One MR, all three parts, which is what .e2e-ref.md asks for: the pin
moves, the pages regenerate, and they land together. Splitting them is
what turns the doc gate red for whoever pushes next.
.e2e-ref e082d027 → d2ebcdf5, which carries e2e!29 — the protocol
version table now has a v3 column, and says what has been RUN rather than
what is present in the source: this engine implements v3, two Android
phones have negotiated it since 0.8, and the iPhone build carries it and
has not shipped, so no iPhone in a conversation today is running a
post-quantum ratchet.
I made this change in engine.leaf first, which is generated. Its own
header says edits there are deleted the next time the build script runs —
and the doc-pages gate PASSED on the hand-edited page, because it checks
whether the source moved since generation, not whether the page was
edited by hand. A green gate would have told me nothing and the fix would
have vanished at the next regeneration.
Gate re-run against a clone checked out at the new pin: 9 generated pages
match their sources. 670 tests.
One MR, all three parts, which is what `.e2e-ref.md` asks for: the pin
moves, the pages regenerate, and they land together. Splitting them is
what turns the doc gate red for whoever pushes next.
`.e2e-ref` e082d027 → d2ebcdf5, which carries e2e!29 — the protocol
version table now has a v3 column, and says what has been RUN rather than
what is present in the source: this engine implements v3, two Android
phones have negotiated it since 0.8, and the iPhone build carries it and
has not shipped, so no iPhone in a conversation today is running a
post-quantum ratchet.
I made this change in `engine.leaf` first, which is generated. Its own
header says edits there are deleted the next time the build script runs —
and the doc-pages gate PASSED on the hand-edited page, because it checks
whether the source moved since generation, not whether the page was
edited by hand. A green gate would have told me nothing and the fix would
have vanished at the next regeneration.
Gate re-run against a clone checked out at the new pin: 9 generated pages
match their sources. 670 tests.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
One MR, all three parts, which is what
.e2e-ref.mdasks for: the pinmoves, the pages regenerate, and they land together. Splitting them is
what turns the doc gate red for whoever pushes next.
.e2e-refe082d027 → d2ebcdf5, which carries e2e!29 — the protocolversion table now has a v3 column, and says what has been RUN rather than
what is present in the source: this engine implements v3, two Android
phones have negotiated it since 0.8, and the iPhone build carries it and
has not shipped, so no iPhone in a conversation today is running a
post-quantum ratchet.
I made this change in
engine.leaffirst, which is generated. Its ownheader says edits there are deleted the next time the build script runs —
and the doc-pages gate PASSED on the hand-edited page, because it checks
whether the source moved since generation, not whether the page was
edited by hand. A green gate would have told me nothing and the fix would
have vanished at the next regeneration.
Gate re-run against a clone checked out at the new pin: 9 generated pages
match their sources. 670 tests.
Co-Authored-By: Claude Opus 5 noreply@anthropic.com
assigned to @paulhitt
restored source branch
docs/take-the-v3-table-updatementioned in commit
8c6c7d9fadchanged the description