Add win-arm64 migrator

[baszalmstra]

Description:

This adds a migrator to add win-arm64 builds. I mostly copied the OSXArm migrator. Im unfamiliar with the codebase so I welcome feedback!

Checklist:

• Pydantic model updated or no update needed

Cross-refs, links to issues, etc:

conda-forge/conda-forge.github.io#1940

[codecov]

Codecov Report

Attention: Patch coverage is 64.70588% with 30 lines in your changes missing coverage. Please review.

Project coverage is 77.00%. Comparing base (26460c8) to head (f9a2ade).
Report is 202 commits behind head on main.

Files with missing lines	Patch %	Lines
conda_forge_tick/migrators/arch.py	57.57%	28 Missing ⚠️
conda_forge_tick/make_migrators.py	33.33%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3194      +/-   ##
==========================================
- Coverage   77.07%   77.00%   -0.08%     
==========================================
  Files         129      129              
  Lines       13991    14074      +83     
==========================================
+ Hits        10784    10837      +53     
- Misses       3207     3237      +30     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[beckermr]

Thank you!

Let's add an additional test that runs the migrator on a feedstock and make sure the pinnings repo is setup properly. Then we should be able to merge!

[baszalmstra]

@beckermr Is there an example for one of the other arch migrators that I could copy?

[beckermr]

Not specifically for an arch migration but there are good examples in tests/test_migrators.py.

[beckermr]

/relock-conda

[isuruf]

No piggy back migrators?

[baszalmstra]

Im not sure what a piggy back migrator is?

[isuruf]

Have a look at the piggy back migrators list in OSX Arm migrator. They do additional stuff other than rerendering.

[beckermr]

If you merge main into your pr, there is now a default set you can use with a function to instantiate them. See the other samples.

[isuruf]

We need to support windows with UpdateCMakeArgsMigrator and GuardTestingMigrator

[isuruf]

No, it's best to add the piggybacks to make it less work for maintainers. Otherwise they would have to figure that on their own.

[wolfv]

Sure, but the first 20-50 packages are probably good anyways, right? I just want us to get started with bootstrapping win-arm64 :)

[isuruf]

No, it's not. I'd rather one person spend time with writing the migrator than 10 people all trying to figure out what went wrong.

[h-vetinari]

We can document what needs to be done in the tracking issue (or a separate one) that we link to in the opening comment of the bot. That way we avoid the "X people have to figure it out themselves" thing, even if some other modifications aren't yet automated.

I agree that we should get the migration started. IMO perfection is the enemy of the good here - we can improve as we scale up (and based on what we learn in the process).

[h-vetinari]

We need to support windows with UpdateCMakeArgsMigrator and GuardTestingMigrator

Although those should probably be considered baseline.

[baszalmstra]

I have been trying to come up with a test but Im really struggling to grasp how to construct all the different objects at play. If someone has some time, I would love some help with this!