Skip to content

Field notes: live-505c wire observations from a second server implementation - #18

Open
M0LTE wants to merge 1 commit into
mainfrom
field-notes-505c-diff
Open

Field notes: live-505c wire observations from a second server implementation#18
M0LTE wants to merge 1 commit into
mainfrom
field-notes-505c-diff

Conversation

@M0LTE

@M0LTE M0LTE commented Jun 11, 2026

Copy link
Copy Markdown
Collaborator

Three new field-notes observations (13–15) from building a second server-side RHPv2 implementation (pdn) and wire-diffing it against the pinned xrouter container (ghcr.io/packethacking/xrouter, image label 505c) and against RHPTEST:

  • 13 — a second listen on an already-listening socket: RHPTEST (tested against v505d) says error 16, the live 505c wire answers idempotent errCode: 0. Either behaviour changed between builds or RHPTEST documents undelivered intent; a 505d container would settle it empirically.
  • 14send on a listening socket: 17 on the wire, 16 per RHPTEST. Same class of conflict; pdn sided with 16, flagged for confirmation.
  • 15 — two listeners can claim one callsign (live wire answers Ok), leaving accept routing undefined; errCode: 9 suggested as the deterministic refusal.

Also a postscript on observation 4 (the hello/helloReply proposal now has a live implementation, with the unknown-type fallback delivering the predicted backwards compatibility), a new v2.1-table row covering 13–15, and a caveat on the primer's RHPTEST listener bullet cross-referencing the conflict.

mkdocs build verified clean; the new cross-page anchor resolves.

🤖 Generated with Claude Code

Three new observations from building a second server-side RHPv2 implementation (pdn) and wire-diffing it against the pinned xrouter container: RHPTEST and the live 505c binary disagree on re-listen (16 vs idempotent Ok) and on send-on-listener (16 vs 17), and a second listener on an already-claimed callsign answers Ok with undefined accept routing (errCode 9 suggested). Plus a postscript on observation 4: the hello/helloReply proposal now has a live implementation, with the unknown-type fallback providing the predicted backwards compatibility. Primer's RHPTEST listener bullet gains a caveat cross-referencing the conflict.

Co-Authored-By: Claude Opus 4.8 (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