Take the flamenet-e2e doc updates and regenerate the pages #223

2026-09-24T18:09:31Z

paulhitt commented

2026-09-24 18:09:31 +00:00

(Migrated from gitlab.com)

The three-part change .e2e-ref.md asks for, in one MR: bump the pin, regenerate with Scripts/build-doc-pages.sh, land together. The source change is already on flamenet-e2e main (!58).

What moved

spec.leaf and audit.leaf pick up three corrections from the claim-register audit.

AUDIT-SCOPE §2.7 used to send a reviewer after "no route accepts a plaintext body". No route is designed to carry one and every shipped client encrypts, but this relay never inspects payload — no length check, no base64 check, no shape check, just the global 4 MiB cap. So it cannot refuse a readable body, and the sentence invited an hour of looking for enforcement that is not there.

SPEC §14.6 lists what sealed sender leaves open: sender IP, recipient visibility, spam. It did not list attachments, which is the one that gives away both ends. POST /e2e/attachments and GET /e2e/attachments/{id} are ordinary authenticated routes, so for any sealed conversation carrying a photo or a voice note an operator watching requests has the pair without touching the envelope.

SPEC §19.4 now says who holds the key that signs the key log. The head signature is what makes a fork evidence between two clients who compare heads, and nothing at all to a client on its own, because this relay holds FN_ISSUER_PRIVATE_KEY. The witnesses field it serves has no submission route and is never written, and no client reads it, so the spec calls it reserved rather than a mechanism.

Why `doc-pages-head` was red, and that it predates this work

doc-pages checks generated pages against the pinned commit and has been green throughout. doc-pages-head checks against e2e HEAD, is allow_failure: true, and exists to say an update is waiting. It was red because the pin was four doc commits behind before any of the audit work started. This clears it.

check-doc-pages.py: 9 generated pages match their sources.
751 tests in 76 suites, no failures.

🤖 Generated with Claude Code

The three-part change `.e2e-ref.md` asks for, in one MR: bump the pin, regenerate with `Scripts/build-doc-pages.sh`, land together. The source change is already on flamenet-e2e `main` (!58). ## What moved `spec.leaf` and `audit.leaf` pick up three corrections from the claim-register audit. **AUDIT-SCOPE §2.7** used to send a reviewer after "no route accepts a plaintext body". No route is *designed* to carry one and every shipped client encrypts, but this relay never inspects `payload` — no length check, no base64 check, no shape check, just the global 4 MiB cap. So it cannot refuse a readable body, and the sentence invited an hour of looking for enforcement that is not there. **SPEC §14.6** lists what sealed sender leaves open: sender IP, recipient visibility, spam. It did not list attachments, which is the one that gives away both ends. `POST /e2e/attachments` and `GET /e2e/attachments/{id}` are ordinary authenticated routes, so for any sealed conversation carrying a photo or a voice note an operator watching requests has the pair without touching the envelope. **SPEC §19.4** now says who holds the key that signs the key log. The head signature is what makes a fork evidence *between two clients who compare heads*, and nothing at all to a client on its own, because this relay holds `FN_ISSUER_PRIVATE_KEY`. The `witnesses` field it serves has no submission route and is never written, and no client reads it, so the spec calls it reserved rather than a mechanism. ## Why `doc-pages-head` was red, and that it predates this work `doc-pages` checks generated pages against the **pinned** commit and has been green throughout. `doc-pages-head` checks against e2e HEAD, is `allow_failure: true`, and exists to say an update is waiting. It was red because the pin was four doc commits behind before any of the audit work started. This clears it. --- `check-doc-pages.py`: **9 generated pages match their sources.** **751 tests in 76 suites, no failures.** 🤖 Generated with [Claude Code](https://claude.com/claude-code)

paulhitt commented

2026-09-25 12:35:19 +00:00

(Migrated from gitlab.com)

mentioned in commit 3f0e24bc0a

mentioned in commit 3f0e24bc0af2c497d0ee6155039fab200d99bcd1

paulhitt (Migrated from gitlab.com) merged commit 3f0e24bc0a into main

2026-09-25 12:35:19 +00:00

Sign in to join this conversation.

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: extant2000/flamenet-relay#223

Take the flamenet-e2e doc updates and regenerate the pages #223

What moved

Why doc-pages-head was red, and that it predates this work

Why `doc-pages-head` was red, and that it predates this work