Jake Lishman

This needs to be an error in practice; there's no way to communicate the creg / variable requirements of an `Instruction`'s inner definition up to the main circuit, so allowing...

Really, `circuit_to_instruction` shouldn't even allow `c_if` in it - in practice, that hasn't ever worked fully correctly when registers are involved.

There should be no need to use `circuit_to_instruction`. If you want to build up circuit parts and then combine them into a larger circuit, you can use `QuantumCircuit.compose` to "inline"...

There's no particular guarantee that any classical handling was working correctly before if `circuit_to_instruction` was used (also via `QuantumCircuit.append`) - the circuit would silently not notice if the classical control...

We didn't get this done for 1.0, so we'll have to revisit for 1.1 and see what's reasonable for us to do within the stability policy.

Thanks for the report - looks like this Qiskit built-in gate needs some special-case logic in QPY (similar to other non-standard built-ins).

Thanks for this report. I agree, this just looks like an oversight in QPY - these multi-controlled gate objects aren't heavily used in most cases (they're more properly compiler syntheses,...

Now rebased over `main`. Apparently I touched a couple of unrelated comments at some point that I mistakenly thought were part of this PR not a previous one, but oh...

@dieris: this PR is generally needed on Qiskit (for Qiskit 1.1 in 3 months' time), but I suspect might have some effects for work you did recently on `qiskit-ibm-provider` (and/or...

Fwiw I have a proof of concept patch to rustworkx and Qiskit locally that solves this, but the issue is to remind us to sort it out properly after the...