https://github.com/pantsbuild/example-django/blob/2b0df09201a53ede07df62a8169ac78ac3203019/helloworld/util/settings_for_tests.py#L23 Is there documentation for this environment variable? It's empty in my tests

Dear Pants community. My team and i use

mypy==0.91

for type checking. When my colleague runs

./pants typecheck <projects/project>

it works pretty well. When i run it, it will solve the constraints.txt file, and it takes 25+ mins. How come it runs so slowly on my computer?

is there any way to get access to the anon stats that is collected for your own projects? (when you have the repo id)

I have a problem upgrading pants from version 2.5 to 2.6

What is the better way of creating a target from a python package which has further sub-packages with BUILD files? I have done it with

sources

set to glob pattern like

dir_name/***/**.py

. With the glob pattern the issue is that the sub-modules can be owned by two targets - one at the top level and the other one from the nested sub packages with BUILD files.

Hey, is there a way to provide the Git python package as dependency to use it in rules? I want to generate a version for a python package, based on Git magic, but I dont get Git installed as dependency to use it in rules. 🤔

Is there some global settings for

tailor

where we can provide list of directories which it should ignore when running

./pants tailor

?

I have another pants+django mystery I could use some help with– I can't figure out why my test database is not being created on test runs. With our non-pants stack we get

Creating test database for alias 'default'...

and a working database. When running the same tests with pants, the django testing framework does not appear to create the test database. Am I missing something obvious here?

Hi, I have a question about how to package the config/static files with a python distribution. Right now I add it as a

resources

dependency explicitly in

python_distribution

target, but it still doesn't get shipped with the source code. What is the way to achieve this?

I'm on 1.27 trying to lint scala with

lint.jvm-dep-check.unnecessary_deps

and

missing_direct_deps

and seeing this:

File "/Users/mitesh/.cache/pants/setup/bootstrap-Darwin-x86_64/1.27.0_py36/lib/python3.6/site-packages/pants/backend/jvm/tasks/jvm_dependency_check.py", line 292, in _compute_unnecessary_deps
    for dep_entries in actual_deps.values():

Exception message: 'list' object has no attribute 'values'

I debugged here and printed out

target

and

actual_deps

and it gives this. The TODO comment seems to acknowledge the bug?

JarLibrary(//:scala-library-synthetic) ['/usr/local/Cellar/openjdk@8/1.8.0+312/libexec/openjdk.jdk/Contents/Home/jre/lib/rt.jar']

Hi all, I'm testing out

pants

and ran into with an interesting failure case with installing a 3rd party dependency (avro-python3==1.10.0). Short version: using

ca_certs_path + indexes.add

for a company pypi repo caused installation of the setup_requires dependencies to fail. After commenting out those options in

pants.toml

the installation would succeed. All other dependencies (including private ones) succeed fine with the ca_certs_path + indexes.add.

pip install 'avro-python3==1.10.0'

works fine either way. Details: thanks to running with

--pex-verbosity=9

(wish that was on the troubleshooting page!) found the following error:

OSError: Could not find a suitable TLS CA certificate bundle, invalid path: cert.pem

Given I provided a full path for

ca_certs_path

I'm wondering what could be stripping off the rest?

./pants --version

Gets stuck just after finishing creating virtualenv. I was trying to do a fresh install. I had cleaned the

~/.cache/pants

I have a minor nit regarding field values I’d like to address. And that is default values vs

None

. 🧵

I have seen that

pex

targets won't update itself by taking into account the changes in the provided 3rd party dependencies. Whenever I run

./pants run <pex_target>

, if it fails because of some missing 3rd party dependency, it keeps failing even after I add that 3rd party dependency to say

requirements.txt

. Is this expected?

Is there a way to make

update-build-files

run only on BUILD files which have changed? In pre-commit and CI I would like to avoid checking every file if we can.

Does the Go support include

go generate

yet? I can't seem to find it in the docs or any issues tracking it.

I am trying out the go support and have run into an error I don't understand. When running basically any pants command, I get this error:

ProcessExecutionFailure: Process 'Run `go list` to download and analyze all third-party Go packages' failed with exit code 1. ...  

...
go: updates to go.mod needed, disabled by -mod=readonly; to update it:
	go mod tidy

Yet

go list

on its own doesn't cause errors. And

go mod tidy

doesn't show any changes.

I need to run a script to trust some additional ca certs before tests/running binaries will work. Is there a recommended way to do that? Currently I’m building a few conda envs (so I have working Pythons of the correct version), putting these on my path and running the script against their envs roots, which seems to cause those certificates to get brought into the test-time environment. Does that make sense, first as a question and second as an approach?

Is it possible to use Conda to install things with Pants? Our current workflow installs a few packages with Conda first (because they’re generally better/more compatible), then the rest with Pip. Would this be a plugin that exposed a “conda_requirements” command?

Hi ! Sorry noob question again : I have a docker image built locally that I want to use with Minikube. When I tried to rebuild it after setting up env variable with

eval $(minikube docker-env)

and try to pants package the dockerfile again, I have this :

ProcessExecutionFailure: Process 'Building docker image recital/n20-dev:latest.' failed with exit code 1.
stdout:
stderr:
2021/12/13 16:15:16 http2: server: error reading preface from client 192.168.49.2:2376: bogus greeting "Client sent an HTTP requ"
read tcp 192.168.49.1:38404->192.168.49.2:2376: read: connection reset by peer

I assume that I have to change docker context in pants.toml (DOCKER_HOST value in [docker]) but I still get the same error.

Docker built . -t {imagetag:version}

doesn’t work either. Thanks for help 🙂

Curious use case for pants… 🧵 🙂

Is there a way to have target local pylint/mypy arguments? For example, I want my core projects to have extremely strict lint and type checking requirements, while allowing things to be a bit looser in leaf projects?

Hi all, quick question about the docker_image target. if there is a way to substitute a common prefix in "registries"? All our AWS ECRs start with the same string - ###.dkr.ecs.us-east-1.amazonaws.com/ - with a different suffix for each image. I'd like to do something like this: registries=["${ECR_PREFIX}/myimage"] Any ideas how that might be accomplished? Similarly, is it possible to add a hook to authenticate against ECR before attempting to push the image? Thanks!

EDIT: Solved in thread... I'm dumb I want to add Pants to our monorepo for lint/format/check with the lowest amount of boilerplate. I'm thinking people won't like every dir being blasted by

BUILD.pants

to sit next to our Bazel

BUILD

files, so my first attempt was to try a (code) root

BUILD.pants

with:

python_sources(name="source", sources=["**/*.py"])
python_tests(name="tests", sources=["**/test_*.py", "**/*_test.py"])

But I'm seeing a hearty dose of

The following imports in a/b/c/d/test_foo.py:../../../../tests have no owners

which points to an import of a module which is just chilling next to the test. --- No source roots at play

I have a package “data” that has a requirements.txt with torch >= 1.8, and a binary “train” that has a requirements.txt with torch == 1.10.1 and depends on data. In my BUILD I have to explicitly do:

pex_binary(
    name="train",
    entry_point="train.py",
    dependencies=[
        "src/data:data",
        "!!src/data:torch",
        "examples/ml_example:torch"
    ]
)

and I still get `The target examples/ml_example/train.py imports

torch.nn.functional

, but Pants cannot safely infer a dependency because more than one target owns this module, so it is ambiguous which to use: ['examples/ml_example:torch', 'src/data:torch']` . If I get rid of

"!!src/data:torch"

pants says:

More than one direct requirement is satisfied by torch 1.10.1:
0. torch
1. torch>=1.10

My goal is to allow varying level of strictness - data should support a wide variety of torches, while other people can do what they want. Is there a better way than doing

"!!src/data:torch"

?

I'm building a python script to format files not tracked by pants (rogue python scripts, markdown, html, json, yaml, etc.) and was wondering if the

requirements.txt

file will install a dependency in CI to give my script access to it from the shell. Now, I could just load

black

as a module but I'd rather have all commands for

black

,

prettier

,

scalafmt

, and any other formatters to be handled the same exact way (subprocess.run). Will

black

be available from the shell w/ the requirements.txt file specifying

black

as a dep?

Is it possible to disable the “cancel if the file system changes”? I know there’s a --loop option, but my usual workflow doing development is run a script, make some (maybe unrelated) changes while I wait for it to crash, read the logs, repeat, so looping would get in the way of that a bit.

Hi, I have question regarding the

platform

option when building

pex_binary

as mentioned in the documentation. Does this mean if I build a

.pex

on macos and copy it to a docker whose base image which isn't mac, then I need to also build explicitly

.pex

compatible with the base docker image? https://www.pantsbuild.org/docs/reference-pex_binary#codeplatformscode

I am trying to create a docker image and it isn't able to identify paths correctly. Please see the attached comments for more I stripped down the docker file to just add the requirements.txt from

3rdparty

. Just this

FROM public.ecr.aws/lambda/python:3.9

RUN yum -y install gcc

COPY ./3rdparty/requirements.txt .
RUN python3 -m ensurepip \
    && pip install --no-cache-dir -r requirements.txt

I keep getting this error

> [3/4] COPY ./3rdparty/requirements.txt .:
------
failed to compute cache key: "/3rdparty/requirements.txt" not found: not found