Don't defer mid-circuit measurements when mcm_method is "device"

[lillian542]

Context:
Some mid-circuit measurement handling is device-specific, like tree-traversal. When a method is device-specific, we want to skip transforming the tape in the standard preprocess pipeline, and instead rely on handling that is implemented in the device class.

This is currently the case for tree-traversal, but the more generic specification "device" defaults to deferred measurements.

Description of the Change:

We handle mcm_method="device" in the mid_circuit_measurements transform the same way we handle mcm_method="tree-traversal" - with a null transform and null_postprocessing.

Benefits:
Generally, we can use "device" the way it seems to have been intended as a method for handling mid-circuit measurements.

More specifically, we can use "device" as an MCM method on null.qubit if we don't want to include transforming and post-processing MCMs in our mock execution, but just want to treat MCMs like any other operation on the tape with null.qubit and skip it. This will be helpful for benchmarking the device processing and post-processing for large-scale FTQC workloads on null.qubit.

[github-actions]

Hello. You may have forgotten to update the changelog!
Please edit doc/releases/changelog-dev.md with:

• A one-to-two sentence description of the change. You may include a small working example for new features.
• A link back to this PR.
• Your name (or GitHub username) in the contributors section.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 99.64%. Comparing base (b2f1f65) to head (c33331c).

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #7243   +/-   ##
=======================================
  Coverage   99.64%   99.64%           
=======================================
  Files         513      513           
  Lines       50172    50172           
=======================================
  Hits        49996    49996           
  Misses        176      176           

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

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[andrijapau]

LGTM! 🚀

I'm just confused about the doc string,

"""Provide the transform to handle mid-circuit measurements.
    If the tape or device uses finite-shot, use the native implementation (i.e. no transform),
    and use the ``qml.defer_measurements`` transform otherwise.
"""

Isn't it true that if the tape is using finite shots then we are using the dynamic_one_shot transform?