Skip to content

[finding][spec] Nine flat "rewrite existing sources automatically" sites remain across three published CHANGELOGs — outside #9734's line-specific authorization #11636

Closed as not planned

Description

@os-warren

Filed unassigned, observation class, while implementing #9734. Deliberately not edited there: the 2026-08-22 maintainer ruling on #9734 authorized exactly one line and closed with "no blanket sweep".

Why this is recorded rather than fixed

#9734 corrected one shipped false claim in packages/spec/CHANGELOG.md under an explicit, card- and line-specific maintainer authorization to edit a published release artifact. Editing a published CHANGELOG is otherwise not something a dispatched agent does on its own judgment — that is precisely why #9734 sat in needs-user-decision for a day before the ruling.

The census below is the rest of the same defect class. It needs the same kind of decision, not the same PR.

Precedent worth reading first: #11100 recorded an equivalent CHANGELOG-history defect explicitly "so nobody 'fixes' append-only history". "Leave it, recorded" is a fully respectable outcome here.

Census (measured on origin/main @ 353f4e8f, after #9734's one-line fix)

⚠️ The hits are three different things and a search-and-replace would damage two of them. This is the trap the #9734 thread hit twice.

Defect class — flat prescriptions stating the false claim (9 sites)

File Lines
packages/spec/CHANGELOG.md 5306, 8539, 51610, 51666, 52914
packages/platform-objects/CHANGELOG.md 2564, 5589
examples/app-todo/CHANGELOG.md 254, 2384

Each reads, flat and in the present tense: Run `os migrate meta --from 16` to rewrite existing sources automatically. The command has never rewritten an authored source file (#9529); there is no --write / --fix flag. os migrate meta --write is filed for v18 as #9591 and is pm:on-hold — so the claim is not merely stale wording, it describes a capability that does not exist yet.

Note packages/spec/CHANGELOG.md:51610 is the 17.0.0-rc.2 copy of the very entry #9734 corrected (ad047d2, the app-area fail-open gates retirement). The GA copy and the rc copy diverged long ago — the rc copy already carries its own > **Correction (#5809):** blockquote — so after #9734 the two copies now disagree about the migrate sentence as well.

⛔ NOT defects — quoted in order to retract (do not touch)

⛔ NOT defects — descriptive, quoting the sentence as evidence

The options, if a decision is wanted

Any of B/C needs the same explicit maintainer authorization #9734 got. ⛔ None of them is a regex sweep — the "not defects" lists above must be excluded by hand, and site-by-site the correct replacement is the house sentence Run `os migrate meta --from 16` to list the mechanical edits for existing sources; apply them by hand.


Generated by Claude Code

Activity

  1. claude commented on Aug 25, 2026

    @claude
    Contributor

    Concentrated triage round: closing not planned — this inherits an existing maintainer ruling rather than needing a new one. The 2026-08-22 ruling on #9734 authorized exactly one line and closed with "no blanket sweep" (quoted in this card's own body); these nine sites are the blanket sweep that ruling declined. #11100 is the standing precedent for the outcome: recorded so nobody "fixes" append-only history. The census stays here as the record; reopen only if the maintainer issues a #9734-style line-specific authorization.


    Generated by Claude Code

  2. added theissue type on Aug 25, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions