put ucx dependency in run_constrained

conda-forge / openmpi-feedstock

A conda-smithy repository for openmpi.

BSD 3-Clause "New" or "Revised" License

9 stars 25 forks source link

put ucx dependency in run_constrained #103

Closed minrk closed 2 years ago

minrk commented 2 years ago

ignoring run_exports

hopefully closes #102

We could move this to an openmpi-ucx metapackage as suggested in #102, but I'm not sure conda install openmpi-ucx is better than conda install openmpi ucx which should have the same result. Unless we actually move the ucx plugin files to the second output, but that would mean we have to find them.

Another alternative is to make the cuda build an actual separate build, rather than trying to keep all the optional features present without dependencies in the default linux build.

conda-forge-linter commented 2 years ago

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 (recipe) and found it was in an excellent condition.

jakirkham commented 2 years ago

Perhaps we should just be solving issue ( https://github.com/conda-forge/ucx-split-feedstock/issues/26 ) and prioritizing CPU there given this feedback?

minrk commented 2 years ago

That makes sense!

minrk commented 2 years ago

I don't really understand the conda-build bug that's preventing pin_compatible from finding ucx in the host dependencies. But it certainly is a conda-build bug.

conda-forge-linter commented 2 years ago

Hi! This is the friendly automated conda-forge-linting service.

I wanted to let you know that I linted all conda-recipes in your PR (recipe) and found some lint.

Here's what I've got...

For recipe:

The outputs section contained an unexpected subsection name. string is not a valid subsection name.
The outputs section contained an unexpected subsection name. track_features is not a valid subsection name.

conda-forge-linter commented 2 years ago

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 (recipe) and found it was in an excellent condition.

minrk commented 2 years ago

Figured out the conda-build bug. It extracts the wrong requirements text for individual outputs when you have template-level if-statements that change the order or number of outputs. Moving our 'extra' output to the end seems to avoid it.

dalcinl commented 2 years ago

@minrk What's your plan for this one? Would you prefer to wait until the conda-build bug is addressed?

minrk commented 2 years ago

@dalcinl No, there's no need to wait for conda-build. The workaround here works just fine.

leofang commented 2 years ago

The Travis ppc64le CI has failed on master. I've restarted, but given its recent poor service quality if it fails again we should switch back to Azure.

leofang commented 2 years ago

The Travis ppc64le CI has failed on master. I've restarted, but given its recent poor service quality if it fails again we should switch back to Azure.

See #108.