[Navigation] Break repeated e1m6 lower-door route loops

Open
  • Is something wrong?
#643 0 comments 0 reactions 0 assignees View on GitHub

Maintainers usually reply within 1 day

Nobody has claimed this yet.

Assessment

Difficulty
5/5
Estimated time
Over a week
Newbie friendliness
8/100
Issue type
Bug
Clarity
Needs clarification
Activity status
Active
Tech stack
c, rust
Domain
ai-infra-agents

Research direction

The payload points at navigation telemetry logs (runs/broad_e6be2_e1m6_3.log, runs/broad_e6fn_e1m6_1.log) and the decision docs (docs/e1m6-decisions.md, runs/e1m6-decisions.results.json), but names no source file or test for the door routing code. Start by reading the logs near nodes 89 and 101 to see the door phase and accepted route. Done would mean repeated lower-door traversal is broken or replanned before prolonged confinement, with upper-platform handoff and water escape still validated.

Written by the indexing model from the issue text.

Description

bug navigation

Observation

The broader e1m6 study identifies lower-door recirculation independent of the upper lift seam. In final runs/broad_e6be2_e1m6_3.log, the native brief reports Romero confined for 33.6 seconds inside a 220-unit box at 146 units/second around -197 1226. Its door-classified stall hotspot is near -112 1273 68, nearest graph node 101, directed reach 95%. Repeated stall targets include 17 1265 24 and -175 1233 24.

The first rejected candidate also confines Romero for 48.6 seconds around -214 1239. That older tape is runs/broad_e6fn_e1m6_1.log; its controller and statistical warning remain separate in runs/e1m6-dock-initial.results.json. The final source does not claim to fix this lower-door loop.

Final decision source SHA256: 71cecc207e97c64d6b8efecceb44985c00b982f280a6447c5c381a6ac15050a3. Final evidence: runs/e1m6-decisions.results.json and docs/e1m6-decisions.md, published in https://github.com/saworbit/argus/pull/645. Speed alone conceals this failure, and broad directed reach does not prove execution through the door.

Next steps and acceptance

  • Record the actual door phase, route/goal ownership and accepted node sequence near nodes 89/101 without steering the observer.
  • Reproduce the loop with production routing and a continuing goal, distinguishing closed-door handling, a body-clearance refusal and route reselection.
  • Break the repeated traversal or replan to a feasible route before prolonged confinement. Do not count a new goal selection or rescue warp as success.
  • Preserve the upper-platform handoff and water escape; validate at both cadences and across the broader maps with unchanged gates.

This is separate from #622's upper-platform transfer and #249's co-op companion problem. No causal regression claim is made from a single ordinary tape.

Dominant language
Rust
Stars
4
Forks
0
Avg merge
30m
Merged PRs (30d)
130

Getting set up

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

More from saworbit/argus

All issues in saworbit/argus

Similar issues

More Rust issues

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.