Raise an error for MCMs nested in Conditionals

[lillian542]

Context:
The current methods for of mid-circuit measurement processing don't support applying a mid-circuit measurement conditionally. This was not within the scope of the implementation for any of these methods, so we don't expect it to work - however, we would prefer to return clear error messages, and we don't want to be returning incorrect results.

Currently, qml.cond(m, qml.measure)(wire) either raises a difficult to parse error (in the case of deferred measurements and analytical execution), or returns incorrect results (in the case of tree-traversal, one-shot, and deferred measurements with shots).

Description of the Change:

We add validation in the midcircuit_measurements transform, checking for conditional mcms and raising an error. Specific mcm_methods can be made to pass this validation check in the future if support for conditional mcms are added to the method; for now, we disallow it completely.

Benefits:
We don't return incorrect results or raise cryptic errors.

Possible Drawbacks:
It adds one more iteration over the operations for validation 🤷‍♀️

Related GitHub Issues:
None, but sc stories [sc-84177] and [sc-84178]

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 99.60%. Comparing base (ab34d13) to head (0389335).
Report is 1 commits behind head on master.

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #7027   +/-   ##
=======================================
  Coverage   99.60%   99.60%           
=======================================
  Files         488      488           
  Lines       46807    46818   +11     
=======================================
+ Hits        46622    46634   +12     
+ Misses        185      184    -1     

☔ 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.

[mlxd]

Thanks @lillian542
If the answer is no to my first question below, then no concerns with the rest, and will be fine to approve.

[andrijapau]

What if the measure is in a cond branch that has other operators around it, will this still detect it?

[mudit2812]

Changing my review based on @mlxd 's observations

[lillian542]

What if the measure is in a cond branch that has other operators around it, will this still detect it?

@andrijapau yes! On the tape we have a series of Conditional ops, each one with a single base.

[mlxd]

Thanks a bunch @lillian542
One suggestion, but happy to approve otherwise.

[mlxd]

Thanks a lot @lillian542