3
0
mirror of https://github.com/larvit/adf-codec.git synced 2026-11-19 14:45:44 +01:00
Files
adf-codec/src/markdown/emit/line-escaping.ts
T

483 lines
24 KiB
TypeScript

import type { Claims, WrittenFlavour } from '../flavour.ts'
import type { InlineReading } from '../inline-content.ts'
import type { InlineSource, InlineToken } from '../inline-tokens.ts'
import type { LineContainer } from '../line-container.ts'
import { delimiterRunToken, highlightFlanking, inlineSource, readInlineToken } from '../inline-tokens.ts'
import { escapesLineClaim, type LinePosition } from '../commonmark/grammar.ts'
import { flavourClaims } from '../flavour.ts'
import { isBareDelimiterRow } from '../pipe-table-syntax.ts'
import { delimiterFlags, isWordCharacter, runLength } from '../commonmark/emphasis-matching.ts'
import { opensLinkDefinition } from '../commonmark/link-reference-definitions.ts'
import { readLineBack } from '../inline-content.ts'
export type DelimiterRole = 'close' | 'open'
// A segment's kind: `backslash` text, `bracketed` text inside `[…]` (link text, alt, a directive's content), `bracketed-link-target` syntax: a link target inside `[…]`, `none` syntax written as is.
export type InlineEscaping = (typeof escapingCodes)[number]
export type NodeRange = { first: number; last: number }
// `depth` is the index into each node's marks array that the run spells.
export type MarkRun = NodeRange & { depth: number }
export type InlineSegment =
| { emphasis: DelimiterRole; escaping: 'none'; highlight?: undefined; nodes: MarkRun; text: string }
| { emphasis?: undefined; escaping: 'none'; highlight: DelimiterRole; nodes: MarkRun; text: string }
| { emphasis?: undefined; escaping: 'none'; highlight?: undefined; nodes: NodeRange; text: string }
| { emphasis?: undefined; escaping: InlineEscaping; highlight?: undefined; nodes?: undefined; text: string }
export type AssembledLine = { line: string; openingLinkAsDirective?: true; unspellableRuns: MarkRun[] }
// How the caller takes a run back: a carry swaps its nodes for one blob, splitting the runs around it; a drop removes its mark, moving no other delimiter where `alone` holds.
export type RunFallback = { kind: 'carry' } | { alone: (run: MarkRun) => boolean; kind: 'drop' }
type ScanLine = { position: LinePosition; start: number; text: string }
// An emphasis run's delimiters, written alike, by index into the placed segments.
type EmphasisPair = { closer: number; delimiter: string; nodes: MarkRun; opener: number }
type Escapings = (index: number) => InlineEscaping | undefined
// Which emphasis pairs fall with the first misread one, each marked in `fallen` once settled.
type PairSweep = {
readonly fallback: RunFallback
readonly fallen: Uint8Array
readonly first: number
// The leaf boundaries where dropping a pair that may fall first can merge the runs meeting there.
readonly merging: ReadonlySet<number>
readonly output: string
// The pair each placed segment delimits, -1 where it delimits none.
readonly pairAt: Int32Array
readonly pairs: readonly EmphasisPair[]
readonly placed: readonly PlacedSegment[]
}
// Where a segment's first character landed in the output.
type PlacedSegment = { segment: InlineSegment; start: number }
// How a pair stands while another falls: `either` where it may fall first.
type Standing = 'either' | 'gone' | 'there'
// The escaped line once reading it back asks for no further escape.
type ReadBack = { output: string; placed: PlacedSegment[]; reading: InlineReading }
// Which of a run's characters escape: every escapable one, or every one before `end`.
type RunEscapes = { end: number; escapable: false } | { escapable: true }
// Which text characters escape wherever they stand: one of them opened a construct when the line was read back.
type TextEscapes = { backticks: boolean; brackets: boolean }
type EscapeWalk = {
readonly claims: Claims
readonly container: LineContainer
readonly escapings: Escapings
readonly headingCloser: number | undefined
readonly inline: InlineSource
}
const escapingCodes: readonly ['backslash', 'bracketed', 'bracketed-link-target', 'none'] = ['backslash', 'bracketed', 'bracketed-link-target', 'none']
const runCharacters = '*_`~'
export function assembleInlineLine(segments: readonly InlineSegment[], container: LineContainer, flavour: WrittenFlavour, fallback: RunFallback): AssembledLine {
return escape(resolveEmphasis(segments), container, flavourClaims[flavour], fallback)
}
function resolveEmphasis(segments: readonly InlineSegment[]): InlineSegment[] {
const resolved = segments.map((segment) => ({ ...segment }))
// Offsets index the pre-swap text: every emphasis spelling this swaps between is one character wide.
const scan = resolved.map((segment) => segment.text).join('')
const offsets: number[] = []
let offset = 0
for (const segment of resolved) {
offsets.push(offset)
offset += segment.text.length
}
const open: number[] = []
for (let index = 0; index < resolved.length; index += 1) {
const segment = resolved[index]
if (segment === undefined) continue
if (segment.emphasis === 'open') open.push(index)
if (segment.emphasis !== 'close') continue
const openerIndex = open.pop()
const opener = openerIndex === undefined ? undefined : resolved[openerIndex]
if (openerIndex === undefined || opener === undefined) continue
const openOffset = offsets[openerIndex] ?? 0
const closeOffset = offsets[index] ?? 0
if (opener.text !== '_') continue
if (!isWordCharacter(scan.charAt(openOffset - 1)) && !isWordCharacter(scan.charAt(closeOffset + 1))) continue
opener.text = '*'
segment.text = '*'
}
return resolved
}
function escape(segments: readonly InlineSegment[], container: LineContainer, claims: Claims, fallback: RunFallback): AssembledLine {
const { output, placed, reading } = readBack(segments, container, claims)
return { ...guardLinkDefinition(output, placed, container), unspellableRuns: unspellableRuns(placed, reading, output, fallback) }
}
function readBack(segments: readonly InlineSegment[], container: LineContainer, claims: Claims): ReadBack {
const scan = segments.map((segment) => segment.text).join('')
// One byte a character, never an array element or Set entry: a line may run to hundreds of millions of characters.
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 claimed = escapeClaims(scan, escapings, container, claims)
let escapes: TextEscapes = { backticks: false, brackets: false }
// At most three passes: each pass that doesn't return turns on at least one more of the two escape kinds.
for (;;) {
const { output, placed } = backslashed(segments, scan, withTextEscapes(scan, escapings, claimed, escapes))
const { faulted, reading } = readLineBack(output, container, claims)
const shown = textOpeners(placed, output.length, reading)
// A text backtick in a directive's slot can open a code span past the slot's `]`: the directive faults before the reader records that span.
const opened = faulted && !shown.backticks && !shown.brackets ? { backticks: true, brackets: true } : shown
const next = { backticks: escapes.backticks || opened.backticks, brackets: escapes.brackets || opened.brackets }
if (next.backticks === escapes.backticks && next.brackets === escapes.brackets) return { output, placed, reading }
escapes = next
}
}
// spec/flavour.md, Canonical form.
function guardLinkDefinition(output: string, placed: readonly PlacedSegment[], container: LineContainer): { line: string; openingLinkAsDirective?: true } {
if (container !== 'paragraph' || !opensLinkDefinition(output)) return { line: output }
return opensWithLink(placed) ? { line: output, openingLinkAsDirective: true } : { line: `\\${output}` }
}
// Only the emitter's link opener is a `[` segment spelling nodes.
function opensWithLink([first]: readonly PlacedSegment[]): boolean {
return first?.segment.nodes !== undefined && first.segment.text === '['
}
function withTextEscapes(scan: string, escapings: Escapings, claimed: Uint8Array, { backticks, brackets }: TextEscapes): Uint8Array {
if (!backticks && !brackets) return claimed
const escaped = claimed.slice()
for (let index = 0; index < scan.length; index += 1) {
const character = scan.charAt(index)
if (((backticks && character === '`') || (brackets && character === '[')) && isEscapable(escapings(index))) escaped[index] = 1
}
return escaped
}
// A construct the reader opened at a text character is one the line does not mean.
function textOpeners(placed: readonly PlacedSegment[], length: number, reading: InlineReading): TextEscapes {
const text = new Uint8Array(length)
for (const [index, { segment, start }] of placed.entries()) {
if (isEscapable(segment.escaping)) text.fill(1, start, placed[index + 1]?.start ?? length)
}
return { backticks: reading.codeSpans.some((at) => text[at] === 1), brackets: reading.links.some((at) => text[at] === 1) }
}
function backslashed(segments: readonly InlineSegment[], scan: string, escaped: Uint8Array): { output: string; placed: PlacedSegment[] } {
// The output joins 4096 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 placed: PlacedSegment[] = []
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)
}
placed.push({ segment, start: start + escapes })
start += segment.text.length
}
return { output: chunks.join(''), placed }
}
function escapeClaims(scan: string, escapings: Escapings, container: LineContainer, claims: Claims): Uint8Array {
const walk: EscapeWalk = {
claims,
container,
escapings,
headingCloser: container === 'heading' ? closingHashes(scan) : undefined,
inline: inlineSource(scan),
}
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)
const run = runEscapes(walk, index, end, line)
if (!run.escapable) escaped.fill(1, index, run.end)
else for (let at = index; at < end; at += 1) if (isEscapable(escapings(at))) escaped[at] = 1
index = end
continue
}
const token = readInlineToken(walk.inline, index, claims)
const claimed = claimsCharacter(walk, index, token, line)
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
}
return escaped
}
// 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): RunEscapes {
const { escapings, inline: { text: scan } } = walk
if (runHolds(escapings, start, end, isSyntax) && runHolds(escapings, start, end, isEscapable)) return { escapable: true }
let index = start
// A backtick run escapes whole: a bare rest would be a shorter backtick string, which can close a span an earlier run opens.
while (index < end && ((index > start && scan.charAt(index) === '`') || claimsRunRest(walk, index, end, line))) index += 1
return { end: index, escapable: false }
}
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, inline }: EscapeWalk, index: number, end: number, line: ScanLine): boolean {
const scan = inline.text
const escaping = escapings(index)
if (escaping === 'bracketed-link-target') return readInlineToken(inline, index, claims).kind === 'code-span'
if (!isEscapable(escaping)) return false
if (claimsLineStart(line, index, container)) return true
const token = scan.charAt(index) === '`' ? readInlineToken(inline, index, claims) : delimiterRunToken(scan, index, end - index, claims)
return token.kind !== 'text'
}
function claimsCharacter({ container, escapings, headingCloser, inline: { text: scan } }: EscapeWalk, index: number, token: InlineToken, line: ScanLine): boolean {
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
switch (token.kind) {
case 'autolink':
case 'code-span':
case 'delimiter-run':
case 'directive':
case 'entity':
case 'escape':
case 'hard-break':
case 'html':
return true
case 'bracket':
return token.image ? isSyntax(escapings(index + 1)) : escaping === 'bracketed'
case 'bracket-close':
return escaping === 'bracketed'
case 'highlight':
return token.opens || token.closes
// A backslash before a line ending reads as a hard break, so no line ending escapes.
case 'line-ending':
case 'text': {
const character = scan.charAt(index)
if (character === '|') return container === 'table-cell'
return character === '{' && scan.charAt(index - 1) === ']' && isSyntax(escapings(index - 1))
}
}
}
// The runs the line does not read back as written: the first emphasis run to close and every run that falls with it, else every misflanked highlight run. One misread emphasis pair can re-pair the delimiters after it.
function unspellableRuns(placed: readonly PlacedSegment[], reading: InlineReading, output: string, fallback: RunFallback): MarkRun[] {
const { pairAt, pairs } = emphasisPairs(placed)
const first = pairs.findIndex((pair) => !readAsWritten(reading, placed, pair))
const firstPair = pairs[first]
if (firstPair === undefined) return misflankedHighlights(placed, output)
const merging = new Set<number>()
if (fallback.kind === 'drop') for (const { nodes } of pairs) if (!fallback.alone(nodes)) merging.add(nodes.first).add(nodes.last + 1)
const sweep: PairSweep = { fallback, fallen: new Uint8Array(pairs.length), first, merging, output, pairAt, pairs, placed }
sweep.fallen[first] = 1
// Each verdict reads the verdicts settled before it, so the sweep settles outward from the first pair.
for (let judged = first + 1; judged < pairs.length; judged += 1) if (fallsWithFirst(sweep, judged)) sweep.fallen[judged] = 1
for (let judged = first - 1; judged >= 0; judged -= 1) if (fallsWithFirst(sweep, judged)) sweep.fallen[judged] = 1
const runs = [firstPair.nodes]
for (const [index, pair] of pairs.entries()) if (index !== first && sweep.fallen[index] === 1) runs.push(pair.nodes)
return runs
}
// The pairs in closing order.
function emphasisPairs(placed: readonly PlacedSegment[]): { pairAt: Int32Array; pairs: EmphasisPair[] } {
const pairAt = new Int32Array(placed.length).fill(-1)
const pairs: EmphasisPair[] = []
const openers: { at: number; nodes: MarkRun }[] = []
for (const [at, { segment }] of placed.entries()) {
if (segment.emphasis === 'open') openers.push({ at, nodes: segment.nodes })
const opener = segment.emphasis === 'close' ? openers.pop() : undefined
if (opener === undefined) continue
pairAt[opener.at] = pairs.length
pairAt[at] = pairs.length
pairs.push({ closer: at, delimiter: segment.text, nodes: opener.nodes, opener: opener.at })
}
return { pairAt, pairs }
}
// No `==` in text can open or close, and highlights never nest, so a highlight run reads back where both its delimiters flank.
function misflankedHighlights(placed: readonly PlacedSegment[], output: string): MarkRun[] {
const misflanked: MarkRun[] = []
for (const { segment, start } of placed) {
if (segment.highlight !== undefined && !highlightFlanks(output, start, segment.highlight) && misflanked.at(-1) !== segment.nodes) misflanked.push(segment.nodes)
}
return misflanked
}
// A pair falls with the first where one run per pass would take it back too: one of its delimiters cannot flank however the unsettled pairs fall, and taking it back moves no delimiter another verdict reads.
function fallsWithFirst(sweep: PairSweep, judged: number): boolean {
if (!fallsCleanly(sweep, judged)) return false
const spellings = pairSpellings(sweep, judged, judged)
return spellings !== undefined && spellings.every((character) => !mayFlank(sweep, judged, character, 'opener') || !mayFlank(sweep, judged, character, 'closer'))
}
// A carry splits every pair holding it, so a carried pair holds no pair that may fall first; a dropped pair keeps its leaves where no run merges at its ends.
function fallsCleanly({ fallback, merging, pairs }: PairSweep, judged: number): boolean {
const pair = pairs[judged]
if (pair === undefined) return false
return fallback.kind === 'drop' ? !mergesAt(merging, pair) : !holdsEarlier(pairs, judged, pair)
}
function mergesAt(merging: ReadonlySet<number>, { nodes }: EmphasisPair): boolean {
return merging.has(nodes.first) || merging.has(nodes.last + 1)
}
// CommonMark judges a delimiter by the characters either side of its whole run of `character`.
function mayFlank(sweep: PairSweep, judged: number, character: string, end: 'closer' | 'opener'): boolean {
const at = sweep.pairs[judged]?.[end] ?? -1
const before = runEdge(sweep, judged, at, -1, character)
const after = runEdge(sweep, judged, at, 1, character)
if (before === undefined || after === undefined) return true
return before.some((left) =>
after.some((right) => {
const flags = delimiterFlags(character, left, right)
return end === 'opener' ? flags.canOpen : flags.canClose
}),
)
}
// The characters that may stand past the run of `character` holding the delimiter at `from`, walking by `step`; `undefined` where a pair beside may move.
function runEdge(sweep: PairSweep, judged: number, from: number, step: -1 | 1, character: string): string[] | undefined {
const found: string[] = []
// A backslash escaping the next text character stands past the last delimiter walked, so the walk tracks where that delimiter ends.
let edge = delimiterEdge(sweep.placed, from, step)
for (let at = from + step; ; at += step) {
const segment = sweep.placed[at]?.segment
const other = sweep.pairAt[at] ?? -1
if (segment === undefined || (other === -1 && segment.text !== '')) break
if (other === -1) continue
const standing = pairStanding(sweep, judged, other)
const spellings = standing === 'gone' ? [] : pairSpellings(sweep, judged, other)
if (standing === undefined || spellings === undefined) return undefined
for (const spelling of spellings) if (spelling !== character) found.push(spelling)
if (standing === 'there' && !spellings.includes(character)) return found
edge = delimiterEdge(sweep.placed, at, step)
}
found.push(sweep.output.charAt(step === 1 ? edge : edge - 1))
return found
}
function delimiterEdge(placed: readonly PlacedSegment[], at: number, step: -1 | 1): number {
const delimiter = placed[at]
if (delimiter === undefined) return 0
return step === 1 ? delimiter.start + delimiter.segment.text.length : delimiter.start
}
// How the other pair stands when the judged one falls. Past the first, every pair closing after the judged one still stands, since that one cannot read as written; `undefined` where a carry or a merge may move the other's delimiters.
function pairStanding({ fallback, fallen, first, merging, pairs }: PairSweep, judged: number, other: number): Standing | undefined {
const pair = pairs[other]
if (other === judged) return 'there'
if (pair === undefined || mergesAt(merging, pair)) return undefined
if (other > judged && judged > first) return 'there'
if (fallback.kind === 'carry') return undefined
return fallen[other] === 1 ? 'gone' : 'either'
}
// The characters the other pair may be written with when the judged one falls: emphasis takes `_`, turning `*` beside a word character.
function pairSpellings(sweep: PairSweep, judged: number, other: number): string[] | undefined {
const pair = sweep.pairs[other]
if (pair === undefined) return undefined
if (pair.delimiter.length > 1) return [pair.delimiter.charAt(0)]
// The carry's punctuation stands where the split pair meets it.
if (sweep.fallback.kind === 'carry' && other !== judged && holdsEarlier(sweep.pairs, judged, pair)) return ['*', '_']
const before = wordBeside(sweep, judged, pair.opener, -1)
const after = wordBeside(sweep, judged, pair.closer, 1)
if (before === undefined || after === undefined) return undefined
const spellings: string[] = []
if (before.includes(true) || after.includes(true)) spellings.push('*')
if (before.includes(false) && after.includes(false)) spellings.push('_')
return spellings
}
// Whether a word character may stand beside the delimiter at `from` as resolveEmphasis reads the line, for each way the unsettled pairs fall.
function wordBeside(sweep: PairSweep, judged: number, from: number, step: -1 | 1): boolean[] | undefined {
const found: boolean[] = []
for (let at = from + step; ; at += step) {
const segment = sweep.placed[at]?.segment
const other = sweep.pairAt[at] ?? -1
if (segment !== undefined && segment.text === '') continue
if (segment === undefined || other === -1) {
found.push(segment !== undefined && isWordCharacter(segment.text.charAt(step === 1 ? 0 : segment.text.length - 1)))
return found
}
const standing = pairStanding(sweep, judged, other)
if (standing === undefined) return undefined
if (standing === 'gone') continue
found.push(false)
if (standing === 'there') return found
}
}
// Pairs close in order, so a pair holding any pair that closes before the judged one holds the one just before it.
function holdsEarlier(pairs: readonly EmphasisPair[], judged: number, pair: EmphasisPair): boolean {
const earlier = pairs[judged - 1]
return earlier !== undefined && pair.nodes.first <= earlier.nodes.first && earlier.nodes.last <= pair.nodes.last
}
function highlightFlanks(output: string, start: number, role: DelimiterRole): boolean {
const flanking = highlightFlanking(output, start)
return role === 'open' ? flanking.opens : flanking.closes
}
function readAsWritten(reading: InlineReading, placed: readonly PlacedSegment[], pair: EmphasisPair): boolean {
const read = reading.emphasis.get(placed[pair.opener]?.start ?? -1)
return read !== undefined && read.closer === placed[pair.closer]?.start && read.width === pair.delimiter.length
}
function isEscapable(escaping: InlineEscaping | undefined): boolean {
return escaping === 'backslash' || escaping === 'bracketed'
}
export function isSyntax(escaping: InlineEscaping | undefined): boolean {
return escaping === 'none' || escaping === 'bracketed-link-target'
}
// A hard break is the one spelling that puts a delimiter row under a row of its own, so only a later line claims.
function claimsLineStart(line: ScanLine, index: number, container: LineContainer): boolean {
if (container !== 'paragraph') return false
if (index === line.start && line.position === 'later' && isBareDelimiterRow(line.text)) return true
return escapesLineClaim(line.text, index - line.start, line.position)
}
// The start of the closing sequence an ATX heading would read at the text's end.
function closingHashes(scan: string): number | undefined {
let start = scan.length
while (scan.charAt(start - 1) === '#') start -= 1
if (start === scan.length) return undefined
return start === 0 || /[ \t]/.test(scan.charAt(start - 1)) ? start : undefined
}
function scanLine(scan: string, start: number): ScanLine {
const end = scan.indexOf('\n', start)
return { position: start === 0 ? 'first' : 'later', start, text: scan.slice(start, end === -1 ? undefined : end) }
}