perf: drop per-element copies in Clarity value deserialization

[francesco-stacks]

Description

This removes the main memory and CPU overheads in Clarity value deserialization. No behavior changes...same bytes accepted and rejected, same resulting values

Three changes:

• construct_parent_list_type no longer collects a TypeSignature for every list element before computing the entry type. It folds while iterating, and skips the fold when the next type equals the accumulator (safe because least_supertype_v2_1(T, T) == T, pinned by a new test).
• The deserializer used to pop and re-push the whole stack frame for every element it read. Frames are now updated in place and only popped when a container completes.
• next_expected_type returned a deep clone of the expected type for every element. It now returns a reference.

On a 1M-element bool list (counting-allocator harness, vs current develop): untyped reads go from 141 MiB / 62 ms to 61 MiB / 28 ms. Typed reads are 16-21% faster with the same peak.

Applicable issues

• fixes #

Additional info (benefits, drawbacks, caveats)

Checklist

• Test coverage for new or modified code paths
• For new Clarity features or consensus changes, add property tests (see docs/property-testing.md)
• Changelog fragment(s) or "no changelog" label added (see changelog.d/README.md)
• Required documentation changes (e.g., rpc/openapi.yaml for RPC endpoints, event-dispatcher.md for new events)
• New clarity functions have corresponding PR in clarity-benchmarking repo

[Copilot]

Pull request overview

Improves performance of Clarity Value deserialization by eliminating avoidable per-element allocations/copies while preserving accepted/rejected bytes and resulting values, and adds targeted tests to pin correctness of the new fast paths.

Changes:

• Stream construct_parent_list_type()/parent_list_type() entry-type computation via a shared fold iterator (with an equality fast-path).
• Update the deserializer to mutate stack frames in place (avoid pop/re-push per element) and make next_expected_type() return references instead of deep clones.
• Add tests for least_supertype_v2_1 idempotence and for tuple truncation during unwind, plus a changelog entry.

Reviewed changes

Copilot reviewed 5 out of 5 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
clarity-types/src/types/signatures.rs	Streams list entry-type folding via a shared iterator-based helper to avoid per-element type collection.
clarity-types/src/types/serialization.rs	Removes per-element frame pop/re-push and avoids per-element deep clones of expected types during deserialization.
clarity-types/src/tests/types/signatures.rs	Adds idempotence/property-style coverage ensuring the new equality fast-path is sound and that streamed/collected list type construction matches.
clarity-types/src/tests/types/serialization.rs	Adds a regression test for EOF while reading the next tuple key during unwind.
changelog.d/clarity-value-deser-perf.changed	Documents the deserialization performance improvement.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[brice-stacks]

This looks great!