Skip to content

Sync public tree to dev (issue 593) - #6

Closed
MathysRennela wants to merge 2 commits into
mainfrom
sync-dev-2026-11-19
Closed

MathysRennela wants to merge 2 commits into
mainfrom
sync-dev-2026-11-19

Conversation

@MathysRennela

Copy link
Copy Markdown
Collaborator

Syncs spec/, machine/, lean/, scripts/, and sanitized root files (including new public design notes) to current dev. Validation: lint, pytest, fixture conformance, priced harness, lake build green.

@vprusso vprusso left a comment •

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Requesting changes. Reviewed at 84da4d2. The content of the sync is consistent: CI passes, no Markdown link in the tree is dead, and the spec/, machine/, and lean/ files match the development tree. The problems are at the edges of the sync.

The sync reverts a merged dependency update

.github/workflows/ci.yml and release.yml go from actions/setup-python v7.1.0 back to v5.6.0. That undoes #4, which is merged on main. Please keep the v7.1.0 pin, and carry it into the source tree so that the next sync does not revert it again.

References to files that are not in this repository

The tree cites 34 distinct files under design/, archive/, notes/, and diary/, about 130 mentions in all, up from about 80 on main. None of those directories ships. Three groups need different handling.

  1. Files that this PR ships at the root and that are still cited at their old path. design/ROADMAP.md, design/SCOPE.md, design/PRIOR_ART.md, and design/GATE_REALIZATION_ROWS.md are cited that way in spec/ and machine/, 17 times in all. spec/JUDGMENT_LAYER.md was updated in three places and the others were not. These are a mechanical rewrite.
  2. ROADMAP.md cites DIARY.md, ../DIARY.md, and ../TODO.md (lines 86, 247, 395, 426, 452, 474, 505, 546, 550, and 646). The ../ prefix is left over from the file's previous location, and neither file ships. Line 426 makes "a dated maintainer decision in ../DIARY.md" the promotion criterion, so a public reader cannot check whether a criterion is met. LEADERBOARD_TAXONOMY.md cites two files under notes/ the same way.
  3. Derivations cited as evidence. spec/JUDGMENT_LAYER.md §2f records four hole-punch facts with the status "derivation in the repository" or "argument in the repository" and gives archive/notes/D2_*.md as the source. In this repository those files do not exist, so the status is false as shipped. The same applies to design/DERIVED_RELATIONS.md (32 mentions, cited as the derivation of the tower price provider and other relations), design/ADMISSIBLE_SEAM_FAMILY.md, design/MAPPING_CONE_CONSTRUCTOR.md, and design/LEFT_IN_PAPERS.md. For each one, either ship the note or reword the citation and the status so that they say the derivation is held outside this repository. Five diary fragments are cited by file name in spec/ as the record of a decision, and the same choice applies to them.

Part of this predates the PR. The PR is the natural place to fix it, since it adds 50 mentions and ships six documents whose purpose is to make the cited context public.

Two scripts that do not belong in the public tree

  • scripts/check_public_correspondence.py requires a checkout of the private development repository and names it.
  • scripts/split_diary.py splits a DIARY.md that this repository does not have.

Neither can run here. Please leave them out.

Release metadata

spec/VERSIONING.md now has a stdCSS-3 / v0.3.0 release row and calls v0.3.0 the current development release. CHANGELOG.md has an empty [Unreleased] section, CITATION.cff says 0.2.0, and the latest tag is v0.2.0. This PR adds four Lean modules, six root documents, and several machine modules. Please add an [Unreleased] entry that lists them, so that the changelog and the release row agree on what is pending.

README.md does not mention the six new root documents. A reader has no path to ROADMAP.md, SCOPE.md, PRIOR_ART.md, LEADERBOARD_TAXONOMY.md, GATE_REALIZATION_ROWS.md, or REDTEAM_FINDINGS.md except through three links in spec/JUDGMENT_LAYER.md.

Smaller points

  • scripts/lint_docs.sh loses its executable bit (100755 to 100644).
  • Issue numbers in the tree go from about 700 mentions to about 1100. They refer to a tracker that public readers cannot open, and a bare #NNN in this repository resolves to an unrelated issue or pull request here. This is existing practice. A stated convention in README.md or CONTRIBUTING.md (for example, that issue numbers in the text refer to the development tracker) would be enough.

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

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants