Design 4: fast-check generators on a fixed seed, in four sub-items
CI / gate (push) Successful in 17s
CI / publish (push) Has been cancelled

This commit was merged in pull request #64.
This commit is contained in:
2026-09-14 15:02:15 +02:00
parent c9fb85f9db
commit f8a5d5e9b3
2 changed files with 28 additions and 19 deletions
+1 -2
View File
@@ -223,8 +223,7 @@ functions, and a branch floor that only ever moves upward. It sits below 100 bec
compared against `undefined` — have a half no valid document reaches. compared against `undefined` — have a half no valid document reaches.
The corpus, all checked in: hand-built fixtures per node and combination; real sanitized ADF from The corpus, all checked in: hand-built fixtures per node and combination; real sanitized ADF from
live Atlassian APIs; property-generated ADF trees; the CommonMark spec suite against live Atlassian APIs; the CommonMark spec suite against `markdownToAdf` and `markdownToHtml`.
`markdownToAdf` and `markdownToHtml`.
`spec/flavour.md` is read as a source too, so the node tables cannot drift from the prose they `spec/flavour.md` is read as a source too, so the node tables cannot drift from the prose they
copy: each `- ` bullet in `## Block nodes`, `## Inline nodes` and `## Marks` declares the nodes copy: each `- ` bullet in `## Block nodes`, `## Inline nodes` and `## Marks` declares the nodes
+27 -17
View File
@@ -44,23 +44,33 @@ proves 12, 13 spells 11's gaps in 12's grammar, and 12 rewrites code 4b and 4c c
- [x] **3j — The carry and the combinations.** - [x] **3j — The carry and the combinations.**
- [x] **3k — The CommonMark spec suite.** - [x] **3k — The CommonMark spec suite.**
- [ ] **4 — Round-trip property tests (`0.2.0`)**, widening 3j's corpus round-trip past the - [ ] **4 — Round-trip property tests (`0.2.0`)**, widening 3j's corpus round-trip past the
documents a human wrote — the thing that proves 2 and 3 beyond them. Editor-normal (§2) is documents a human wrote — the thing that proves 2 and 3 beyond them.
finished here, on 3i's merging — `toEditorNormal(doc)` and the equality the round-trip **Settled** (the maintainer, 2026-09-13): `fast-check` generates and shrinks. The gate runs a
asserts, which over normalized input is the canonical serializer's compact spelling — fixed seed, the properties together adding about five seconds per engine; an environment
rather than staying spelled inline as `?? []` at every reader. The variable raises the runs and randomizes the seed for local digging, and a counterexample
reading half is `nodeContent`/`nodeAttrs`/`nodeMarks` over the ~28 sites spelling it found becomes a round-trip fixture. The generators draw from the node tables — each node's
inline today, which also lifts the branch floor §10 keeps below 100 for exactly those content model and attribute vocabulary as `adf/` records them, which 11 holds to Atlassian's
halves. schema — and misplace a share of nodes so the carry (§3) is exercised; no JSON Schema walker
Generators emit editor-normal ADF (§2). Real sanitized ADF from live Atlassian APIs lands enters the tests. `toEditorNormal` stays internal. 2e5's collision test goes, since a
here too (§10), in `corpus/real-payloads/`: an ADF→markdown→ADF check with no expected collision already fails the round-trip on the same fixtures; the fixture-duplicate test
markdown, the payloads supplied by the maintainer. This subsumes 2e5's collision property — stays.
a document that round-trips proves no other document shares its spelling — so decide here - [ ] **4.1 — Editor-normal and the node accessors.** `toEditorNormal(doc)` in
whether that gate stays as the parser-free, faster-failing signal or goes; the half holding `src/adf/editor-normal.ts`, on 3i's merging: adjacent text nodes carrying identical marks
no fixture duplicates is hygiene rather than a round-trip claim, and stays either way. merged, an empty `attrs`, `marks` or `content` the absent key (§2), and the round-trip
**Settled** (the maintainer, 2026-09-13): the generators draw from the node tables — each tests compare through it. `nodeContent`/`nodeAttrs`/`nodeMarks` replace the 49 inline
node's content model and attribute vocabulary as `adf/` records them, which 11 holds to `?? []`/`?? {}` reads in `src/` (27 `content`, 12 `marks`, 10 `attrs`), and the branch floor
Atlassian's schema — and misplace a share of nodes so the carry (§3) is exercised; no JSON rises to what the suite then measures.
Schema walker enters the tests. - [ ] **4.2 — The ADF property.** `fast-check` joins `devDependencies`, AGENTS.md §5 naming what
it earns — shrinking a failing document to the nodes that break it — and §10 the properties
beside the corpus. A generated editor-normal document either refuses in `adfToMarkdown`
with a `ConvertError` or reads back through `markdownToAdf` to an equal document, and
nothing throws, under Node, Deno and Bun alike. 2e5's collision test is deleted.
- [ ] **4.3 — The markdown property.** Generated markdown through `markdownToAdf` never throws,
and the runs fit the budget; where it parses and `adfToMarkdown` spells the result, that
spelling parses and emits to itself byte for byte (§2).
- [ ] **4.4 — The real payloads.** `corpus/real-payloads/` holds the maintainer's sanitized
payloads, each round-tripped ADF→markdown→ADF with no expected markdown. It waits on the
maintainer placing the files.
- [ ] **4b — The block walk's retry (`0.2.0`).** `emitBlock` walks a subtree twice wherever - [ ] **4b — The block walk's retry (`0.2.0`).** `emitBlock` walks a subtree twice wherever
`readableBlock` reads it whole and then gives up — a list item whose first line reads back `readableBlock` reads it whole and then gives up — a list item whose first line reads back
as a thematic break — and the walk below does the same, so the cost doubles per level: as a thematic break — and the walk below does the same, so the cost doubles per level: