Skip to content

docs: bring the repo's account of itself in line with what is true - #9

Merged
hyperpolymath merged 1 commit into
mainfrom
docs/state-of-play
Jul 29, 2026
Merged

docs: bring the repo's account of itself in line with what is true#9
hyperpolymath merged 1 commit into
mainfrom
docs/state-of-play

Conversation

@hyperpolymath

Copy link
Copy Markdown
Owner

Documentation pass for humans and machines. Everything below was measured or
path-checked; nothing is asserted. No code touched — the diff is 5 files, all docs and
manifests.

The machine-readable side was the worse half

.machine_readable/6a2/STATE.a2ml was stale and partly corrupt:

  • named the project "gql-dt", dated 2026-02-01, claimed 45% completion
  • listed already-completed milestones as critical next actions — "Create
    bridge/lith_types.zig", "Begin Milestone 6 (GQL Parser)"
  • its purpose field held a fragment of a build error rather than a purpose:
    "References to removed .tracked and .confidence constructors, missing API updates"

Rewritten with measured figures, plus four sections it didn't have: [verification]
(each gate and its result), [caveat] (a green proof gate ≠ nothing assumed),
[history-landmines] (the squash-extract, the sweep branches, the floating @main pins),
and [evidence] (the command reproducing every claim).

0-AI-MANIFEST.a2ml identified the repo as "GQL-DT — Dependently-Typed Glyph Query
Language"
. It is GNPL; GQLdt is the layer beneath. The purpose now states the two-layer
structure explicitly, so an agent doesn't conclude the GqlDt namespace is a mistake.

Its canonical-locations block was worse — all six paths were dead:

Claimed Reality
.machine_readable/{STATE,META,ECOSYSTEM,ROADMAP}.scm converted to .a2ml under 6a2/, never repointed
AI.a2ml never existed
justfile never existed

Replaced with 20 paths, each verified to exist (20/20 checked), plus an (evidence)
block and a (caveat-for-agents) warning against reporting this repo as "verified" on the
strength of a green build.

Human-readable

docs/STATE-OF-PLAY.adoc (new) — the human twin of STATE.a2ml, kept deliberately
parallel so drift between them is visible. What is established and by which command;
the gates and why they're trusted; what is known missing; the two open decisions; the
sequencing.

It also lists the four gates this repo has shipped and removed that could not fail — the
tautological naming check, the || echo on lake test, suites whose main : IO Unit
always exited 0 while printing "All tests passed!", and the Containerfile masking both
its steps. That history is why the standard is "a gate is not trusted until it has gone
red on a seeded fault".

docs/proof-debt.md §(b) claimed this repo has "no executable test coverage" and
that a @[test_driver] was "a prerequisite for moving any item into this section". That
prerequisite is now met (163 checks) — so the claim was stale in my own favour, which is
the direction that matters least and gets caught least.

Corrected while keeping the section empty: coverage existing is not the same as a given
axiom being covered, and the policy wants a citable refutation budget. D4's entry updated
likewise — its blocker is no longer "no coverage at all" but "this path is not
property-tested", which is materially different and now reachable.

Verification

  • 20/20 manifest paths confirmed present
  • diff touches no .lean, .zig or lakefilegit diff --name-only shown above
  • main at 9de0713 independently verified green from a clean checkout earlier today:
    lake build exit 0, lake test exit 0, 163 checks

🤖 Generated with Claude Code

Documentation pass for humans and for machines. Everything here was measured or
path-checked; nothing is asserted.

## Machine-readable

`.machine_readable/6a2/STATE.a2ml` was stale and partly corrupt: it named the
project "gql-dt", was dated 2026-02-01, claimed 45% completion, listed
already-completed milestones ("Create bridge/lith_types.zig", "Begin Milestone
6") as CRITICAL NEXT ACTIONS, and its `purpose` field held a fragment of a build
error — "References to removed .tracked and .confidence constructors" — rather
than a purpose.

Rewritten with measured figures, plus sections that did not exist:
[verification] naming each gate and its result, [caveat] stating that a green
proof gate does not mean nothing is assumed, [history-landmines] recording the
squash-extract and the sweep branches, and [evidence] giving the command that
reproduces every claim.

`0-AI-MANIFEST.a2ml` identified the repo as "GQL-DT / Dependently-Typed Glyph
Query Language". It is GNPL; GQLdt is the layer beneath. Corrected, and the
purpose now states the two-layer structure explicitly so an agent reading it
does not conclude the `GqlDt` namespace is a mistake.

Its `canonical-locations` block was worse: **all six paths were dead**. The
.scm state/meta/ecosystem/roadmap files were converted to .a2ml under
`.machine_readable/6a2/` and never repointed, and neither `AI.a2ml` nor
`justfile` has ever existed here. Replaced with 20 paths, each verified to exist,
plus an (evidence) block and a (caveat-for-agents) warning against reporting this
repository as "verified" on the strength of a green build.

## Human-readable

`docs/STATE-OF-PLAY.adoc` (new) — the human twin of STATE.a2ml. What is
established and *by which command*, the gates and why they are trusted (each was
canary-tested), what is known missing, the two open decisions, and the
sequencing. It also lists the four gates this repo has shipped and removed that
could not fail, because that history is the reason the standard is what it is.

`docs/proof-debt.md` §(b) said this repo has "no executable test coverage" and
that adding a `@[test_driver]` was "a prerequisite for moving any item into this
section". That prerequisite is now met — 163 checks. Corrected, while keeping
the section empty: coverage existing is NOT the same as a given axiom being
covered, and the policy wants a citable refutation budget. D4's entry updated
likewise — its blocker is no longer "no coverage at all" but "this path is not
property-tested", which is a materially different and now-reachable problem.

`README.adoc` links both new documents from the design-documents table.

No code changes; `lake test` still 163/163.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
@gitar-bot

gitar-bot Bot commented Jul 29, 2026

Copy link
Copy Markdown

Note

Automatic reviews are paused because your trial's included automatic processing has been used for this period. Upgrade now, or comment "Gitar review" to run a review anytime.
Learn more

Code Review ✅ Approved

Updates documentation and machine-readable manifests to correct stale status claims, fix dead file paths, and align the repository description with actual state. No issues found.

Auto-approved and auto-merge armed: No blocking issues found.
Please see Auto-approve Docs for details on setting custom approval criteria. — merges when pipeline and required approvals pass.

Options

Display: compact → Showing less information.

Comment with these commands to change the behavior for this request:

Compact
gitar display:verbose         

Was this helpful? React with 👍 / 👎 | Gitar

@gitar-bot

gitar-bot Bot commented Jul 29, 2026

Copy link
Copy Markdown

⚠️ Gitar auto-approved this PR but could not enable auto-merge: auto-merge is disabled for this repository — enable "Allow auto-merge" in the repository settings.

@gitar-bot gitar-bot Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

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

Gitar has auto-approved this PR and enabled auto-merge (configure)

@gitar-bot gitar-bot Bot added the gitar-approved Added by Gitar label Jul 29, 2026
@hyperpolymath
hyperpolymath merged commit e199288 into main Jul 29, 2026
17 of 19 checks passed
@hyperpolymath
hyperpolymath deleted the docs/state-of-play branch July 29, 2026 07:33
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

gitar-approved Added by Gitar

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant