Add render-diff and render-fixture subpath exports + MDX↔MDXish regression suite by jamestclark · Pull Request #1489 · readmeio/markdown

jamestclark · 2026-05-26T22:21:48Z

🎯 What does this PR do?

Adds two new subpath exports plus a 13-fixture regression suite that gates any future rendering change in the @readme/markdown engine.

New public surface

// Engine-free HTML diff — 152 KB
import { diff } from '@readme/markdown/render-diff';
import type { Change, DiffResult, Severity, Preset } from '@readme/markdown/render-diff';

// Fixture-based render helper — 4.04 MB (includes engines)
import { renderFixture, loadFixture } from '@readme/markdown/render-fixture';
import type { RenderContext, Engine, FixtureRenderResult } from '@readme/markdown/render-fixture';

The two subpaths intentionally split along the engine boundary so a consumer that only wants to diff two HTML strings (e.g. CI gating on rendering parity across @readme/markdown version bumps) doesn't have to pull in the full 4 MB engine bundle. Pairing them gives a complete render-and-compare loop in ~10 lines of consumer code:

import { renderFixture, loadFixture } from '@readme/markdown/render-fixture';
import { diff } from '@readme/markdown/render-diff';

const { body, ctx } = loadFixture('./docs/install/');
const { html: fresh } = renderFixture(body, ctx, 'mdxish');
const persisted = readFileSync('./snapshots/install.html', 'utf8');
const result = diff(persisted, fresh, { preset: 'minimal' });

if (result.status === 'differ' && result.severity !== 'cosmetic') {
  // regression — gate CI / surface diff
}

Why the diff tool is engine-agnostic

The public API uses neutral left/right naming (diff(leftHtml, rightHtml, opts), Change.left/Change.right), not mdx/mdxish. Suite B in this PR uses it for MDX↔MDXish comparison; other consumers use the same tool for before/after a version bump or any two-input HTML comparison. The 'minimal' preset is explicitly designed for same-engine before/after diffing (skips span-flatten + adjacent-text-merge that would mask real structural changes). 'cross-engine' (the default) does the fuller normalization needed for MDX↔MDXish.

Regression suite — 13 fixtures

Two suites share the fixture corpus:

Suite A (__tests__/regression/snapshots.test.ts) renders each fixture through both engines separately and toMatchSnapshot()s the raw HTML. 26 entries total (13 fixtures × 2 engines). Byte-level changes in either engine surface as a failing snapshot on the PR that caused them.
Suite B (__tests__/regression/equivalence.test.ts) renders each fixture through both engines, runs diff() on the two outputs, and toMatchSnapshot()s the structured diff. As MDX↔MDXish parity improves, snapshots shrink toward { status: 'match' }.

Both suites auto-discover fixture directories via readdirSync — adding a new fixture takes zero code change. The fixture format (body.md + context.json + optional components/*.mdx) is documented in __tests__/regression/fixtures/README.md.

The 13 fixtures break into baselines and regression coverage:

Baselines (the two arms of DiffResult + the data-plumbing surface):

convergent — match branch (plain markdown)
divergent — differ branch (HTMLBlock safeMode)
rich — variables + glossary + custom component

Regression fixtures (each grounded in a real merged bug fix from the past few months):

tables-with-html (PR fix: code elements stripped from td elements #1466, fix: don't parse content of <code> elements in table cells #1467, fix(mdxish): preserve attributes on raw <table> rows/cells #1463, fix(mdxish): keep md images inline when in tableCells #1469, fix: dont normalize emphasis syntax within html code elements #1471)
magic-blocks-table (PR fix(mdxish): multiple consecutive magic blocks under a list item not rendering & out of order #1451, fix(mdxish): consecutive magic blocks not rendering if they have html tags in their content #1452)
unclosed-tags (PR fix(mdxish): bail orphaned <Tag> openers before they eat sibling blocks #1474, fix(mdxish): preserve raw-content HTML tags when unclosed on a line #1477, fix(mdxish): stop orphan closer tags from breaking table parsing #1480, fix(mdxish): handle orphan void closers in JSX table cells #1482)
callouts-and-glossary (PR fix: guard against empty Glossary tag causing crash #1408, fix(mdxish): fix FA Emojis in Callouts #1421, fix(mdxish): legacy glossary syntax in callout title crashing #1441, fix(Callout): only apply callout title style to headings #1454)
jsx-attribute-entities (PR fix(mdxish): render Image captions containing entity-encoded JSX #1461, fix(mdxish): decode html entities in jsx attribute values #1462)
variables-everywhere (PR fix(mdxish): handle user vars on standalone lines in tables #1423, fix(mdxish): skip variable resolution in Mermaid code blocks #1459, fix: dont normalize emphasis syntax within html code elements #1471)
embeds (PR fix: render iframe/youtube/pdf/jsfiddle embeds without embedly html #1476)
consecutive-emojis-fa (PR fix(mdxish): allow rendering consecutive emojis #1390, fix(mdxish): add gemoji support to api-header title parser #1416, fix(mdxish): fix FA Emojis in Callouts #1421, fix(mdxish): gemoji & expressions not rendering in components #1449)
compact-headings (PR fix(mdxish): Add compact heading preprocessor #1428)
htmlblock-with-script (PR fix(mdxish): scope \n unescaping in HTMLBlock content #1457 — scoped to non-runScripts path)

Each fixture has its own README documenting what it covers and what regressing it would mean.

Build verification

npm run build:verify runs both subpath check scripts in CI:
- scripts/verify-render-diff.cjs — dist-grep ensures dist/render-diff.node.js contains none of 5 engine identifiers (bundle isolation); self-ref probe confirms the subpath resolves and diff() works
- scripts/verify-render-fixture.cjs — positive control ensures dist/render-fixture.node.js contains all 5 engine identifiers (would-be-broken if missing); self-ref probe confirms both engines render via the subpath
Source-level guard __tests__/lib/render-diff/no-engine-imports.test.ts prevents any file under lib/render-diff/ from importing engine modules directly (defense-in-depth for the dist-grep)
Bundlewatch budgets: render-diff 185 KB (actual 152 KB), render-fixture 4.5 MB (actual 4.04 MB)

Commit-by-commit

Commit	What	LOC
`feat(render-diff)`	render-diff subpath + tests + verify script + CI step	+993
`feat(render-fixture)`	render-fixture subpath + tests + verify script	+456
`test(regression)`	13 fixtures + Suite A + Suite B + 39 snapshots	+1487

Each commit builds and verifies in isolation (bisectable).

🧪 QA tips

npm install
npm run build:verify    # builds dist/, runs both subpath verify scripts
npm test                # full vitest suite; new tests under __tests__/regression and __tests__/lib/render-{diff,fixture}

Specifically:

dist/render-diff.node.js is ~152 KB and contains no engine identifiers (grep mdxishAstProcessor dist/render-diff.node.js should be empty)
dist/render-fixture.node.js is ~4.04 MB and contains all engine identifiers (grep -c mdxishAstProcessor dist/render-fixture.node.js should be > 0)
node -e "console.log(require('@readme/markdown/render-diff').diff('<p>x</p>', '<p>x</p>'))" prints { status: 'match' }
node -e "const {renderFixture} = require('@readme/markdown/render-fixture'); console.log(renderFixture('# hi', { variables: { defaults: [], user: {} }, glossary: [], components: [] }, 'mdxish').html.includes('hi'))" prints true
Adding a new directory under __tests__/regression/fixtures/ is auto-discovered by both suites on the next npm test run
No regressions in existing npm test suite (full run: 1059 pass on this branch)

📸 Screenshot or Loom

N/A — no user-visible UI change. The new surface is a pair of programmatic subpath exports and a CI test suite.

Engine-agnostic HTML diff tool published as its own subpath so consumers can compare two rendered HTML strings without pulling in the engine (compile/run/mdxish/renderMdxish). Public API: import { diff } from '@readme/markdown/render-diff'; import type { Change, DiffOptions, DiffResult, Preset, Severity } from '@readme/markdown/render-diff'; diff(leftHtml, rightHtml, opts?) → { status: 'match' } | { status: 'differ', severity, changes[] } `leftHtml` and `rightHtml` are neutral param names so the same tool serves MDX↔MDXish, before/after a @readme/markdown version bump, or any other two-input comparison. Each `Change` carries `left`/`right` fields matching the inputs. Canonicalization pipeline (transplanted from the reference differ in readmeio/readme PR #18479, re-targeted to htmlparser2/domhandler): whitespace collapse, class sort, attribute normalization, noise-attr drop, heading-counter strip, void-tag handling, text-equivalent merging. Bottom-up SHA-1 content hash fast-path short-circuits to match when both trees hash equal. Two presets: 'cross-engine' (default, full normalization) and 'minimal' (skips span-flatten + adjacent-text- merge — suitable for same-engine before/after). Packaging: - webpack serverConfig multi-entry adds 'render-diff' → dist/render-diff.node.js - package.json exports map adds './render-diff' (types-first ordering) - bundlewatch budget: 185 KB (actual 152 KB) - scripts/verify-render-diff.cjs runs in CI via build:verify: (1) dist-grep — bundle MUST NOT contain any of 5 engine identifiers (2) self-ref probe — require('@readme/markdown/render-diff') resolves - .github/workflows/ci.yml adds 'Verify subpath exports' step Source-level guard (__tests__/lib/render-diff/no-engine-imports.test.ts) prevents any file under lib/render-diff/ from importing engine modules directly. Combined with the bundle-level dist-grep, this gives two layers of defense against engine leakage. Unit tests (__tests__/lib/render-diff/differ.test.ts) cover match arms, differ arms, severity scoring, preset partition, fast-path, and determinism.

Pairs with the engine-free render-diff subpath. Where render-diff refuses to load engine code (152 KB), render-fixture deliberately includes the full engine (4.04 MB) — its job is to render markdown through MDX or MDXish for fixture-based regression testing. Public API: import { renderFixture, loadFixture } from '@readme/markdown/render-fixture'; import type { RenderContext, Engine, FixtureRenderResult } from '@readme/markdown/render-fixture'; const { body, ctx } = loadFixture('./fixtures/install/'); const { html, error } = renderFixture(body, ctx, 'mdxish'); Motivating consumer: an `owl qa` style CLI that clones customer projects and detects rendering regressions across @readme/markdown version bumps. Pairing with `diff()` from the render-diff subpath gives a complete regression loop in ~10 lines of consumer code. `renderFixture` is deterministic: Date.now is frozen to 2024-01-01 and Math.random is replaced with a seeded LCG for the duration of each render, then restored in a finally block. The helper asserts sync return — if any of compile/run/mdxish/renderMdxish becomes async in the future, the determinism guarantee breaks loudly rather than silently producing flaky output. Packaging: - webpack serverConfig multi-entry adds 'render-fixture' → dist/render-fixture.node.js - package.json exports map adds './render-fixture' (types-first ordering) - bundlewatch budget: 4.5 MB (actual 4.04 MB) - verify-exports script extends to also run scripts/verify-render-fixture.cjs: (1) positive-control: bundle MUST contain all 5 engine identifiers (opposite of render-diff's negative check) (2) self-ref probe: require('@readme/markdown/render-fixture') resolves and both engines successfully render a smoke-test markdown body PKG-04 (render-diff engine-free) constraint preserved unchanged — its check still runs and still passes. render-fixture intentionally violates that constraint because it has to.

Two test suites that share a fixture corpus: - Suite A (__tests__/regression/snapshots.test.ts) renders each fixture through both engines separately and toMatchSnapshot()s the raw HTML — one snapshot entry per (fixture, engine), 26 total. Byte-level changes in either engine surface as a failing snapshot on the PR that caused them. - Suite B (__tests__/regression/equivalence.test.ts) renders each fixture through both engines, runs diff() on the two outputs, and toMatchSnapshot()s the structured diff. As MDX↔MDXish parity improves, the snapshot shrinks toward { status: 'match' }. Both suites auto-discover fixture directories via readdirSync — adding a new fixture takes zero code change. Fixture format (documented in __tests__/regression/fixtures/README.md): - body.md — markdown source - context.json — variables + glossary (engine inputs) - components/*.mdx — optional custom components by tag name Fixture corpus (13 directories): Baselines (the two arms of DiffResult + the data-plumbing surface): - convergent — match branch (plain markdown) - divergent — differ branch (HTMLBlock safeMode) - rich — variables + glossary + custom component Regression fixtures grounded in merged bug fixes: - tables-with-html — PR #1466, #1467, #1463, #1469, #1471 - magic-blocks-table — PR #1451, #1452 - unclosed-tags — PR #1474, #1477, #1480, #1482 - callouts-and-glossary — PR #1408, #1421, #1441, #1454 - jsx-attribute-entities — PR #1461, #1462 - variables-everywhere — PR #1423, #1459, #1471 - embeds — PR #1476 - consecutive-emojis-fa — PR #1390, #1416, #1421, #1449 - compact-headings — PR #1428 - htmlblock-with-script — PR #1457 (scoped to non-runScripts path) Each fixture README documents what it covers and what regressing it would mean. Three regression fixtures intentionally render empty on the MDX side (magic-blocks-table, unclosed-tags, variables-everywhere) because they exercise MDXish-flavored syntax strict MDX rejects — the MDXish snapshot is the regression contract for those. 39 snapshot entries committed: 26 in snapshots.test.ts.snap (Suite A) + 13 in equivalence.test.ts.snap (Suite B).

The Change.kind values emitted by alignChildren were the reverse of what types.ts documents: - 'missing' should mean present in left, absent in right - 'extra' should mean present in right, absent in left The differ emitted the opposite at every call site, with the helper's left/right field placement compensating so the data was self-consistent but the public discriminant was wrong. Tightens the differ.test.ts case to assert kind strictly (it previously accepted either label) and adds the sibling test for the missing case. Regenerates the equivalence suite snapshots (all 30 entries flip from missing → extra; field shapes unchanged). Also folds in the pre-staged "Suite B:" describe rename so the regenerated snapshot keys match the test names.

…ns tests" Matches the parallel rename of Suite B → "MDX↔MDXish equivalence" and refreshes the snapshot keys accordingly.

The new exports block introduced in a665b8a inadvertently narrowed the package's import surface: without exports, Node falls back to filesystem lookup so any deep import works; with exports, every path not listed throws ERR_PACKAGE_PATH_NOT_EXPORTED. The previous block only listed '.', './render-diff', and './render-fixture', which would have broken existing consumers doing imports like '@readme/markdown/styles/main.css' or '@readme/markdown/components/...'. Adds an explicit './package.json' entry (commonly read by tooling) and a './*' → './*' wildcard fallback. Specific entries still take precedence in Node's resolver, so the new subpath exports continue to hit their bundled .node.js targets.

…tution renderFixture previously passed variables only to renderMdxish(), which runs after the MDXish parse step. That meant variablesCodeResolver — the plugin in mdxish() that resolves <<...>> and {user.*} inside code and inline-code nodes — saw `variables: undefined` and fell back to uppercase placeholders (APIKEY, USER.REGION). So MDXish-side fixture snapshots locked in the unsubstituted state, even though variables-everywhere is the dedicated fixture for variable-substitution regressions (PRs #1423, #1459, #1471). Pass ctx.variables to both mdxish() and renderMdxish() so the regression suite exercises real end-to-end substitution. Re-record snapshots: - rich (mdxish) and variables-everywhere (mdxish) now resolve apiKey to sk_test_abc123 and region to us-east-1 inside code. - The equivalence snapshot for `rich` now correctly reports the MDX↔MDXish engine divergence on code-block substitution (MDX treats code as opaque; MDXish runs variablesCodeResolver). That divergence is real — captured rather than masked. - Variables without configured values (e.g. `<<name>>`) still fall back to uppercase, as the resolver intends. Update the variables-everywhere README: the "Known wiring gap" section is replaced with a "Substitution coverage" section documenting the new end-to-end coverage.

… case The structural HTML differ produced false matches (and one false positive) in its canonicalization pipeline: - Per-text-node trim() dropped significant inline-boundary whitespace, so `hello <strong>world</strong>` ("hello world") canonicalized identically to `hello<strong>world</strong>` ("helloworld"). - Adjacent-text merge inserted a space, so comment-split fragments with no surrounding whitespace (`foobar`, "foobar") falsely matched "foo bar". - xmlMode preserves attribute- AND tag-name case, so case-sensitive checks (NOISE_ATTRS, attrIgnore, class/style/id, aria-hidden, plus INLINE/VOID and heading-id classification) silently bypassed for any uppercase name. Fixes: - Collapse whitespace runs but stop trimming text nodes; drop whitespace only at block boundaries via normalizeBoundaryWhitespace. It reaches through inline wrappers (trimInlineEdge) so block-edge whitespace inside <strong>/<a>/… collapses, while genuinely significant mid-block inline spaces are preserved. - Merge adjacent text verbatim instead of fabricating a separator space. - Lowercase attribute and tag names for every classification/membership check and as the canonical key. Adds regression tests for each finding. One equivalence snapshot (consecutive-emojis-fa) is updated to drop a previously fabricated space, matching the actual rendered HTML.

eaglethrost

Nice thanks for setting this up! It feels really nice to setup these regression suites & also make them importable elsewhere.

The fixtures & render diff logic looks good to me! I might add some more specific nits later on or iterate on the assertions, but it's already a good v0.

The bundling & exports change work overall but have some questions:

Curious to know where we get the 4MB size from? Does that mean the current consumers now install 4MB of data? It seems like quite a big jump from the current bundle size
Confirming that we would not need to change any consumer code (e.g. how it imports markdown)?
Since we export the diff logic, does this open up the possibility to use it in the monorepo pipeline, and run it with fixtures of real customer projects?

eaglethrost · 2026-06-09T03:02:07Z

This is a good starting example, just noting many of the recent engine issues are table related so I anticipate we would expand this further!

Yeah for sure, the table rendering issues are probably the highest priority gap to close in this fixture data. Some PRs with recent issues we've had that we could extract some fixture data from:

Multi-paragraph table cells

Unclosed tags in JSX <Table> cells

<tr> rows without <tbody> wrapper dropped

<HTMLBlock> inside <Table> not rendering

Legacy vars + thead-only tables + MDX fallback

# Conflicts: # package.json

kevinports

Overall this is looking great!

Just had a few comments on fixture data. I think it's worth adding some coverage for some of the Table rendering issues we've had now.

kevinports · 2026-06-24T17:43:34Z

I didn't imagine we'd have a/b regression snapshots across engines like this but this will be helpful as we continue migrations!

kevinports · 2026-06-24T17:49:03Z

One other divergent example I can think of is the recent change in Callout compiling for mdxish #1498

mdx compiles to a blockquote syntax:

> 📘 Heading > > Body

mdxish now compiles to the JSX component:

<Callout theme="info" icon="📘"> ### Heading Body </Callout>

kevinports · 2026-06-24T17:54:09Z

Yeah for sure, the table rendering issues are probably the highest priority gap to close in this fixture data. Some PRs with recent issues we've had that we could extract some fixture data from:

Multi-paragraph table cells

Unclosed tags in JSX <Table> cells

<tr> rows without <tbody> wrapper dropped

<HTMLBlock> inside <Table> not rendering

Legacy vars + thead-only tables + MDX fallback

Addresses PR review requests for more table-rendering coverage and a callout example, each fixture grounded in a merged bug fix: - jsx-table-multiline-cells (#1445) — multi-paragraph cells preserved - jsx-table-unclosed-cells (#1465) — asymmetric/unclosed cell tags recovered - table-unwrapped-rows (#1458, #1411) — rows missing <tr>/<tbody> wrappers - htmlblock-in-table (#1484) — <HTMLBlock> inside a <Table> cell - legacy-vars-in-table (#1458) — legacy <<vars>> in raw table cells - callout-icons (#1498) — blockquote + FA-class-icon callout render Also refreshes the divergent, htmlblock-with-script, and jsx-attribute-entities snapshots to reflect engine output changes pulled in from the origin/next merge (invalid <p> wrappers removed around block elements; <figcaption> now a direct child of <figure>). Claude-Session: https://claude.ai/code/session_01GPTShf49qTsVP1AxSbpRJk

jamestclark · 2026-06-25T23:13:27Z

@eaglethrost good questions on the bundling — answers below:

1. Where does the 4 MB come from, and does that mean current consumers install 4 MB?

It's the engine bundle, and it isn't new weight. dist/main.node.js — what every consumer already imports via @readme/markdown — is already ~4.1 MB; the engines were always that size. The split is:

render-diff.node.js — 152 KB, engine-free
render-fixture.node.js — ~4.1 MB, re-bundles both engines (it has to — its job is to render)

Importing @readme/markdown doesn't pull in either subpath, so there's no runtime/import-size change for existing consumers. Only code that explicitly imports @readme/markdown/render-fixture (i.e. test/CI tooling) pays the 4 MB, and render-diff on its own is 152 KB. CI guards enforce the split: verify-render-diff.cjs asserts the diff bundle contains none of the engine identifiers, and verify-render-fixture.cjs asserts the fixture bundle contains all of them.

2. Confirming consumers don't need to change how they import markdown?

Correct — the changes are purely additive. The . export is unchanged, and cb57a777 ("preserve deep-import surface via exports wildcard") deliberately added "./*": "./*" + a ./package.json entry so existing deep imports like @readme/markdown/styles/main.css keep resolving. (Without that, introducing an exports block would have thrown ERR_PACKAGE_PATH_NOT_EXPORTED for any path not explicitly listed.)

3. Does exporting the diff logic open up using it in the monorepo pipeline against real customer projects?

Yes — that's exactly the intended use. render-diff is engine-agnostic: neutral left/right API (diff(leftHtml, rightHtml, opts)), no engine dependency, 152 KB, with the engine isolation enforced in CI. So the pipeline can render real customer docs and diff the HTML across @readme/markdown version bumps for parity gating. The 'minimal' preset is built for same-engine before/after diffing; pair it with render-fixture when the pipeline also needs to do the rendering.

"should handle very long content without crashing" parses a 10k-char [block:callout] body (~2s in isolation). Under full-suite parallel contention in CI it exceeded the 5s default and timed out. It's a pure not.toThrow() guard with no latency assertion, so give it a generous 30s timeout — still catches a genuine hang/backtracking regression while tolerating worker contention. Claude-Session: https://claude.ai/code/session_01GPTShf49qTsVP1AxSbpRJk

The fixture used a Font Awesome class icon (far fa-car-bolt), whose render is changing on `next`: the centralized <Icon> component renders FA icons as <i> instead of <span>. PR CI checks the branch merged with the current `next` tip, so the committed <span> snapshot mismatched CI's <i> output. Switch to an emoji icon — the reviewer's own #1498 example: a blockquote `> 📘` callout paired with the equivalent <Callout icon="📘"> JSX. Emoji icons render as <span> identically before and after the <Icon> change, so the fixture is stable across the merge while still pinning blockquote-vs-JSX render parity. README and index updated to match. Claude-Session: https://claude.ai/code/session_01GPTShf49qTsVP1AxSbpRJk

After merging the latest next (new Icon component + engine growth), the main.node.js gzip bundle reached 950.09KB — 90 bytes over the 950KB budget that #1502 had just set, leaving next itself red on this check. Bump to 975KB for headroom. The render-diff/render-fixture subpath budgets are unaffected and remain well within limits. Claude-Session: https://claude.ai/code/session_01GPTShf49qTsVP1AxSbpRJk

jamestclark · 2026-06-26T00:01:05Z

@kevinports @eaglethrost
Added some test fixtures for the tables & callouts.

3a80823

eaglethrost · 2026-06-26T01:10:33Z

Thanks for the answers James, looks good!

rafegoldberg · 2026-06-29T22:11:10Z

This PR was released!

🚀 Changes included in v14.10.2

jamestclark added 3 commits May 26, 2026 15:14

jamestclark marked this pull request as draft May 26, 2026 22:22

jamestclark changed the title ~~Regressions testing~~ Add render-diff and render-fixture subpath exports + MDX↔MDXish regression suite May 26, 2026

jamestclark added 6 commits May 27, 2026 12:29

test(regression): rename Suite A describe to "Render engine regressio…

60e3c06

…ns tests" Matches the parallel rename of Suite B → "MDX↔MDXish equivalence" and refreshes the snapshot keys accordingly.

update snapshots

ff283c5

jamestclark marked this pull request as ready for review June 4, 2026 23:59

jamestclark requested review from a team, darrenyong, eaglethrost, jbury, kevinports and maximilianfalco June 4, 2026 23:59

eaglethrost reviewed Jun 9, 2026

View reviewed changes

erunion requested review from a team and removed request for a team, darrenyong and jbury June 16, 2026 21:11

Merge remote-tracking branch 'origin/next' into tide-helmet

40d8e4a

# Conflicts: # package.json

kevinports reviewed Jun 24, 2026

View reviewed changes

jamestclark added 3 commits June 25, 2026 16:23

Merge remote-tracking branch 'origin/next' into tide-helmet

16bd140

eaglethrost approved these changes Jun 26, 2026

View reviewed changes

jamestclark merged commit 5736bc0 into next Jun 29, 2026
8 checks passed

jamestclark deleted the tide-helmet branch June 29, 2026 17:29

rafegoldberg added the released label Jun 29, 2026

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Add render-diff and render-fixture subpath exports + MDX↔MDXish regression suite#1489

Add render-diff and render-fixture subpath exports + MDX↔MDXish regression suite#1489
jamestclark merged 15 commits into
nextfrom
tide-helmet

jamestclark commented May 26, 2026 •

edited

Loading

Uh oh!

eaglethrost left a comment

Uh oh!

eaglethrost Jun 9, 2026

Uh oh!

kevinports Jun 24, 2026

Uh oh!

kevinports left a comment

Uh oh!

kevinports Jun 24, 2026

Uh oh!

kevinports Jun 24, 2026

Uh oh!

kevinports Jun 24, 2026

Uh oh!

jamestclark commented Jun 25, 2026

Uh oh!

jamestclark commented Jun 26, 2026 •

edited

Loading

Uh oh!

eaglethrost commented Jun 26, 2026

Uh oh!

Uh oh!

rafegoldberg commented Jun 29, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

Uh oh!

Conversation

jamestclark commented May 26, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

🎯 What does this PR do?

New public surface

Why the diff tool is engine-agnostic

Regression suite — 13 fixtures

Build verification

Commit-by-commit

🧪 QA tips

📸 Screenshot or Loom

Uh oh!

eaglethrost left a comment

Choose a reason for hiding this comment

Uh oh!

eaglethrost Jun 9, 2026

Choose a reason for hiding this comment

Uh oh!

kevinports Jun 24, 2026

Choose a reason for hiding this comment

Uh oh!

kevinports left a comment

Choose a reason for hiding this comment

Uh oh!

kevinports Jun 24, 2026

Choose a reason for hiding this comment

Uh oh!

kevinports Jun 24, 2026

Choose a reason for hiding this comment

Uh oh!

kevinports Jun 24, 2026

Choose a reason for hiding this comment

Uh oh!

jamestclark commented Jun 25, 2026

Uh oh!

jamestclark commented Jun 26, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

eaglethrost commented Jun 26, 2026

Uh oh!

Uh oh!

rafegoldberg commented Jun 29, 2026

This PR was released!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

jamestclark commented May 26, 2026 •

edited

Loading

jamestclark commented Jun 26, 2026 •

edited

Loading