Warning: This build has drifted.
The coverage report for this pull request build may be inaccurate because its base commit is no longer the HEAD of its target branch.
This means it includes changes from outside the original pull request, including, potentially, unrelated coverage changes.

    • Learn more: For more information on this, see Tracking coverage changes for pull request builds.
    • Fix now: For a quick fix, rebase this PR at GitHub. Your next report should be accurate.
    • Prevent going forward: To avoid this issue with future PRs, see these Recommended CI Configurations.

fix(cypher): optional shortestPath and shortestPath-as-value converge on one BFS (#581, #721)

OPTIONAL MATCH p = shortestPath(...) previously rewrote the clause to a
plain MATCH and swallowed every error into a fabricated {"result": null}
row with the wrong columns. Clause-only and anchored forms
(MATCH (a) OPTIONAL MATCH p = shortestPath((a)-...->(c))) now run the
shared shortestPath BFS with left-outer-join semantics: one row per seed,
projected values when a path exists, nulls when none, errors propagated.

shortestPath(...)/allShortestPaths(...) are now valid in value position
(RETURN length(shortestPath((a)-[:R*]->(b))), WITH, SET projections):
the projected-pattern validator masks the path-function arguments instead
of rejecting them, and the shared row evaluator runs the same BFS per row.
length() of an unbound path returns null instead of a fabricated 0.

Convergence: MATCH, OPTIONAL MATCH, and value forms all traverse through
e.shortestPath/e.allShortestPaths and project through buildPathContext;
clause-position detection shares extractShortestPathCall with
isShortestPathClause (new); the validator's mask reuses indexASCIIFold
(zero-alloc ASCII fold shared with equalFoldASCII).

Performance (M2 Max, -cpu=1 -count=3, BenchmarkStatementRouting):
  autocommit/simple_match_limit  main 161 allocs, 17.4-18.8 us/op
                                 branch 161 allocs, 17.5-17.7 us/op
  explicit_tx/simple_match_limit main 88 allocs, 11.5-12.0 us/op
                                 branch 88 allocs, 11.4 us/op
  No allocation or latency regression; extractShortestPathCall's
  upperASCII(funcName) copies were removed with constant upper-case
  literals after memprofile diffing showed the +2 allocs they added.

Tests: optional_shortest_path_test.go pins #581's five Neo4j outcomes
plus value forms in RETURN/WITH, autocommit and explicit transactions,
each run twice; full pkg/cypher suite, -race, and TCK ratchet
(7794/7794, 100%) pass.