3
0
mirror of https://github.com/larvit/adf-codec.git synced 2026-11-18 22:25:44 +01:00

Answer prose round 1: 0 🔴, 8 🟡 (0 misread); reword the decision, the changelog bullet and the plan sentence, keep contributor text out of the Size row

This commit is contained in:
2026-11-15 13:45:04 +01:00
committed by Lillem4n
parent f78c12cd5f
commit d3832a85b6
4 changed files with 19 additions and 16 deletions
+1 -1
View File
@@ -66,7 +66,7 @@ In `docs/decisions.md`:
- Firefox reads the build
- The coverage floors
- The size ratchet
- The published size is a bundler's output, checked within a tenth
- The build ships unminified; the README's size is a minified bundle's, checked within a tenth
- Properties on a fixed seed
- The CommonMark suite checks three ways
- The flavour spec is read as a source
+1 -1
View File
@@ -2,7 +2,6 @@
## Unreleased
- The README states the package's size: 71 kB minified and gzipped as a bundler ships it, 161 kB packed and 601 kB unpacked, and a size column in the comparison table; the release pipeline fails where a figure is a tenth off.
- `adfToHtml` writes a media single, a media group and a caption as `<figure class="adf-media-single">` holding an
`<img>` and a `<figcaption>`, and as `<div class="adf-media-group">` of `<span class="adf-file">` files, where
`0.3.1` wrote the carry element, and `htmlToAdf` reads them back. A link around the image is the media's link
@@ -55,6 +54,7 @@
returns `not-an-adf-document`, where `0.3.1` ran the getter, looped or threw. The copy keeps own
enumerable keys alone, so an inherited or non-enumerable key drops. A document of more than
10,000,000 JSON values returns `unsupported-node-shape`.
- The README states the package's minified and gzipped size, its tarball and its unpacked size, and gives each compared converter's minified and gzipped size in a column of the comparison table. The release fails where the package's stated size is more than a tenth off the build's.
## 0.3.1
+13 -11
View File
@@ -924,20 +924,22 @@ config gone missing fails the leg instead of falling back to oxlint's own defaul
`--deny-warnings`, since a rule from a category this config never names arrives as a warning it
exits 0 on.
## The published size is a bundler's output, checked within a tenth
## The build ships unminified; the README's size is a minified bundle's, checked within a tenth
2026-11-15, the implementer's recommendation, open to the maintainer. Goal 3. Valid while competitors advertise a minified and gzipped figure.
2026-11-15, the implementer's recommendation, open to the maintainer. Goal 3, and goal 8, which asks a
consumer to configure nothing. Valid while competitors advertise a minified and gzipped figure.
The README states the package's size as what a bundler ships, minified and gzipped, beside the tarball
and the unpacked size, and in a column of the comparison table. That is the figure the other packages
advertise (marklassian's "12kb" measures 13 kB here). `bundle-size/` measures it with esbuild, in its own
package as §Stryker in its own package has it: the bundle of `export * from` the entry, dependencies
included, for Node. The comparison measures every package with the same call.
The README states the package's size as a bundler produces it: the whole package and its dependencies
bundled, minified and gzipped. Competitors advertise that figure (marklassian's "12kb" measures 13 kB
here), so it is the one that compares. `bundle-size/` measures it with esbuild. That package is its own, as
§Stryker in its own package has it, and the comparison table's `converter-comparison/` measures each
package with the same function. The README also states the tarball and the unpacked size.
The build stays tsc's unminified output, since goal 8 asks nothing of a consumer, and a minified build
leaves their debugger and stack traces nothing to read. The gate fails where a README figure is more than
a tenth off the measure: a tighter bound fails most pull requests, and a tenth still catches a feature
that doubles a part of the package. The release job needs the gate, so a stale figure stops a release.
The build stays tsc's unminified output: a consumer has to add no step to read a stack trace or step
through the package in a debugger, and the build takes no minifier. The gate fails where a stated figure
for this package is more than a tenth off the measure. A tighter bound fails most pull requests, while a
tenth still fails a release in which the package has grown by a tenth since its figure was last stated.
The release job needs the gate, so a stale figure stops a release.
## Properties on a fixed seed
+4 -3
View File
@@ -117,12 +117,13 @@ Consumers bundle the package, Forge apps with webpack among them, and one report
dependency" warning from another converter. Nothing in the gate bundles it today. Install the
tarball as `package-tests` does, bundle an entry file that round-trips one document through
`adfToLosslessMarkdown` and `losslessMarkdownToAdf`, and run each bundle in Node. A warning fails the
leg as an error does: webpack's "Critical dependency" is a warning. The bundlers live in their own
package, as Stryker does (§Stryker in its own package), with pinned versions. Bun builds from the
leg as an error does: webpack's "Critical dependency" is a warning. The bundlers live in
`bundle-size/`'s package, beside esbuild, as Stryker lives in its own (§Stryker in its own package), with
pinned versions. Bun builds from the
image the gate already pulls. No leg needs an account or a network beyond the install.
Once the leg holds, the README may say the package bundles with those four; it says nothing about
Forge's runtime, which no leg tests. The bundlers join esbuild in `bundle-size/`'s package.
Forge's runtime, which no leg tests.
### 42. Trim a text leaf's trailing blanks in linear time.