[Python] Nested kernel with `exp_pauli` failed

[1tnguyen]

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

The following code failed with error error: 'func.call' op kernel call was not inlined, possible recursion in call tree

import cudaq

@cudaq.kernel           
def kernel(
        num_qubits: int,
        coefficients: list[complex] = None,
        words: list[cudaq.pauli_word] = None,
    ):
    K = 5
    t = 1.0
    # create the qubit vector
    qubits = cudaq.qvector(num_qubits)
    
  
    # determin delta for one trotter step
    dt = t / K
    
    # Apply K Trotter steps
    for _ in range(K): 
        trotter(qubits, dt, coefficients, words)
   
@cudaq.kernel
def trotter(
            qubits: cudaq.qview,
            dt: float, 
            coefficients: list[complex],
            words: list[cudaq.pauli_word]
        ):
   
    for i in range(len(coefficients)):
        exp_pauli(coefficients[i].real * dt, qubits, words[i])   # this crashes jupyter kernel on draw

num_qubits = 4
words =  [cudaq.pauli_word("XXXX")]
coefficients = [1.0]

counts = cudaq.sample(kernel, num_qubits, coefficients, words)

This used to work, as reported by @cuMikeBrown. Also, confirmed that it worked in 0.9.1.

Steps to reproduce the bug

See above

Expected behavior

Code runs

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

0.9.1

Environment

• CUDA-Q version: Latest main
• Python version: 3.10
• C++ compiler: 11
• Operating system: Ubuntu

Suggestions

No response

[bmhowe23]

git bisect reveals this was broken in #2542 (cc @schweitzpgi)

Also, just as an FYI, I believe the comment about "# this crashes jupyter kernel on draw" is no longer accurate once #2678 went in.