Add overload to handle the case when the user writes:

[schweitzpgi]

  cudaq::control(Kernel{}, {ctrl1, ctrl2}, args...);

Get vector<reference_wrapper> through the bridge.

Add simple test.

[amccaskey]

Will you check that the following makes it through the bridge with this PR:

struct givens {
  void operator()(double lambda, cudaq::qubit &q, cudaq::qubit &r) __qpu__ {
    ry(M_PI_2, q);
    ry(M_PI_2, r);
    z<cudaq::ctrl>(q, r);
    ry(lambda, q);
    ry(-lambda, r);
    z<cudaq::ctrl>(q, r);
    ry(-M_PI_2, q);
    ry(-M_PI_2, r);
  }
};

__qpu__ void qnppx(double theta, cudaq::qubit &q, cudaq::qubit &r,
                   cudaq::qubit &s, cudaq::qubit &t) {
  x<cudaq::ctrl>(r, q);
  x<cudaq::ctrl>(s, t);
  cudaq::control(givens{}, {q, t}, theta, r, s);
  x<cudaq::ctrl>(r, q);
  x<cudaq::ctrl>(s, t);
}

[schweitzpgi]

Will you check that the following makes it through the bridge with this PR:

struct givens {
  void operator()(double lambda, cudaq::qubit &q, cudaq::qubit &r) __qpu__ {
    ry(M_PI_2, q);
    ry(M_PI_2, r);
    z<cudaq::ctrl>(q, r);
    ry(lambda, q);
    ry(-lambda, r);
    z<cudaq::ctrl>(q, r);
    ry(-M_PI_2, q);
    ry(-M_PI_2, r);
  }
};

__qpu__ void qnppx(double theta, cudaq::qubit &q, cudaq::qubit &r,
                   cudaq::qubit &s, cudaq::qubit &t) {
  x<cudaq::ctrl>(r, q);
  x<cudaq::ctrl>(s, t);
  cudaq::control(givens{}, {q, t}, theta, r, s);
  x<cudaq::ctrl>(r, q);
  x<cudaq::ctrl>(s, t);
}

Thanks for the test. I'm getting a failure on this code. Investigating. Update: added fix.

[amccaskey]

Thanks for the bug fix, the code snippet works for me now.

[amccaskey]

Approved pending CI

[bettinaheim]

Rel notes: see also #35.