New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Use correct CLI entrypoint #146
Conversation
Hi! This is the friendly automated conda-forge-linting service. I just wanted to let you know that I linted all conda-recipes in your PR ( |
@conda-forge-admin, please rerender |
Hi! This is the friendly automated conda-forge-webservice. I tried to rerender for you, but it looks like there was nothing to do. This message was generated by GitHub actions workflow run https://github.com/conda-forge/dask-core-feedstock/actions/runs/3322902694. |
Perhaps we should use The |
- dask docs --help | ||
- dask info --help | ||
- dask info versions --help |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Tests are failing because the CLI requires click
; are we okay with adding click
as a core dependency for the Dask conda packages? Or are these tests better suited for the dask
metapackage
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yes think that is reasonable
We should then also added it to Dask's setup.py
and the nightly package
Would reuse Distributed click >=6.6
requirement
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I initially suggested adding click
as a required dependency of dask
(pip) / dask-core
(conda) but there was some push back against that. I personally don't have strong thoughts one way or the other. I agree we should have these. We currently have these types of checks over in the dask-feedstock
due to the current places that click
is installed
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yeah think we still want some kind of check in this feedstock to make sure the entry point added here works
We could just add click
to test/requires
as is done with pip
below
That said, I think not including click
as a requirement in Dask itself while adding entry points is a subpar user experience. Think we should bring this up again and point back to this issue (and its resolution) as a reason to change it
Was going to ask about this, as it seems like this error handling is implemented in cli.py as well, so wasn't sure if
It's not, we would need to make changes to the core dependencies if we wanted Dask's CLI to work out of the box after installing |
cc @jrbourbeau |
|
Thanks for the context @douglasdavis 😄 in that case, using Some follow up questions:
|
In the CLI PR we decided not to make Click a hard dependency. It seems that it may be easiest to just make it a dependency.
It looks like I shouldn't have opened conda-forge/dask-feedstock#200 - based on what @jrbourbeau and @jakirkham were doing I think the the entry point should be part of |
The only thing I can think of is maybe the command still returned exit code 0 |
It was doing that at the time. Maybe because of these? Don't see that behavior with recent packages (that dropped those) $ dask scheduler ; echo $?
Usage: dask [OPTIONS] COMMAND [ARGS]...
Try 'dask -h' for help.
Error: No such command 'scheduler'.
2 |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks all for digging in here!
Folks seem to think dask = dask.cli:run_cli
is the right thing to do here. I'm inclined to remove the dask docs --help
tests for the time being (we have some coverage still as those are included as test in dask-feedstock
) to see if the entypoint update fixes things. Adding click
as a dask-core
dependency seems like a relevant, but separate, conversation. Thoughts?
As noted above, think if we are adding an entry point here (especially given the issues we have had around it), we should have a test here as well. In terms of |
@jrbourbeau I agree with your inclination (kick the can down the road on the Click dep w.r.t this PR, but at this point I would say for the upcoming release I think it's a good idea to just have a Click dep). I'll also add that I think |
This is a passing run triggered by the removal of those entrypoints, so think it might be something else. I'm generally in favor of @jakirkham's suggestion to add |
@conda-forge-admin, please re-render |
(oops missed re-rendering was already done above 🤦♂️) |
Hi! This is the friendly automated conda-forge-webservice. I tried to rerender for you, but it looks like there was nothing to do. This message was generated by GitHub actions workflow run https://github.com/conda-forge/dask-core-feedstock/actions/runs/3323505831. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM. Thanks all! 🙏
Just tried out the new build and |
Nice! Also working for me 🎉 Think we will want to add these changes to the Am looking into making |
Closes #145
Closes #144
Fixes dask/distributed#7176
Checklist
0
(if the version changed)conda-smithy
(Use the phrase@conda-forge-admin, please rerender
in a comment in this PR for automated rerendering)