PYTHON-5993 Add OpenTelemetry operation spans for cursor getMores

[blink1073]

PYTHON-5993

Last of four PRs splitting #2964. Base is PYTHON-5947-otel-3-unified.

	Branch	Contents
1	PYTHON-5947-otel-1-operations	operation spans
2	PYTHON-5947-otel-2-transactions	transaction spans
3	PYTHON-5947-otel-3-unified	unified runner and vendored fixtures
4	PYTHON-5947-otel-4-getmore	getMore spans (this PR)

Changes in this PR

Gives each caller-driven getMore an operation span of its own, which the specification requires: an application may do unrelated work between batches, so nesting every getMore under the call that created the cursor would misrepresent the timing.

Implements DRIVERS-3598 (specifications#1973), which settles this: the spec previously said only that a command span nests under "the corresponding driver operation span", which is unambiguous only while an operation sends one command. Both readings passed the suite because no fixture exercised getMore.

• One operation span per caller-driven batch, sibling to the cursor-creating call's span rather than nested under it.
• A call that creates a cursor and drains it itself, such as list_collection_names or index_information, keeps a single span covering every getMore it sends.
• db.mongodb.cursor_id holds the id the driver sent, even when the reply returns 0, and is omitted rather than reported as 0 for a cursor-creating command that leaves no cursor open.
• No span is scoped to a cursor's lifetime, so a cursor that is never exhausted leaves nothing unfinished.
• The client bulk-write results cursor reuses the enclosing bulkWrite span instead of adding spurious siblings.
• A command cursor exhausted by its first batch closes its span promptly rather than at garbage collection.
• Vendored the spec's getMore fixture, held back from PR 3, plus its prose tests 3 and 4 for what the unified format cannot express.

Change streams deliberately get neither treatment. A change stream can tail indefinitely, so a span covering its lifetime would never end or be exported; each getMore gets a short-lived sibling span instead. There is a comment on the code saying so.

The fixture is a local deviation until specifications#1973 merges. get_more.json is not on mongodb/specifications master yet, so running resync-specs.sh before that lands will delete it. The file here is byte-identical to the spec PR. That PR's tree-wide ignoreExtraSpans rule, which this fixture's negative assertions rely on, is in PR 3.

Completes the split: this branch's tree is identical to #2964 apart from a .gitignore entry for local worktrees, which the stack drops.

Test Plan

• 14 new tests each side: a span per caller-driven getMore, single-span coverage for internally drained cursors, bulk-write cursor nesting, cursors abandoned part-way, change-stream namespaces, and the spec's prose tests 3 and 4.
• The getMore fixture, which passes unmodified. It caught a real divergence on its first run: the driver had used a single cursor-lifetime span, the reading DRIVERS-3598 rules out.
• Locally against a replica set: all 196 otel tests across the four PRs pass. A wider sweep of the async suite plus the sync otel modules gives 3,319 passed; the 23 non-passing are Atlas-only suites, mockupdb not being installed, and one timing-sensitive test, and otel reproduces the same set unchanged.
• just lint clean.

Checklist

Checklist for Author

• Did you update the changelog (if necessary)?
• Is there test coverage?
• Is any followup work tracked in a JIRA ticket? If so, add link(s).

Checklist for Reviewer

• Does the title of the PR reference a JIRA Ticket?
• Do you fully understand the implementation? (Would you be comfortable explaining how this code works to someone else?)
• Is all relevant documentation (README or docstring) updated?