test: cover Taproot standardness rules in mempool_accept #35882

IsWitnessStandard rejects tapscript stack items over 80 bytes and an empty control block, and no test asserts bad-witness-nonstandard for either. With the 80-byte check removed, feature_taproot.py failed on only 6 of the ten seeds I tried, and with the empty control block check removed it can't fail at all, since none of its spenders has an empty control block. Without that check the node segfaults reading control_block[0].

This adds a case for each to mempool_accept.py, by mutating the witness of the P2TR script-path spend the file already builds. Each case fails with its policy.cpp branch disabled and passes with it restored.

<!--e57a25ab6845829454e8d69fc972939a-->

The following sections might be updated with supplementary metadata relevant to reviewers and maintainers.

<!--006a51241073e994b41acfe9ec718e94-->

External sites

• Code coverage & benchmarks: https://corecheck.dev/bitcoin/bitcoin/pulls/35882
• Ralph review <img src=https://review.fish.foo/assets/bitcoin/bitcoin/35882.png alt="Ralph review status"> https://git.fish.foo/bitcoin/bitcoin/pulls/35882#author-ralph

<!--021abf342d371248e50ceaed478a90ca-->

Reviews

See the guideline and AI policy for information on the review process. A summary of reviews will appear here.

<!--5faf32d7da4f0f540f40219e4f7537a3-->

These mutants already seem to be killed by feature_taproot.py. Why add these here?

You're right about the annex. feature_taproot.py kills that mutant on all ten seeds I tried, but the 80-byte one on only 6, since tapscript/input81limit isn't in spenders_taproot_nonstandard().

None of its spenders has an empty control block, and without the control_block.empty() check testmempoolaccept on the new case segfaults the node. Dropped the annex case.