fix(fleet): normalize formatting before the sweep idempotency checks

[swackhamer]

Problem

Second follow-up to #691/#693. Measured 2026-08-11 in production, right
after #693's path-scoped fix deployed: PR #694 duplicated #692
anyway (byte-identical diffs, md5-verified).

The scoped diff #693 added did find a difference — but it was pure
rustfmt whitespace (a 3-line match arm collapsed to one line),
nothing semantic:

-            "note" => metadata.insert("DjVu:Note".to_string(), TagValue::new_string(value.clone())),
+            "note" => metadata.insert(
+                "DjVu:Note".to_string(),
+                TagValue::new_string(value.clone()),
+            ),

Root cause: #692's branch had already been through
format_sweep_branch's fmt-and-commit step in its own round. This
round's comparison ran on raw, pre-fmt worker output, because both
idempotency checks (the origin_ref one and #693's open-PR one) are
commit-to-commit diffs, and at the point they run, this round's branch
has never been through cargo fmt yet — that historically only
happened much later, right before the push. Unformatted content can
never tree-match a PR that already went through fmt, no matter how
many times the same gap gets re-solved with functionally identical
code.

Fix

Run format_sweep_branch once, early — right before the
origin_ref zero-delta check — so both idempotency checks compare
already-formatted content. The existing late call (after the evidence
table / judgment queue are built) is now usually a no-op ("already
cargo-fmt clean"); kept in place because commits_contributed resolves
each squad's contribution from merge_infos boundaries recorded
before any fmt call, so an early fmt commit can never leak into the
evidence table — but building the evidence table itself must not move
earlier than it already is.

Testing

New regression test reproduces the exact #694-vs-#692 shape: the
already-open PR's branch carries the fmt-normalized form of a fix,
this round's squad commit carries the identical fix pre-fmt. Confirmed
the test fails against the pre-fix code and passes with this
fix (verified via stash-and-rerun). One existing assertion (which of
the two format_sweep_branch calls reports "committed") updated to
match the new timing.

Full suite: test_overlord_sweep.py, 103 tests, all green.