Skip to content

docs(study): session-router graph-directed analysis (#1279)#6

Merged
Xepayac merged 7 commits intodevfrom
1279-study-session-router
Apr 4, 2026
Merged

docs(study): session-router graph-directed analysis (#1279)#6
Xepayac merged 7 commits intodevfrom
1279-study-session-router

Conversation

@Xepayac
Copy link
Copy Markdown
Owner

@Xepayac Xepayac commented Apr 4, 2026

Summary

  • Lab-book style capability study documenting the full SESSION_ROUTER lifecycle
  • Living document — entries will be added as work continues
  • Tracks TRUGS-DEVELOPMENT#1279

Contents

Test plan

  • Review narrative accuracy against commit history
  • Verify all commit hashes referenced exist

🤖 Generated with Claude Code

Xepayac and others added 7 commits April 3, 2026 18:20
…279)

Lab-book style paper documenting the full lifecycle:
curiosity → 3 TRUG analysis passes → 3 upstream PRs → complete
rewrite → 3 audit cycles → proposal → continued development.

Tracks TRUGS-DEVELOPMENT#1279.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Recursive AAA specification → GitHub Copilot in Codespace → Go
implementation. Clean room wall: spec author saw upstream, implementer
sees only AAA. Proves AAA as cross-agent specification format.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Full study sent to Session Foundation before clean room begins.
Two outcomes, both acceptable: they challenge with full evidence
(burden is theirs), or silence = acquiescence.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Phase 1: Complete C++ rewrite as GPL-3.0 gift to Session Foundation.
Phase 2: Clean room Go reimplementation via AAA + Copilot, owned by
TRUGS LLC. Phase 1 builds understanding and credibility. Phase 2
produces the business asset.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Token dilution context in motivation. State of the project: zero tests
on active namespace, zero fuzz coverage, acknowledged vulns in TODOs,
broken exit mode, 2 maintainers on 37K lines. The offer and the
alternative. Go rewrite under Apache 2.0.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
We set out to write tests for untested code. What we found:
- W1: call_get deadlock trap (undocumented inline execution)
- W3: cross-loop race window in outbound control stream creation
- W5: two different control stream paths (outbound vs inbound)
- W6: datagram handler use-after-free (no canary protection)
- W8: find_cc legacy path leaks requests (upstream bug)
- 106 tests designed across 11 categories
- "These are the findings of someone trying to write tests for
  code that has never been tested."

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
@Xepayac Xepayac merged commit 606a3e3 into dev Apr 4, 2026
Xepayac added a commit that referenced this pull request Apr 4, 2026
…pat)

Keep both: test_wire_compat outside QUIC guard (no QUIC needed),
test_link/exit/node inside QUIC guard (need oxen-libquic).

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant