spikerj/spikersoft-issues
1
0
Fork 0
Code Issues 196 Pull Requests Actions Packages Projects Releases Wiki Activity

[GameServer][Bug] Hierarchical A* never worked: region identity collapsed to (0,0,0) via a placeholder IntToRegion — long-range pathfinding fell back to straight lines through obstacles #646

New Issue
Closed
opened 2026-07-17 07:21:03 +00:00 by spikerj · 1 comment
spikerj commented 2026-07-17 07:21:03 +00:00
Owner
Copy Link
Copy Source

Found via SonarQube S1172 triage (2026-07-17) — the rule flagged IntToRegion's unused 'hash' parameter; reading it revealed the function is a placeholder that ALWAYS returns (0,0,0) ('This is a hack... Placeholder'), and it sits in the middle of the abstract A* loop: every dequeued region 'was' the origin, so the search could never advance. AdaptivePathfinder selects HierarchicalAStar for exactly the hard cases (30+ obstacles, long paths, :204), so long-range pathfinding around dense obstacle fields effectively produced degenerate results since the class was written.

Fix (PR incoming): the hash round-trip is unnecessary — the region tuple is now the priority-queue element directly (also eliminating the HashCode.Combine collision risk the 'proper bidirectional map' TODO would have carried). RegionToInt/IntToRegion deleted.

Regression guard: new HierarchicalAStarTests — a wall with a side gap forces the abstract search to actually traverse regions; verified FAILING on the pre-fix code (falls back to a 2-waypoint straight line through the wall) and passing after. Full GameServer suite (915) passes; slnf clean.

**Found via SonarQube S1172 triage (2026-07-17)** — the rule flagged IntToRegion's unused 'hash' parameter; reading it revealed the function is a placeholder that ALWAYS returns (0,0,0) ('This is a hack... Placeholder'), and it sits in the middle of the abstract A* loop: every dequeued region 'was' the origin, so the search could never advance. AdaptivePathfinder selects HierarchicalAStar for exactly the hard cases (30+ obstacles, long paths, :204), so long-range pathfinding around dense obstacle fields effectively produced degenerate results since the class was written. **Fix (PR incoming):** the hash round-trip is unnecessary — the region tuple is now the priority-queue element directly (also eliminating the HashCode.Combine collision risk the 'proper bidirectional map' TODO would have carried). RegionToInt/IntToRegion deleted. **Regression guard:** new HierarchicalAStarTests — a wall with a side gap forces the abstract search to actually traverse regions; verified FAILING on the pre-fix code (falls back to a 2-waypoint straight line through the wall) and passing after. Full GameServer suite (915) passes; slnf clean.
spikerj commented 2026-07-17 07:47:56 +00:00
Author
Owner
Copy Link
Copy Source

Resolved in spikersoft-backend PR #351 (merged, 7/7 checks green). Region tuples are now the abstract-search queue element directly — the placeholder IntToRegion that collapsed every dequeued region to (0,0,0) is deleted along with the hash round-trip and its HashCode.Combine collision risk. The new HierarchicalAStarTests wall-with-side-gap guard is proven non-vacuous: it fails on the pre-fix code (2-waypoint straight line through the wall) and passes after. Long-range pathfinding through dense obstacle fields does real hierarchical search for the first time. Closing.

Resolved in spikersoft-backend PR #351 (merged, 7/7 checks green). Region tuples are now the abstract-search queue element directly — the placeholder IntToRegion that collapsed every dequeued region to (0,0,0) is deleted along with the hash round-trip and its HashCode.Combine collision risk. The new HierarchicalAStarTests wall-with-side-gap guard is proven non-vacuous: it fails on the pre-fix code (2-waypoint straight line through the wall) and passes after. Long-range pathfinding through dense obstacle fields does real hierarchical search for the first time. Closing.
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
agentic
Actionable surfaced by the agentic architecture & security review
 bug
Something is not working
 duplicate
This issue or pull request already exists
 enhancement
New feature
 help wanted
Need some help
 invalid
Something is wrong
 question
More information is needed
 sonarqube
Code quality issue surfaced by SonarQube (bug / code smell)
 wontfix
This won't be fixed
No labels
Milestone
No items
No Milestone
Projects
Clear projects
No projects
Assignees
Clear assignees
spikerj (Joseph Spiker)
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: spikerj/spikersoft-issues#646