Create a secure internal path for custom prefixes in registers.

[raynelfss]

Summary

The following commits add the internal method _new_with_prefix, which allows for the one-time replacement of the set prefix name and use of the same instance counter.

Details and comments

Prior implementations would replace the Register's prefix attribute in place which is an unsafe operation. The following commits add a secure path for a provisional replacement of a register's prefix name to fix changed unsafe behavior from #13860 and attempts to fix #14003.

[qiskit-bot]

One or more of the following people are relevant to this code:

• @Qiskit/terra-core

[coveralls]

Pull Request Test Coverage Report for Build 13925424331

Warning: This coverage report may be inaccurate.

This pull request's base commit is no longer the HEAD commit of its target branch. This means it includes changes from outside the original pull request, including, potentially, unrelated coverage changes.

• For more information on this, see Tracking coverage changes with pull request builds.
• To avoid this issue with future PRs, see these Recommended CI Configurations.
• For a quick fix, rebase this PR at GitHub. Your next report should be accurate.

Details

• 16 of 17 (94.12%) changed or added relevant lines in 3 files are covered.
• 961 unchanged lines in 19 files lost coverage.
• Overall coverage decreased (-0.06%) to 88.061%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
qiskit/circuit/quantumcircuit.py	1	2	50.0%

Files with Coverage Reduction	New Missed Lines	%
crates/accelerate/src/gate_metrics.rs	1	97.06%
crates/accelerate/src/unitary_synthesis.rs	1	94.69%
qiskit/transpiler/passes/optimization/template_matching/template_substitution.py	1	93.27%
crates/accelerate/src/synthesis/clifford/bm_synthesis.rs	2	99.22%
crates/accelerate/src/synthesis/clifford/greedy_synthesis.rs	3	97.92%
crates/qasm2/src/lex.rs	3	92.73%
crates/accelerate/src/remove_identity_equiv.rs	5	94.44%
crates/accelerate/src/gate_direction.rs	7	96.95%
qiskit/circuit/tools/pi_check.py	9	91.3%
qiskit/circuit/parameterexpression.py	10	96.74%

Totals
Change from base Build 13813281684:	-0.06%
Covered Lines:	72592
Relevant Lines:	82434

---

💛 - Coveralls

[ElePT]

It would be nice to have a unit test for this path. The reproducing code that I found looks quite hacky, and this is probably why the issue hadn't surfaced until now. It's a test for the Aer estimator primitive that builds a circuit, adds measurements (via circuit.measure_all()), and then calls a part of the aer estimator code that does:

  circuit = self._circuits[i].copy()
  circuit.measure_all()
  .... # stuff
  circuit.remove_final_measurements()

Calling measure_all() twice after a shallow copy doesn't look like a great thing to do, but I guess we shouldn't hit this error if someone does it. I can see if I can isolate this into a more reasonable example.

[kevinhartman]

Looks good to me after @ElePT's comment is addressed!

[ElePT]

Thanks a lot! The PR looks good to me too. Here is the minimal reproducible code for the issue, I can't add it as a direct suggestion but I think it's enough to test this fix:

from qiskit import QuantumCircuit

qc = QuantumCircuit(1)
qc.measure_all()
qc.measure_all()

[kevinhartman]

Looks like you've got a linter error in CI as well 🙂

[ElePT]

LGTM, I found a tiny typo in the test comment but it's not important.