fix: extend catch_unwind to cover lock_pypi_packages for proper panic handling

[nichmor]

fixes: #5315

Summary

• Moves the lock_pypi_packages call inside the catch_unwind block to properly catch panics that occur during wheel metadata fetching
• Moves last_initialization_error() method from LazyBuildDispatch to LazyBuildDispatchDependencies since LazyBuildDispatch is moved into the async block and can't be accessed after the catch_unwind

This fixes panics like:

thread 'main2' panicked at crates/pixi_core/src/lock_file/resolve/build_dispatch.rs:447:17
Error: × failed to solve the pypi requirements of environment 'myagv' for platform 'linux-aarch64'
  ╰─▶ build dispatch initialization failed: there is no compatible Python interpreter for 'osx-arm64'

Testing

Tested by cloning the repository from the issue, and instead of using vendored conda package

rpi-gpio = {path = "./vendor/rpi-gpio"}

I'm using the one from pypi

[feature.myagv.pypi-dependencies]
RPi-GPIO = "==0.7.1"

When trying to lock, it produces the right error:

AI Disclosure

• This PR contains AI-generated content.
  • I have tested any AI-generated content in my PR.
  • I take responsibility for any AI-generated content in my PR.

Tools: Claude

Checklist:

• I have performed a self-review of my own code
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• I have added sufficient tests to cover my changes.
• I have verified that changes that would impact the JSON schema have been made in schema/model.py.

[wolfv]

Looks good to me and I think the error message is much better! I'll let @tdejager have another look at the code.

[tdejager]

Nice! Maybe it would be good to move the last error out of the dependencies structure altogether, as it does not really count as a dependency.

[tdejager]

Not sure if it would be doable to pull this out, or we need to rename the struct ;)

[tdejager]

Also make sure to fill out the AI disclosure fully :)

[tdejager]

Actually, now that I look at the description, I'm unsure what actually changed in the output. Is the description correct, from the original issue it seems that the miette error was never printed?

[nichmor]

Actually, now that I look at the description, I'm unsure what actually changed in the output. Is the description correct, from the original issue it seems that the miette error was never printed?

the error itself from locking was never printed and only panicked