diff --git a/CHANGELOG.md b/CHANGELOG.md index 9e1caa0..36aa60e 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -91,6 +91,10 @@ list lines ran the call out of memory. - Fix the markdown readers past the longest string or array the JavaScript engine builds: they return `unsupported-node-shape` where they threw. +- Fix a paragraph of tens of millions of characters: `adfToLosslessMarkdown` and + `adfToPortableMarkdown` write one up to 536870888 characters within Node's default heap, where 80 + million characters ran out of memory and 20 million `*` threw `RangeError`. `losslessMarkdownToAdf` + no longer throws on such a paragraph inside a directive. - Fix deeply nested compact JSON in an `adf:` fence: `losslessMarkdownToAdf` refuses it as `unsupported-node-shape`, as it refuses any compact JSON there, where about a megabyte of it threw. - Add `adfToPortableMarkdown` and `portableMarkdownToAdf`, a lossy pair converting ADF to and from diff --git a/src/markdown/emit/adf-to-markdown.test.ts b/src/markdown/emit/adf-to-markdown.test.ts index 2150f70..389c015 100644 --- a/src/markdown/emit/adf-to-markdown.test.ts +++ b/src/markdown/emit/adf-to-markdown.test.ts @@ -70,6 +70,12 @@ test('refuses text past V8\'s longest string', () => { assert.equal(markdown(result), tooLong) }) +// One escape past the most entries V8's Set holds. +test('escapes more characters in a line than a Set holds', () => { + const stars = '*'.repeat(2 ** 24 + 1) + assert.equal(markdown(adfToLosslessMarkdown(document(paragraph({ text: stars, type: 'text' })))), `${'\\*'.repeat(2 ** 24 + 1)}\n`) +}) + test('refuses a document version the markdown cannot carry', () => { assert.equal(code(adfToLosslessMarkdown({ type: 'doc', version: 2 })), 'unsupported-document-version') }) diff --git a/src/markdown/emit/line-escaping.ts b/src/markdown/emit/line-escaping.ts index 941ef81..9052be7 100644 --- a/src/markdown/emit/line-escaping.ts +++ b/src/markdown/emit/line-escaping.ts @@ -32,15 +32,20 @@ type EmittedDelimiter = { closes: boolean; offset: number; pair: number; width: type EmittedRun = { canClose: boolean; canOpen: boolean; character: string; delimiters: EmittedDelimiter[]; length: number; start: number } +// One byte a character, never an array element or Set entry: a line may run to hundreds of millions of characters. +type Escapings = (index: number) => InlineEscaping | undefined + type EscapeWalk = { readonly claims: Claims readonly container: LineContainer - readonly escapings: readonly InlineEscaping[] + readonly escapings: Escapings readonly headingCloser: number | undefined readonly linkClose: number readonly scan: string } +const escapingCodes: readonly InlineEscaping[] = ['backslash', 'bracketed', 'bracketed-link-target', 'none'] + const followsLinkText = /[([]/ const runCharacters = '*_`~' @@ -80,16 +85,19 @@ function resolveEmphasis(segments: readonly InlineSegment[]): InlineSegment[] { function escape(segments: readonly InlineSegment[], container: LineContainer, claims: Claims): AssembledLine { const scan = segments.map((segment) => segment.text).join('') - const escapings: InlineEscaping[] = [] - for (const segment of segments) for (let index = 0; index < segment.text.length; index += 1) escapings.push(segment.escaping) - const escaped = escapeClosedRuns(scan, escapings, escapeClaims(scan, escapings, container, claims)) - const placements: number[] = [] - let output = '' - for (let index = 0; index < scan.length; index += 1) { - if (escaped.has(index)) output += '\\' - placements.push(output.length) - output += scan.charAt(index) + const codes = new Uint8Array(scan.length) + let start = 0 + for (const segment of segments) { + codes.fill(escapingCodes.indexOf(segment.escaping), start, start + segment.text.length) + start += segment.text.length } + const escapings: Escapings = (index) => { + const code = codes[index] + return code === undefined ? undefined : escapingCodes[code] + } + const escaped = escapeClaims(scan, escapings, container, claims) + escapeClosedRuns(scan, escapings, escaped) + const { output, placements } = backslashed(segments, scan, escaped) if (container === 'paragraph' && opensLinkDefinition(output)) { if (segments[0]?.nodes !== undefined) return { line: output, openingLinkAsDirective: true, unspellableRuns: [] } return { line: `\\${output}`, unspellableRuns: unspellableRuns(segments, output, placements) } @@ -97,7 +105,37 @@ function escape(segments: readonly InlineSegment[], container: LineContainer, cl return { line: output, unspellableRuns: unspellableRuns(segments, output, placements) } } -function escapeClaims(scan: string, escapings: readonly InlineEscaping[], container: LineContainer, claims: Claims): ReadonlySet { +// The output, and where each segment's first character landed in it. +function backslashed(segments: readonly InlineSegment[], scan: string, escaped: Uint8Array): { output: string; placements: number[] } { + // Joined a few thousand pieces at a time: an array element per escape would cost more than the character it escapes. + const chunks: string[] = [] + let pieces: string[] = [] + let from = 0 + for (let index = escaped.indexOf(1); index !== -1; index = escaped.indexOf(1, index + 1)) { + pieces.push(scan.slice(from, index), '\\') + from = index + if (pieces.length < 4096) continue + chunks.push(pieces.join('')) + pieces = [] + } + pieces.push(scan.slice(from)) + chunks.push(pieces.join('')) + const placements: number[] = [] + let escapes = 0 + let next = escaped.indexOf(1) + let start = 0 + for (const segment of segments) { + while (next !== -1 && next <= start) { + escapes += 1 + next = escaped.indexOf(1, next + 1) + } + placements.push(start + escapes) + start += segment.text.length + } + return { output: chunks.join(''), placements } +} + +function escapeClaims(scan: string, escapings: Escapings, container: LineContainer, claims: Claims): Uint8Array { const walk: EscapeWalk = { claims, container, @@ -106,20 +144,20 @@ function escapeClaims(scan: string, escapings: readonly InlineEscaping[], contai linkClose: lastLinkClose(scan, escapings), scan, } - const escaped = new Set() + const escaped = new Uint8Array(scan.length) let line = scanLine(scan, 0) let index = 0 while (index < scan.length) { while (index > line.start + line.text.length) line = scanLine(scan, line.start + line.text.length + 1) if (runCharacters.includes(scan.charAt(index))) { const end = index + runLength(scan, index) - for (const claimed of runEscapes(walk, index, end, line)) escaped.add(claimed) + runEscapes(walk, index, end, line, escaped) index = end continue } const token = readInlineToken(scan, index, claims) const claimed = claimsCharacter(walk, index, token, line) - if (claimed) escaped.add(index) + if (claimed) escaped[index] = 1 // The reader takes a `==` whole, so its second `=` starts nothing unless the first is escaped. index += !claimed && token.kind === 'highlight' ? token.width : 1 } @@ -127,25 +165,27 @@ function escapeClaims(scan: string, escapings: readonly InlineEscaping[], contai } // A text character beside one of the emitter's delimiters would lengthen that delimiter's run, so every text character in such a run escapes. -function runEscapes(walk: EscapeWalk, start: number, end: number, line: ScanLine): number[] { +function runEscapes(walk: EscapeWalk, start: number, end: number, line: ScanLine, escaped: Uint8Array): void { const { escapings, scan } = walk - const indexes: number[] = [] - const run = escapings.slice(start, end) - if (run.some(isSyntax) && run.some(isEscapable)) { - for (let index = start; index < end; index += 1) if (isEscapable(escapings[index])) indexes.push(index) - return indexes + if (runHolds(escapings, start, end, isSyntax) && runHolds(escapings, start, end, isEscapable)) { + for (let index = start; index < end; index += 1) if (isEscapable(escapings(index))) escaped[index] = 1 + return } for (let index = start; index < end; index += 1) { // A backtick run escapes whole: a bare rest would be a shorter backtick string, which can close a span an earlier run opens. const claimed = (index > start && scan.charAt(index) === '`') || claimsRunRest(walk, index, end, line) if (!claimed) break - indexes.push(index) + escaped[index] = 1 } - return indexes +} + +function runHolds(escapings: Escapings, start: number, end: number, kind: (escaping: InlineEscaping | undefined) => boolean): boolean { + for (let index = start; index < end; index += 1) if (kind(escapings(index))) return true + return false } function claimsRunRest({ claims, container, escapings, scan }: EscapeWalk, index: number, end: number, line: ScanLine): boolean { - const escaping = escapings[index] + const escaping = escapings(index) if (escaping === 'bracketed-link-target') return readInlineToken(scan, index, claims).kind === 'code-span' if (!isEscapable(escaping)) return false if (claimsLineStart(line, index, container)) return true @@ -154,7 +194,7 @@ function claimsRunRest({ claims, container, escapings, scan }: EscapeWalk, index } function claimsCharacter({ container, escapings, headingCloser, linkClose, scan }: EscapeWalk, index: number, token: InlineToken, line: ScanLine): boolean { - const escaping = escapings[index] + const escaping = escapings(index) if (escaping === 'bracketed-link-target') return token.kind === 'directive' if (!isEscapable(escaping)) return false if (index === headingCloser || claimsLineStart(line, index, container)) return true @@ -169,7 +209,7 @@ function claimsCharacter({ container, escapings, headingCloser, linkClose, scan case 'html': return true case 'bracket': - if (token.image) return isSyntax(escapings[index + 1]) + if (token.image) return isSyntax(escapings(index + 1)) return escaping === 'bracketed' || index < linkClose case 'bracket-close': return escaping === 'bracketed' @@ -180,14 +220,14 @@ function claimsCharacter({ container, escapings, headingCloser, linkClose, scan case 'text': { const character = scan.charAt(index) if (character === '|') return container === 'table-cell' - return character === '{' && scan.charAt(index - 1) === ']' && isSyntax(escapings[index - 1]) + return character === '{' && scan.charAt(index - 1) === ']' && isSyntax(escapings(index - 1)) } } } // CommonMark reads no escape inside a code span, so a backtick string left beside an escape still closes a span that an earlier bare run of its length opens. -function escapeClosedRuns(scan: string, escapings: readonly InlineEscaping[], claimed: ReadonlySet): ReadonlySet { - const escaped = new Set(claimed) +// The walk runs backwards and escapes only the run it just read, so every index it reads still holds the claims alone. +function escapeClosedRuns(scan: string, escapings: Escapings, escaped: Uint8Array): void { // The lengths of the backtick strings escapes leave: a bare run of one of these lengths would open a span one of them closes, so it is escaped too. const formed = new Set() let end = scan.length - 1 @@ -200,19 +240,18 @@ function escapeClosedRuns(scan: string, escapings: readonly InlineEscaping[], cl while (scan.charAt(start - 1) === '`') start -= 1 let segmentEnd = end for (let index = end; index > start; index -= 1) { - if (!claimed.has(index)) continue + if (escaped[index] !== 1) continue formed.add(segmentEnd - index + 1) segmentEnd = index - 1 } - if (segmentEnd !== end || claimed.has(start)) formed.add(segmentEnd - start + 1) - else if (escapings[start] !== 'none' && formed.has(end - start + 1)) { - for (let index = start; index <= end; index += 1) escaped.add(index) + if (segmentEnd !== end || escaped[start] === 1) formed.add(segmentEnd - start + 1) + else if (escapings(start) !== 'none' && formed.has(end - start + 1)) { + escaped.fill(1, start, end + 1) // Each escaped backtick then stands alone, a string of length one. formed.add(1) } end = start - 1 } - return escaped } // One emphasis run, the innermost, or every highlight run the line cannot spell. @@ -226,10 +265,8 @@ function unspellableRuns(segments: readonly InlineSegment[], output: string, pla // No `==` in text can open or close, and highlights never nest, so a pair reads back where each delimiter flanks. function unreadHighlights(segments: readonly InlineSegment[], output: string, placements: readonly number[]): MarkRun[] { const unread: MarkRun[] = [] - let cursor = 0 - for (const segment of segments) { - const start = placements[cursor] ?? 0 - cursor += segment.text.length + for (const [index, segment] of segments.entries()) { + const start = placements[index] ?? 0 if (segment.highlight === undefined) continue const flanking = highlightFlanking(output, start) const flanks = segment.highlight === 'open' ? flanking.opens : flanking.closes @@ -272,10 +309,8 @@ function emittedRuns(segments: readonly InlineSegment[], placements: readonly nu const runs: EmittedRun[] = [] const nodes: MarkRun[] = [] const open: number[] = [] - let cursor = 0 - for (const segment of segments) { - const start = placements[cursor] ?? 0 - cursor += segment.text.length + for (const [index, segment] of segments.entries()) { + const start = placements[index] ?? 0 if (segment.emphasis === undefined) continue const closes = segment.emphasis === 'close' const pair = closes ? (open.pop() ?? nodes.length) : nodes.length @@ -336,9 +371,9 @@ function scanLine(scan: string, start: number): ScanLine { } // A `]` the emitter spelled sits inside a construct that binds before link text does. -function lastLinkClose(scan: string, escapings: readonly (InlineEscaping | undefined)[]): number { +function lastLinkClose(scan: string, escapings: Escapings): number { for (let cursor = scan.length - 1; cursor >= 0; cursor -= 1) { - if (scan.charAt(cursor) !== ']' || isSyntax(escapings[cursor])) continue + if (scan.charAt(cursor) !== ']' || isSyntax(escapings(cursor))) continue if (followsLinkText.test(scan.charAt(cursor + 1))) return cursor } return -1 diff --git a/src/markdown/parse/markdown-to-adf.test.ts b/src/markdown/parse/markdown-to-adf.test.ts index dc03e7e..ad1e2d9 100644 --- a/src/markdown/parse/markdown-to-adf.test.ts +++ b/src/markdown/parse/markdown-to-adf.test.ts @@ -341,6 +341,15 @@ test('refuses the directive form of a blockquote opening with a code span', () = assert.equal(content(losslessMarkdownToAdf('!adf:blockquote\n` `` `\n!adf:/blockquote\n')), 'unsupported-node-shape: blockquote takes the CommonMark spelling, not the directive form') }) +// Some 1000 characters short of V8's longest string, which the quote's `> ` on each of its 600 lines passes. +test('refuses a blockquote directive whose CommonMark spelling would pass V8\'s longest string', () => { + const fence = `\`\`\`\n${`${'a'.repeat(894_782)}\n`.repeat(300)}\`\`\`\n` + assert.equal( + content(losslessMarkdownToAdf(`!adf:blockquote\n${fence}\n${fence}!adf:/blockquote\n`)), + 'unsupported-node-shape: the markdown for this document, or for a directive\'s plain CommonMark, would be longer than 536870888 characters, the longest this library writes', + ) +}) + test('names the directive name no node reads back to', () => { assert.equal(code(losslessMarkdownToAdf('!adf:widget info\nx\n!adf:/widget\n')), 'unknown-directive-name') assert.equal(content(losslessMarkdownToAdf('!adf:widget\n')), 'unknown-directive-name: the directive name widget reads back to no node; \\!adf: keeps the prefix literal') diff --git a/todo.md b/todo.md index ef873c4..7ad133d 100644 --- a/todo.md +++ b/todo.md @@ -25,7 +25,6 @@ | ID | Release | Exempt | Item | R | S | A | G | Goals | Score | |---|---|---|---|---|---|---|---|---|---| -| 108 | 0.2.0 | defect | **Spell a line in memory linear in its length, so a paragraph shorter than V8's longest string converts within Node's default heap.** | 4 | 6 | 5 | 8 | 1, 9 | 20.5 | | 103 | 0.2.0 | | **Fold `unspellable-whitespace` into `unsupported-node-shape`, so a line ending and a null character in a content slot take one code.** | 2 | 2 | 5 | 6 | 1 | 19.5 | | 84 | 0.2.0 | | **Read GFM's pipe table without leading pipes, and with alignment colons, in `portableMarkdownToAdf`, dropping the alignment.** | 3 | 4 | 7 | 7 | 4, 7 | 19.4 | | 52 | 0.2.0 | | **Spell `colwidth` as a comma list, `colwidth="340,420"`.** | 3 | 3 | 5 | 6 | 6 | 11.5 | @@ -75,17 +74,6 @@ ## Details -### 108. Spell a line in memory linear in its length, so a paragraph shorter than V8's longest string converts within Node's default heap. - -The line escaper keeps an array element per character twice (`escapings`, `placements` in `escape`, -`line-escaping.ts`) and a `Set` entry per escaped character, so `adfToLosslessMarkdown` holds about -50 bytes a character: one 20-million-character text node took 1.2 GB on Node 24, and one of 80 -million exhausted its 2.2 GB default heap. Twenty million `*` throw `RangeError: Set maximum size -exceeded`, which `withinEngineLimits` lets through. `losslessMarkdownToAdf` reaches both through a -directive's CommonMark ask, and reading 537 MB of markdown, a `!adf:bulletList` of two code blocks -just under the length bound, took 3.4 GB on Bun, too much for the gate to test its length refusal: -add that test with this fix. Found while shipping item 106, 2026-10-11. - ### 103. Fold `unspellable-whitespace` into `unsupported-node-shape`, so a line ending and a null character in a content slot take one code. `slotFault` in `directive-syntax.ts` gives a newline in an `emoji`, `mention` or `status` content