Translating cudaq::compute_action to quake if the action contains a for-loop

[oryarnitzky-classiq]

Required prerequisites

• 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

It seems like cudaq-quake fails to translate calls to cudaq::compute_action to quake, when the action operand contains a for loop. The following code example demonstrates this:

#include <cudaq.h>

struct entry {
  auto operator()() __qpu__ {
    cudaq::qubit p;
    cudaq::qubit q;
    cudaq::compute_action(
    [&]() { x(p); },
    [&]() {
        for (size_t i = 0; i < 1; ++i) { y(q); }
    });
  }
};

int main() {
  auto counts = cudaq::sample(entry{});
  counts.dump();
  return 0;
}

Actually having quantum logic in the code is not needed, the following also fails:

#include <cudaq.h>

struct entry {
  auto operator()() __qpu__ {
    cudaq::compute_action(
    [&]() {  },
    [&]() {
        for (size_t i = 0; i < 1; ++i) {  }
    });
  }
};

int main() {
  auto counts = cudaq::sample(entry{});
  counts.dump();
  return 0;
}

Running the command cudaq-quake --emit-llvm-file a.cpp fails with the following error:
error: 'quake.compute_action' op operand #0 must be Proxy for a C++ callable instance in cc, but got 'i64'

When inspecting the quake code, it seems like the quake.compute_action doesn't access the correct argument, but instead if accesses the result of the addition from the loop step.

Steps to reproduce the bug

1. Create a file called a.cpp with the following content:

#include <cudaq.h>

struct entry {
  auto operator()() __qpu__ {
    cudaq::qubit p;
    cudaq::qubit q;
    cudaq::compute_action(
    [&]() { x(p); },
    [&]() {
        for (size_t i = 0; i < 1; ++i) { y(q); }
    });
  }
};

int main() {
  auto counts = cudaq::sample(entry{});
  counts.dump();
  return 0;
}

2. Try to run cudaq-quake --emit-llvm-file a.cpp

Expected behavior

The command should succeed and an appropriate quake file should be created

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

Not a regression

Environment

• CUDA Quantum version:
• Python version:
• C++ compiler:
• Operating system:

Suggestions

No response