Skip to content

chore(corpus): advance the vendored corpus pin to the corpus that carries the wave-3 guidance, and regenerate the site #179

Closed

Description

@aitrawog

The gap

This repo vendors the corpus as the marketplace submodule and regenerates the site from it. The pinned corpus is 1c156511, and the page it publishes at src/content/docs/recipes/internals/git-workflow/umbrella-mechanics.md still teaches the old cost of a pin wave:

A PIN WAVE needs the FULL SET: bump fails loud on a path it could not advance, so CHARLY_INIT_MODE=all charly task init comes FIRST

That is no longer what the product does. The wave-3 producer change landed upstream — opencharly/plugin-task#41 (squash 1d8342fc, root tag v2026.296.1911, module tag candy/plugin-task/v0.2026296.1911) — and the bump now advances a path with no checkout IN THE INDEX (a leaf to the tip its remote serves for its branch=, an edge to the twin the source's tree records), while a path WITH a checkout is still rolled in that checkout, so the two keep agreeing. It still refuses, by name, every path it cannot advance.

The consumer leg inside the corpus is opencharly/marketplace#565 (the refs list advanced to layer-charly-internals:v2026.296.1907, the tag that carries the wave-3 guidance, and the corpus regenerated). Once that lands, the corpus publishes the new rule, and this repo must take the consumer leg of its own: advance the vendored pin and regenerate the site, so the published page teaches what the product does — a wave needs no full set, and a wave's order is sync -> init -> verify.

The fix

  • advance the marketplace submodule pin to the corpus's merged main;
  • regenerate the site, so src/content/docs/recipes/internals/git-workflow/umbrella-mechanics.md carries the new wording;
  • prove the regeneration is a no-op on the committed tree (the repo's own drift gate).

Coordination

The marketplace path in this repo is also touched by the open PR #178 (the site regeneration at marketplace 76b8c06). This leg lands after #178, and never over it.

Related

Activity

  1. aitrawog commented on Oct 23, 2026

    @aitrawog
    ContributorAuthor

    CLAIM

    I am taking this leg: the vendored marketplace pin advance and the site regeneration that carries it.

    What is done. The producer is merged and tagged (opencharly/plugin-task#41, root tag v2026.296.1911, module tag candy/plugin-task/v0.2026296.1911), and the consumer leg inside the corpus is opencharly/marketplace#565 — the refs list advanced to layer-charly-internals:v2026.296.1907 and the corpus regenerated, so the published reference teaches that a wave needs no full set and that a wave's order is sync -> init -> verify.

    What is next, and in what order. This leg lands after the open PR #178, which touches the same marketplace path in this repo (the site regeneration at marketplace 76b8c06). The sequence is: #565 merges and is tagged → the corpus's merged main is known → this branch advances the pin to it → the site is regenerated → the repo's own drift gate proves the regeneration is a no-op on the committed tree.

    Nothing is pushed over #178. The same-path overlap is why this issue waits instead of branching now.

    Agent: charly · session light-init
    Assisted-by: Claude Code minimax/minimax-m3.1-flash:cloud (fully tested and validated)

  2. self-assigned this
    on Oct 23, 2026
  3. aitrawog commented on Oct 23, 2026

    @aitrawog
    ContributorAuthor

    CLEARED

    The block on this leg is gone: #178 merged (f3bd7cc, tag-on-merge
    1ab65a7), so the marketplace path in this repo is free and this leg branched from
    the merged main — not over it.

    What is now true. The branch feat/corpus-wave3-docs is one commit ahead of
    origin/main and zero behind; the corpus pin advances 76b8c06 → af08be9
    (the corpus's merged main, opencharly/marketplace#565), and the site is regenerated
    from it. The PR is #180.

    What would unblock it. Nothing further on this leg: #178 landing was the only
    condition, and it is met.

    Agent: charly · session light-init
    Assisted-by: Claude Code minimax/minimax-m3.1-flash:cloud (fully tested and validated)

  4. aitrawog commented on Oct 23, 2026

    @aitrawog
    ContributorAuthor

    STATUS — the leg is open as #180.

    • Head ceb6bf1b8bb001344b15ac090e5a71f5b4c76ebf, tree 50cff75ffae8bcd32fc89f448b43247a123c0391
    • Branch feat/corpus-wave3-docs, 1 commit ahead of origin/main, 0 behind
    • Diff vs origin/main: marketplace gitlink 76b8c06 → af08be9, plus the one
      regenerated page — src/content/docs/recipes/internals/git-workflow/umbrella-mechanics.md
      (+45/−29); 2 files changed, 46 insertions(+), 30 deletions(-)

    Done. The corpus pin advanced to the corpus's merged main (opencharly/marketplace#565),
    the site was regenerated with the deploy workflow's exact inputs (charly at
    bafd8dbc, umbrella at e631aacf, plugin-docs at v2026.289.0814), a second
    regeneration on the committed tree is a no-op (the repo's own drift gate: "committed
    tree matches the pinned inputs (regeneration is a no-op)"
    ), and the Astro build produced
    1699 pages. The retired claim — "a pin wave needs the full set" — is gone from the
    published page, and the new one — "a pin wave needs NO full set", order sync → init →
    verify
    — is in its place; the claim-keyed grep is pasted in the PR body.

    Next. Read the validator's verdict on #180 in full, fix or rebut every finding there
    before any next push, and land it. No other repo's leg is waiting on this one.

    Next update due: when the validator reports on #180, or when the next round of
    findings lands.

    Agent: charly · session light-init
    Assisted-by: Claude Code minimax/minimax-m3.1-flash:cloud (fully tested and validated)

  5. aitrawog commented on Oct 23, 2026

    @aitrawog
    ContributorAuthor

    RESOLVED — the published site now carries the wave-3 umbrella mechanics, and the leg is landed and tagged.

    What landed. #180 — merge commit f4984c10d4ab5c60764fa43ec052af35042d160d, merge-time tag v2026.296.2037 (the docs repo's main is now 5e267990, the tag-on-merge changelog commit).

    The page now says what the product does. A pin wave needs no full set (the bump advances a path with no checkout in the index), the wave's order is sync → init → verify, and a fresh session worktree runs sync to a green exit with its pins living in the index until the tree is initialized. The retired claim is gone from the published page, and the claim-keyed before/after grep is in the PR body.

    Gated. A second regeneration on the committed tree is a no-op ("content: committed tree matches the pinned inputs (regeneration is a no-op)", exit 0), the site builds (1699 page(s) built in 51.41s, exit 0), and the validator's verdict on the final body is a PASS — read in full, its one non-blocking observation (the #37 vs #41 attribution) was fixed in the body, not filed away, and the validator was re-run once over the corrected body.

    The sibling legs. The umbrella's own wave-order leg landed as opencharly/opencharly#641 (merge 6eddd829, tag v2026.296.2037) on the same wave, and its tracker opencharly/opencharly#588 is closed with its own RESOLVED record. The one finding this wave raised — the recurring stale box/ubuntu twin — is filed as opencharly/charly#1048 and is CLAIMED; it is the umbrella's distro-ubuntu edge's producer leg, and it is not closed by this issue.

    Agent: charly

    Assisted-by: Claude Code minimax/minimax-m3.1-flash:cloud (fully tested and validated)

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

Labels

No labels
No labels

Type

No type

Projects

No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions