264 lines
7.7 KiB
JavaScript
264 lines
7.7 KiB
JavaScript
/**
|
||
* @typedef {import('micromark-util-types').Code} Code
|
||
* @typedef {import('micromark-util-types').Construct} Construct
|
||
* @typedef {import('micromark-util-types').Event} Event
|
||
* @typedef {import('micromark-util-types').Point} Point
|
||
* @typedef {import('micromark-util-types').Resolver} Resolver
|
||
* @typedef {import('micromark-util-types').State} State
|
||
* @typedef {import('micromark-util-types').Token} Token
|
||
* @typedef {import('micromark-util-types').TokenizeContext} TokenizeContext
|
||
* @typedef {import('micromark-util-types').Tokenizer} Tokenizer
|
||
*/
|
||
|
||
import {push, splice} from 'micromark-util-chunked'
|
||
import {classifyCharacter} from 'micromark-util-classify-character'
|
||
import {resolveAll} from 'micromark-util-resolve-all'
|
||
/** @type {Construct} */
|
||
export const attention = {
|
||
name: 'attention',
|
||
tokenize: tokenizeAttention,
|
||
resolveAll: resolveAllAttention
|
||
}
|
||
|
||
/**
|
||
* Take all events and resolve attention to emphasis or strong.
|
||
*
|
||
* @type {Resolver}
|
||
*/
|
||
// eslint-disable-next-line complexity
|
||
function resolveAllAttention(events, context) {
|
||
let index = -1
|
||
/** @type {number} */
|
||
let open
|
||
/** @type {Token} */
|
||
let group
|
||
/** @type {Token} */
|
||
let text
|
||
/** @type {Token} */
|
||
let openingSequence
|
||
/** @type {Token} */
|
||
let closingSequence
|
||
/** @type {number} */
|
||
let use
|
||
/** @type {Array<Event>} */
|
||
let nextEvents
|
||
/** @type {number} */
|
||
let offset
|
||
|
||
// Walk through all events.
|
||
//
|
||
// Note: performance of this is fine on an mb of normal markdown, but it’s
|
||
// a bottleneck for malicious stuff.
|
||
while (++index < events.length) {
|
||
// Find a token that can close.
|
||
if (
|
||
events[index][0] === 'enter' &&
|
||
events[index][1].type === 'attentionSequence' &&
|
||
events[index][1]._close
|
||
) {
|
||
open = index
|
||
|
||
// Now walk back to find an opener.
|
||
while (open--) {
|
||
// Find a token that can open the closer.
|
||
if (
|
||
events[open][0] === 'exit' &&
|
||
events[open][1].type === 'attentionSequence' &&
|
||
events[open][1]._open &&
|
||
// If the markers are the same:
|
||
context.sliceSerialize(events[open][1]).charCodeAt(0) ===
|
||
context.sliceSerialize(events[index][1]).charCodeAt(0)
|
||
) {
|
||
// If the opening can close or the closing can open,
|
||
// and the close size *is not* a multiple of three,
|
||
// but the sum of the opening and closing size *is* multiple of three,
|
||
// then don’t match.
|
||
if (
|
||
(events[open][1]._close || events[index][1]._open) &&
|
||
(events[index][1].end.offset - events[index][1].start.offset) % 3 &&
|
||
!(
|
||
(events[open][1].end.offset -
|
||
events[open][1].start.offset +
|
||
events[index][1].end.offset -
|
||
events[index][1].start.offset) %
|
||
3
|
||
)
|
||
) {
|
||
continue
|
||
}
|
||
|
||
// Number of markers to use from the sequence.
|
||
use =
|
||
events[open][1].end.offset - events[open][1].start.offset > 1 &&
|
||
events[index][1].end.offset - events[index][1].start.offset > 1
|
||
? 2
|
||
: 1
|
||
const start = Object.assign({}, events[open][1].end)
|
||
const end = Object.assign({}, events[index][1].start)
|
||
movePoint(start, -use)
|
||
movePoint(end, use)
|
||
openingSequence = {
|
||
type: use > 1 ? 'strongSequence' : 'emphasisSequence',
|
||
start,
|
||
end: Object.assign({}, events[open][1].end)
|
||
}
|
||
closingSequence = {
|
||
type: use > 1 ? 'strongSequence' : 'emphasisSequence',
|
||
start: Object.assign({}, events[index][1].start),
|
||
end
|
||
}
|
||
text = {
|
||
type: use > 1 ? 'strongText' : 'emphasisText',
|
||
start: Object.assign({}, events[open][1].end),
|
||
end: Object.assign({}, events[index][1].start)
|
||
}
|
||
group = {
|
||
type: use > 1 ? 'strong' : 'emphasis',
|
||
start: Object.assign({}, openingSequence.start),
|
||
end: Object.assign({}, closingSequence.end)
|
||
}
|
||
events[open][1].end = Object.assign({}, openingSequence.start)
|
||
events[index][1].start = Object.assign({}, closingSequence.end)
|
||
nextEvents = []
|
||
|
||
// If there are more markers in the opening, add them before.
|
||
if (events[open][1].end.offset - events[open][1].start.offset) {
|
||
nextEvents = push(nextEvents, [
|
||
['enter', events[open][1], context],
|
||
['exit', events[open][1], context]
|
||
])
|
||
}
|
||
|
||
// Opening.
|
||
nextEvents = push(nextEvents, [
|
||
['enter', group, context],
|
||
['enter', openingSequence, context],
|
||
['exit', openingSequence, context],
|
||
['enter', text, context]
|
||
])
|
||
|
||
// Always populated by defaults.
|
||
|
||
// Between.
|
||
nextEvents = push(
|
||
nextEvents,
|
||
resolveAll(
|
||
context.parser.constructs.insideSpan.null,
|
||
events.slice(open + 1, index),
|
||
context
|
||
)
|
||
)
|
||
|
||
// Closing.
|
||
nextEvents = push(nextEvents, [
|
||
['exit', text, context],
|
||
['enter', closingSequence, context],
|
||
['exit', closingSequence, context],
|
||
['exit', group, context]
|
||
])
|
||
|
||
// If there are more markers in the closing, add them after.
|
||
if (events[index][1].end.offset - events[index][1].start.offset) {
|
||
offset = 2
|
||
nextEvents = push(nextEvents, [
|
||
['enter', events[index][1], context],
|
||
['exit', events[index][1], context]
|
||
])
|
||
} else {
|
||
offset = 0
|
||
}
|
||
splice(events, open - 1, index - open + 3, nextEvents)
|
||
index = open + nextEvents.length - offset - 2
|
||
break
|
||
}
|
||
}
|
||
}
|
||
}
|
||
|
||
// Remove remaining sequences.
|
||
index = -1
|
||
while (++index < events.length) {
|
||
if (events[index][1].type === 'attentionSequence') {
|
||
events[index][1].type = 'data'
|
||
}
|
||
}
|
||
return events
|
||
}
|
||
|
||
/**
|
||
* @this {TokenizeContext}
|
||
* @type {Tokenizer}
|
||
*/
|
||
function tokenizeAttention(effects, ok) {
|
||
const attentionMarkers = this.parser.constructs.attentionMarkers.null
|
||
const previous = this.previous
|
||
const before = classifyCharacter(previous)
|
||
|
||
/** @type {NonNullable<Code>} */
|
||
let marker
|
||
return start
|
||
|
||
/**
|
||
* Before a sequence.
|
||
*
|
||
* ```markdown
|
||
* > | **
|
||
* ^
|
||
* ```
|
||
*
|
||
* @type {State}
|
||
*/
|
||
function start(code) {
|
||
marker = code
|
||
effects.enter('attentionSequence')
|
||
return inside(code)
|
||
}
|
||
|
||
/**
|
||
* In a sequence.
|
||
*
|
||
* ```markdown
|
||
* > | **
|
||
* ^^
|
||
* ```
|
||
*
|
||
* @type {State}
|
||
*/
|
||
function inside(code) {
|
||
if (code === marker) {
|
||
effects.consume(code)
|
||
return inside
|
||
}
|
||
const token = effects.exit('attentionSequence')
|
||
|
||
// To do: next major: move this to resolver, just like `markdown-rs`.
|
||
const after = classifyCharacter(code)
|
||
|
||
// Always populated by defaults.
|
||
|
||
const open =
|
||
!after || (after === 2 && before) || attentionMarkers.includes(code)
|
||
const close =
|
||
!before || (before === 2 && after) || attentionMarkers.includes(previous)
|
||
token._open = Boolean(marker === 42 ? open : open && (before || !close))
|
||
token._close = Boolean(marker === 42 ? close : close && (after || !open))
|
||
return ok(code)
|
||
}
|
||
}
|
||
|
||
/**
|
||
* Move a point a bit.
|
||
*
|
||
* Note: `move` only works inside lines! It’s not possible to move past other
|
||
* chunks (replacement characters, tabs, or line endings).
|
||
*
|
||
* @param {Point} point
|
||
* @param {number} offset
|
||
* @returns {undefined}
|
||
*/
|
||
function movePoint(point, offset) {
|
||
point.column += offset
|
||
point.offset += offset
|
||
point._bufferIndex += offset
|
||
}
|