@takazudo/zfb-md-wasm 2.13.0 → 2.14.0

package/README.md

@@ -31,12 +31,12 @@ migration. `./highlight` is also backward-compatible. New `./render` and
 `zfb-render` but intentionally retain `zfb-content` and `syntect-fancy`.
 `./parse` is not syntect-free.
 
-| Entry | gzip-9 wasm (2.13.0) | Exact runtime values | Exact exported types |
+| Entry | gzip-9 wasm (2.14.0) | Exact runtime values | Exact exported types |
 | --- | ---: | --- | --- |
-| `.` | 1,457,796 B | `init`, `compile`, `renderHtml`, `parseToAst`, `highlightCode`, `version`, `__forceTrapForTests`, `__getTrapRecoveryStateForTests`, `toMdastRoot`, `ZfbMdWasmTrapError`, `ZfbMdWasmTrapRecoveryLimitError`, `MdastAdapterError` | Full current compile/render/parse/raw-mdast/highlight surface |
-| `./highlight` | 758,250 B | `init`, `highlightCode`, `version`, `__forceTrapForTests`, `__getTrapRecoveryStateForTests`, `ZfbMdWasmTrapError`, `ZfbMdWasmTrapRecoveryLimitError` | `HighlightRole`, `HighlightCodeOptions`, `HighlightCodeResult`, `HighlightDiagnostic`, `HighlightDiagnosticSource` |
-| `./render` | 1,010,801 B | `init`, `renderHtml`, `version`, `ZfbMdWasmTrapError`, `ZfbMdWasmTrapRecoveryLimitError`, `__forceTrapForTests`, `__getTrapRecoveryStateForTests` | `RenderHtmlResult`, `Diagnostic`, `DiagnosticSource`, `ZfbMdWasmOptions`, `ParseDialect`, `PipelineOptions`, `GfmOptions`, `CodeHighlightMode`, `CodeHighlightOptions`, `MarkdownFeaturesConfig`, `JsxRuntime`, `HighlightRole` |
-| `./parse` | 276,433 B | `init`, `parseToAst`, `toMdastRoot`, `MdastAdapterError`, `version`, `ZfbMdWasmTrapError`, `ZfbMdWasmTrapRecoveryLimitError`, `__forceTrapForTests`, `__getTrapRecoveryStateForTests` | `ParseToAstResult`, `ParseToAstOptions`, `ParseDialect`, `FrontmatterPolicy`, `ParsePipelineOptions`, `Diagnostic`, `DiagnosticSource`, `AstPoint`, `AstPosition`, `RawMdastData`, `MarkdownRsStop`, `MdastNode`, `MdastRoot`, `UnknownMdastNode`, `Root`, `Paragraph`, `Heading`, `ThematicBreak`, `Blockquote`, `List`, `ListItem`, `Html`, `Code`, `Definition`, `Text`, `DirectiveNodeBase`, `ContainerDirective`, `LeafDirective`, `TextDirective`, `Emphasis`, `Strong`, `InlineCode`, `Break`, `Link`, `Image`, `ReferenceKind`, `LinkReference`, `ImageReference`, `FootnoteDefinition`, `FootnoteReference`, `TableAlign`, `Table`, `TableRow`, `TableCell`, `Delete`, `Yaml`, `MdxFlowExpression`, `MdxTextExpression`, `MdxJsxFlowElement`, `MdxJsxTextElement`, `MdxJsxAttributeContent`, `MdxJsxAttribute`, `MdxJsxAttributeValueExpression`, `MdxJsxExpressionAttribute` |
+| `.` | 1,443,041 B | `init`, `compile`, `renderHtml`, `parseToAst`, `highlightCode`, `version`, `__forceTrapForTests`, `__getTrapRecoveryStateForTests`, `toMdastRoot`, `ZfbMdWasmTrapError`, `ZfbMdWasmTrapRecoveryLimitError`, `MdastAdapterError` | Full current compile/render/parse/raw-mdast/highlight surface |
+| `./highlight` | 758,247 B | `init`, `highlightCode`, `version`, `__forceTrapForTests`, `__getTrapRecoveryStateForTests`, `ZfbMdWasmTrapError`, `ZfbMdWasmTrapRecoveryLimitError` | `HighlightRole`, `HighlightCodeOptions`, `HighlightCodeResult`, `HighlightDiagnostic`, `HighlightDiagnosticSource` |
+| `./render` | 1,010,796 B | `init`, `renderHtml`, `version`, `ZfbMdWasmTrapError`, `ZfbMdWasmTrapRecoveryLimitError`, `__forceTrapForTests`, `__getTrapRecoveryStateForTests` | `RenderHtmlResult`, `Diagnostic`, `DiagnosticSource`, `ZfbMdWasmOptions`, `ParseDialect`, `PipelineOptions`, `GfmOptions`, `CodeHighlightMode`, `CodeHighlightOptions`, `MarkdownFeaturesConfig`, `JsxRuntime`, `HighlightRole` |
+| `./parse` | 261,099 B | `init`, `parseToAst`, `toMdastRoot`, `MdastAdapterError`, `version`, `ZfbMdWasmTrapError`, `ZfbMdWasmTrapRecoveryLimitError`, `__forceTrapForTests`, `__getTrapRecoveryStateForTests` | `ParseToAstResult`, `ParseToAstOptions`, `ParseDialect`, `FrontmatterPolicy`, `ParsePipelineOptions`, `Diagnostic`, `DiagnosticSource`, `AstPoint`, `AstPosition`, `RawMdastData`, `MarkdownRsStop`, `MdastNode`, `MdastRoot`, `UnknownMdastNode`, `Root`, `Paragraph`, `Heading`, `ThematicBreak`, `Blockquote`, `List`, `ListItem`, `Html`, `Code`, `Definition`, `Text`, `DirectiveNodeBase`, `ContainerDirective`, `LeafDirective`, `TextDirective`, `Emphasis`, `Strong`, `InlineCode`, `Break`, `Link`, `Image`, `ReferenceKind`, `LinkReference`, `ImageReference`, `FootnoteDefinition`, `FootnoteReference`, `TableAlign`, `Table`, `TableRow`, `TableCell`, `Delete`, `Yaml`, `MdxFlowExpression`, `MdxTextExpression`, `MdxJsxFlowElement`, `MdxJsxTextElement`, `MdxJsxAttributeContent`, `MdxJsxAttribute`, `MdxJsxAttributeValueExpression`, `MdxJsxExpressionAttribute` |
 
 The focused entries own private resource pairs:
 
@@ -683,8 +683,8 @@ suite gates exact-match). Deliberate limitations of the browser build:
 - **Choose a focused artifact for non-compile calls.** The root remains the
   compatibility entry and carries the complete compiler graph. `./highlight`
   keeps its public API and resources while the post-#2449/#2450 graph is
-  proven SWC-free — and its payload shows it: 758,250 B gzip-9 versus
-  1,457,796 B for root. `./render` and `./parse` are also SWC-free and omit
+  proven SWC-free — and its payload shows it: 758,247 B gzip-9 versus
+  1,443,041 B for root. `./render` and `./parse` are also SWC-free and omit
   `zfb-render`; parse intentionally retains `zfb-content`/`syntect-fancy`, so
   it is not syntect-free. Use `./render` or `./parse` when a consumer does
   not need `compile`; root and highlight callers otherwise require no
@@ -718,36 +718,36 @@ three-step repair sequence:
 2. Run `node scripts/assert-zfb-md-wasm-budgets.mjs --build-log "$BUILD_LOG" --dist crates/zfb-md-wasm/npm/dist --update-manifest`.
 3. Run `node scripts/assert-md-wasm-size-docs.mjs --fix`, then `pnpm format:mdx`.
 
-These are the shipped **2.13.0** artifact rows — optimized final wasm after
+These are the shipped **2.14.0** artifact rows — optimized final wasm after
 wasm-bindgen and wasm-opt, Node `gzipSync(..., { level: 9 })`, and glue
 bytes/gzip:
 
 | Entry/graph | final wasm | gzip-9 | glue | glue gzip-9 |
 | --- | ---: | ---: | ---: | ---: |
-| root (full) | 3,273,252 B | 1,457,796 B | 14,998 B | 4,199 B |
-| highlight | 1,476,740 B | 758,250 B | 8,758 B | 2,637 B |
-| render | 2,082,385 B | 1,010,801 B | 8,772 B | 2,661 B |
-| parse | 624,976 B | 276,433 B | 11,159 B | 3,797 B |
+| root (full) | 3,294,286 B | 1,443,041 B | 14,998 B | 4,199 B |
+| highlight | 1,476,740 B | 758,247 B | 8,758 B | 2,637 B |
+| render | 2,082,385 B | 1,010,796 B | 8,772 B | 2,661 B |
+| parse | 646,115 B | 261,099 B | 11,159 B | 3,797 B |
 
 The #2447 decision snapshot measured the split package at 3,638,607 B versus
 2,314,818 B for the root-plus-highlight package. Locked gzip-9 ceilings are
 root 1,600,000 B, highlight 820,000 B, render 1,100,000 B, and parse
 325,000 B; the complete packed tarball ceiling is 3,900,000 B. All four ship
-inside their ceilings, with 142,204 B (root), 61,750 B (highlight), 89,199 B
-(render), and 48,567 B (parse) of headroom. These are 2.13.0 measurements, not
+inside their ceilings, with 156,959 B (root), 61,753 B (highlight), 89,204 B
+(render), and 63,901 B (parse) of headroom. These are 2.14.0 measurements, not
 permanent promises — re-measure against the version you actually install.
 The clean four-step production ceiling is 210 seconds; the selected #2447
 median was 155.015 s [153.496, 165.977].
 
 Gating `swc_core` out of the highlight graph (#2449/#2450) was a
-**provability win, not a size win**. The shipped highlight artifact is only
-7,965 B smaller than #2447's SWC-retaining baseline (1,484,705 B →
-1,476,740 B; gzip-9 767,009 B → 758,244 B, −8,765 B) — wasm-opt was already
+**provability win, not a size win**. The #2447 SWC-retaining baseline was
+1,484,705 B raw and 767,009 B gzip-9; the #2450 result was 7,965 B smaller
+raw and 8,765 B smaller gzip-9 (758,244 B) — wasm-opt was already
 dead-stripping the unreachable `swc_core`, and #2450's exact-parity and
 no-`swc_core` assertions turned that emergent property into a guaranteed one.
 The delta that matters to a highlight-only consumer is root versus
-highlight: the highlight artifact is 1,796,512 B smaller raw and 699,546 B
-smaller gzip-9, landing at about 45% of root's raw bytes and 52% of its
+highlight: the highlight artifact is 1,817,546 B smaller raw and 684,794 B
+smaller gzip-9, landing at about 45% of root's raw bytes and 53% of its
 gzipped bytes.
 
 Every shipped final wasm came in under its #2447 candidate measurement: root

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@takazudo/zfb-md-wasm",
-  "version": "2.13.0",
+  "version": "2.14.0",
   "private": false,
   "type": "module",
   "description": "zfb's md/mdx -> JS/HTML conversion pipeline compiled to WebAssembly, for browser-side dynamic conversion (CMS live preview with parity to zfb's production output).",
@@ -82,6 +82,7 @@
   },
   "scripts": {
     "build": "node scripts/build.mjs",
+    "pretest": "node scripts/assert-consumer-artifacts.mjs",
     "test": "pnpm typecheck:consumer && vitest run",
     "test:watch": "vitest",
     "typecheck": "tsc --noEmit",