Hi guys, what’s the best way how to create Django migrations with Pants? Since everything is run in the sandbox I need to copy migrations manually which is quite tedious with 20+ django apps. Also I have issue with detecting changes unless the django app is specified explicitly. I’m composing django apps from shared apps (e.g.

comapany.payments

,

comapany.contrib

) and service specific (e.g.

my_service.auth

). Any help is welcome 🙂

did you check the example repo? there are migrations added as dependencies for tests - I would assume something similar is possible for running migrations?

Yop, it’s only mentioning how to run migrations, but nothing about creating them. It would be nice for this particular use case if the command won’t run in sandbox or if it would be possible to define some shell target which would move the files from sandbox to the project

export a venv like you'd do for IDE integration maybe? 🤔

Not sure how it would work. I guess you can run the django makemigrations outside of pants with the generated env, but it doesn’t feel right 🙂

yeah that's the idea

in our repo, we set

run_goal_use_sandbox=False

on the targets for our

manage.py

files. then we

pants run path/to/manage.py -- makemigrations

There is: https://www.pantsbuild.org/docs/reference-python#default_run_goal_use_sandbox I think the confusing thing about the example-django repo is its just old. It uses a

pex_binary

target for

manage.py

and the

run_goal_use_sandbox=False

does not exist for

pex_binary

targets; only

python_source(s)

in this run-a-thing context.

Awesome, this is what I was looking for. Thanks guys

You cannot use it for

pex_binary

- that was my point. If you cannot convert your

pex_binary

target to a

python_source

target and use that instead (or define both ~side-by-side), then you may have to file a feature request.