Unexpected results in `global` register for kernels that both `mx` and `mz`

[bmhowe23]

Required prerequisites

• Consult the security policy. If reporting a security vulnerability, do not report the bug using this form. Use the process described in the policy to report the issue.
• Make sure you've read the documentation. Your issue may be addressed there.
• Search the issue tracker to verify that this hasn't already been reported. +1 or comment there if it has.
• If possible, make a PR with a failing test to give us a starting point to work on!

Describe the bug

Putting both mx and mz in kernels can result in incomplete results in the __global__ register.

Steps to reproduce the bug

Consider the following kernel:

import cudaq

@cudaq.kernel
def kernel():
    qubits_a = cudaq.qvector(2)
    qubit_b = cudaq.qubit()
    mz(qubits_a)
    mx(qubit_b)

counts = cudaq.sample(kernel)
print(counts)

That will produce something like this:

{ 0:490 1:510 }

Expected behavior

The specification/docs say that

if a qubit is measured in the middle of a kernel and subsequent operations change the state of the qubit, the qubit will be implicitly re-measured at the end of the kernel, and that re-measured value is the value that will appear in the global register.

It is unclear if "re-measured" means measured in the same basis as the original measurement or not, but regardless, the example above shows that the measurements for qubits_a can be lost. They should be in the resulting global register.

Is this a regression? If it is, put the last known working version (or commit) here.

Not a regression

Environment

• CUDA-Q version: latest (fe1396b at the time of the creation of this issue)
• Python version: 3.10
• C++ compiler:
• Operating system:

Suggestions

No response