Restore prefix-penalty scoring semantics and remove dead scoring allocation

[Copilot]

Prefix-penalty logic in Fuzzy had become a permanent no-op (0 multiplier + 0 cap), so late-starting matches were not downgraded versus start-aligned matches. The associated test also derived expectations from those same constants, making it unable to detect regressions.

• Scoring behavior: make prefix penalty real again

  • Set unmatchedPrefixLetterPenalty to -1 and maxPrefixPenalty to -5 in Fuzzy.
  • Preserves intended behavior: matches that begin deeper in the subject now score lower on prefix positioning.

• Scoring hot path cleanup

  • Removed unused matchedIndices list and writes in CalculateScore.
  • Eliminates per-call allocation/work that had no effect on outputs.

• Test hardening: independent expectations + observable ranking

  • Rewrote prefix-penalty unit expectation to fixed values (no recomputation from production constants).
  • Added cap-coverage test for long prefixes.
  • Added end-to-end scoring-order test to assert early match outranks late match.

// Before: no-op
internal const int unmatchedPrefixLetterPenalty = 0;
internal const int maxPrefixPenalty = 0;

// After: effective penalty with cap
internal const int unmatchedPrefixLetterPenalty = -1;
internal const int maxPrefixPenalty = -5;

• Fixes Prefix-penalty scoring is a permanent no-op, and its test can't catch that #66

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud