docs: move the connect onboarding design to agmsg-cloud - #776
Open
fujibee wants to merge 2 commits into
Open
Conversation
…els that point at nothing koit's ruling: the OSS repo should not carry the hosted service's content, and duplication is fine if it makes each side self-contained. ## Moved remote-connect-onboarding.md (55KB, names the service 20 times) is in JugemuAI/agmsg-cloud now, under docs/ref/. Its six relative links were rewritten as absolute URLs there, since they pointed at files in this repo. adr/ref/0007 listed it under "Normative design and specifications" while it carried a SUPERSEDED banner. That entry is gone; the section still names docs/design/remote-sync.md, which is what is normative. ## The labels, and why they are gone rather than moved The instruction was to write the content of that document's section 8 into device-pairing.md, so the latter stops depending on a document that is leaving. device-pairing.md said: Replaces the NOT READY subsection of remote-connect-onboarding.md section 8, which recorded five findings from an adversarial review. A2 (one-directional authentication) was closed by ... Revision 1 addressed A1, B, E1 and E2 ... There is nothing to move. Measured before editing: "five findings" in that document 0 definitions of A1 / A2 / E1 / E2 there 0 (4 hits, all E2EE or E1-E3) definitions anywhere under docs/ 0 So those labels are named in exactly one place -- the sentence citing them -- and the thing they cite does not record them. Writing them out would have meant inventing five findings and attributing them to a review, which is the one thing this cleanup must not do. What had content is kept as content: one-directional authentication is still described, without its label. What had no content anywhere is gone. The only loss is the knowledge that labels by those names existed, and they had already stopped pointing at anything. ## Still here device-pairing.md stays. Nine code sites name it, and it is now readable without opening a document in another repository -- which was the stated stop condition.
… has happened
Both raised in review.
## A spelling my derivation did not cover
`scripts/key.sh:56` said "(remote-connect onboarding design)" -- a
reference to the document being removed, written with a hyphen inside the
first pair and a space before the last word. The sweep used three
spellings (basename with .md, bare basename, suffix-stripped) and none of
them match that.
Searched by WORD SEQUENCE now rather than by spelling:
remote[-_ ]?connect[-_ ]?onboarding (case-insensitive)
2 hits, both `feat/remote-connect-onboarding` branch names in
comments -- not references to the document
Third time in this cleanup that one form of a reference was handled and
another was not looked at. A spelling list is a guess about how people
wrote it; the word sequence is the thing itself.
## The claim that depends on landing order
`docs/design/remote-sync.md` said the design "has moved to the
agmsg-cloud repository". This PR removes the file; the receiving PR is
separate and not landed. If this one lands first, the tree asserts a move
that has not happened, and nothing in it can be checked.
It now says the document does not belong in this repository, which is
true of the state this PR creates and does not depend on when the other
one lands.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Declared reviewers: 1
What this PR is, on its own
It removes
docs/design/remote-connect-onboarding.mdfrom this repository and repairs everything that pointed at it. The reason it is removed does not depend on any other PR: the document is about the hosted service, and the ruling for this cleanup is that the OSS repo does not carry the hosted service's content.A companion PR exists in the private
agmsg-cloudrepo. It is open and unmerged, a public reader cannot follow it, and nothing here is true or false because of it. It is named for the record only, and no sentence in this PR or in the tree it produces says the document "is now" anywhere.An earlier revision of this body and of
docs/design/remote-sync.mdsaid the document "has moved to the agmsg-cloud repository". Both were wrong in the same way — asserting a state that would only exist if something else landed first — and both are gone.What replaces it
docs/design/remote-sync.mdis what is normative for this subject, and it existed before this PR.adr/ref/0007listed the removed document under "Normative design and specifications" while it carried a SUPERSEDED banner. That entry is gone; the section still namesremote-sync.md.device-pairing.mdstays. Nine code sites name it, and it now reads without opening a document in another repository — the stated stop condition for this move.The five labels are dropped, and this is the part to read
The task was to write section 8's content into
device-pairing.md, so that document stops depending on one that is leaving. It said:There was nothing to move. Measured before editing anything:
remote-connect-onboarding.mddocs/A1/A2/E1/E2E1-E3)The labels are named in exactly one place — the sentence that cites them — and the document they cite does not record them. Writing them out would have meant inventing five findings and attributing them to a review, which is the single thing a cleanup like this must not do.
So: what had content is kept as content — one-directional authentication is still described, without its label. What had no content anywhere is gone. The only thing lost is the knowledge that labels by those names existed, and they had already stopped pointing at anything.
The residual-reference measurement, redone
The first sweep searched three spellings — basename with
.md, bare basename, suffix stripped — and scoped todocs/. It missedscripts/key.sh:56, which wrote the name asremote-connect onboarding: a hyphen inside the first pair, a space before the last word. A spelling list is a guess about how people wrote it.Redone over the whole public tree, by word sequence rather than spelling:
Zero references to the document.
key.sh:56now points atdocs/design/remote-sync.md, which exists in this tree.No mutation
This PR changes one comment line and prose. It does not change a line that executes, so there is no mutation to run and no red to report —
bash -n scripts/key.shis the whole of what was checkable. Recorded because "no mutation given" and "no mutation definable" read the same in a review queue.exact head:
6b0ebec507230840b4ae82ea90d24e9e533fa339