapp --watch

and then select a test based on a test name regex pattern (see attached screenshot). Is something like this available? If no watch mode is available, can I filter for a regex name and then select from a list of matches?

Hey, I'm having several issues with regards to 3rdparty dependencies and how they behave with different lockfile generators and different goals. Maybe someone has an idea what I could try.

I'm looking at CI w/ pants in GHA. Please sanity check my thoughts: If there are multiple python versions in the matrix: https://github.com/pantsbuild/example-python/blob/main/.github/workflows/pants.yaml#L25-L26 Can the cache be shared between the python versions? I'm thinking that the cache key needs to include the matrix.python-version https://github.com/pantsbuild/example-python/blob/main/.github/workflows/pants.yaml#L36 And, shouldn't the log include the matrix var(s) as well? Otherwise only one of the log uploads will succeed. https://github.com/pantsbuild/example-python/blob/main/.github/workflows/pants.yaml#L60

So, I often have pants running in one terminal, and then I open a file in vim in another terminal. vim creates its temporary file which triggers pants file change detection and then it restarts whatever pants goal I had running in the other terminal. Is there a way to get pants to ignore vim's swap files? I tried adding

*.swp

to

[GLOBAL].pants_ignore

but that didn't seem to do anything. If I don't want to interrupt pants' process, then I make sure not to save until it's done. But, I can only do that if I opened the file before pants started.

I couldve swore there was an issue regarding having an option

.add

at multiple config values overriding each other, but can't find it 😕 E.g.

pants.toml - [GLOBAL].pants_ignore.add = [...]

And

.pants.rc - [GLOBAL].pants_ignore.add = [...]

the second case overwrites the first 😕

hey folks 🙂 I have a draft of a blog post that • shows some of the techniques and commands you may use to lint your monorepository configuration and build files. • shows how Pants can help you find issues in your application design and dependency graph thanks to its dependency inference and the concept of file ownership through targets. Feedback and comments are most welcome! https://docs.google.com/document/d/1T5BWB-NRZUode8zT4-oZg37WD1Ti8DB301ng-IebblY/edit?usp=sharing

Is there a way to use environment variables from within your

BUILD

files? For context, looking to configure the

repositories

field of a

python_distribution