Fix schema validators after collection pre-load changes length

[nyxst4ck]

Summary

• Avoid zipping deserialized many=True items with the original raw collection when a per-item @validates_schema hook did not request pass_original.
• Add a regression test for @pre_load(pass_collection=True) changing collection length before per-item schema validation.
• Add @nyxst4ck to AUTHORS.rst per the contribution guidelines.

Root cause

For many=True schema validation, _invoke_schema_validators always used zip(data, original_data, strict=True) for per-item validators. If a collection-level pre-load hook filtered or otherwise changed the number of items, per-item schema validators that did not use pass_original still failed with ValueError from the strict zip, even though original_data is unused in that case.

Test Plan

• uv run pytest tests/test_decorators.py -k collection_pre_load_can_change_length_before_item_schema_validation -q
• uv run pytest tests/test_decorators.py -q
• uv run pytest -q -k "not test_from_timestamp_with_overflow_value"
• uv run tox -e lint

Also ran uv run pytest -q; it had one unrelated Windows/Python assertion failure in tests/test_utils.py::test_from_timestamp_with_overflow_value because the platform error message was Error converting value to datetime instead of the expected regex.