fix(compiler): support full range of entity decoding in browser builds

BREAKING CHANGE: compiler options have been adjusted.
    - new option `decodeEntities` is added.
    - `namedCharacterReferences` option has been removed.
    - `maxCRNameLength` option has been rmeoved.
This commit is contained in:
Evan You
2020-04-08 18:51:25 -04:00
parent 8c17535a47
commit 1f6e72b110
11 changed files with 245 additions and 1809 deletions

View File

@@ -5,12 +5,10 @@ import {
TextNode,
ErrorCodes,
ElementTypes,
InterpolationNode
InterpolationNode,
AttributeNode
} from '@vue/compiler-core'
import {
parserOptionsStandard as parserOptions,
DOMNamespaces
} from '../src/parserOptionsStandard'
import { parserOptions, DOMNamespaces } from '../src/parserOptions'
describe('DOM parser', () => {
describe('Text', () => {
@@ -170,6 +168,77 @@ describe('DOM parser', () => {
content: `foo${nbsp}${nbsp}bar`
})
})
// https://html.spec.whatwg.org/multipage/parsing.html#named-character-reference-state
test('HTML entities compatibility in text', () => {
const ast = parse('&ampersand;', parserOptions)
const text = ast.children[0] as TextNode
expect(text).toStrictEqual({
type: NodeTypes.TEXT,
content: '&ersand;',
loc: {
start: { offset: 0, line: 1, column: 1 },
end: { offset: 11, line: 1, column: 12 },
source: '&ampersand;'
}
})
})
// https://html.spec.whatwg.org/multipage/parsing.html#named-character-reference-state
test('HTML entities compatibility in attribute', () => {
const ast = parse(
'<div a="&ampersand;" b="&amp;ersand;" c="&amp!"></div>',
parserOptions
)
const element = ast.children[0] as ElementNode
const text1 = (element.props[0] as AttributeNode).value
const text2 = (element.props[1] as AttributeNode).value
const text3 = (element.props[2] as AttributeNode).value
expect(text1).toStrictEqual({
type: NodeTypes.TEXT,
content: '&ampersand;',
loc: {
start: { offset: 7, line: 1, column: 8 },
end: { offset: 20, line: 1, column: 21 },
source: '"&ampersand;"'
}
})
expect(text2).toStrictEqual({
type: NodeTypes.TEXT,
content: '&ersand;',
loc: {
start: { offset: 23, line: 1, column: 24 },
end: { offset: 37, line: 1, column: 38 },
source: '"&amp;ersand;"'
}
})
expect(text3).toStrictEqual({
type: NodeTypes.TEXT,
content: '&!',
loc: {
start: { offset: 40, line: 1, column: 41 },
end: { offset: 47, line: 1, column: 48 },
source: '"&amp!"'
}
})
})
test('Some control character reference should be replaced.', () => {
const ast = parse('&#x86;', parserOptions)
const text = ast.children[0] as TextNode
expect(text).toStrictEqual({
type: NodeTypes.TEXT,
content: '†',
loc: {
start: { offset: 0, line: 1, column: 1 },
end: { offset: 6, line: 1, column: 7 },
source: '&#x86;'
}
})
})
})
describe('Interpolation', () => {

View File

@@ -0,0 +1,133 @@
import { ParserOptions } from '@vue/compiler-core'
import namedCharacterReferences from './namedChars.json'
// lazy compute this to make this file tree-shakable for browser
let maxCRNameLength: number
export const decodeHtml: ParserOptions['decodeEntities'] = (
rawText,
asAttr
) => {
let offset = 0
const end = rawText.length
let decodedText = ''
function advance(length: number) {
offset += length
rawText = rawText.slice(length)
}
while (offset < end) {
const head = /&(?:#x?)?/i.exec(rawText)
if (!head || offset + head.index >= end) {
const remaining = end - offset
decodedText += rawText.slice(0, remaining)
advance(remaining)
break
}
// Advance to the "&".
decodedText += rawText.slice(0, head.index)
advance(head.index)
if (head[0] === '&') {
// Named character reference.
let name = ''
let value: string | undefined = undefined
if (/[0-9a-z]/i.test(rawText[1])) {
if (!maxCRNameLength) {
maxCRNameLength = Object.keys(namedCharacterReferences).reduce(
(max, name) => Math.max(max, name.length),
0
)
}
for (let length = maxCRNameLength; !value && length > 0; --length) {
name = rawText.substr(1, length)
value = (namedCharacterReferences as Record<string, string>)[name]
}
if (value) {
const semi = name.endsWith(';')
if (
asAttr &&
!semi &&
/[=a-z0-9]/i.test(rawText[name.length + 1] || '')
) {
decodedText += '&' + name
advance(1 + name.length)
} else {
decodedText += value
advance(1 + name.length)
}
} else {
decodedText += '&' + name
advance(1 + name.length)
}
} else {
decodedText += '&'
advance(1)
}
} else {
// Numeric character reference.
const hex = head[0] === '&#x'
const pattern = hex ? /^&#x([0-9a-f]+);?/i : /^&#([0-9]+);?/
const body = pattern.exec(rawText)
if (!body) {
decodedText += head[0]
advance(head[0].length)
} else {
// https://html.spec.whatwg.org/multipage/parsing.html#numeric-character-reference-end-state
let cp = Number.parseInt(body[1], hex ? 16 : 10)
if (cp === 0) {
cp = 0xfffd
} else if (cp > 0x10ffff) {
cp = 0xfffd
} else if (cp >= 0xd800 && cp <= 0xdfff) {
cp = 0xfffd
} else if ((cp >= 0xfdd0 && cp <= 0xfdef) || (cp & 0xfffe) === 0xfffe) {
// noop
} else if (
(cp >= 0x01 && cp <= 0x08) ||
cp === 0x0b ||
(cp >= 0x0d && cp <= 0x1f) ||
(cp >= 0x7f && cp <= 0x9f)
) {
cp = CCR_REPLACEMENTS[cp] || cp
}
decodedText += String.fromCodePoint(cp)
advance(body[0].length)
}
}
}
return decodedText
}
// https://html.spec.whatwg.org/multipage/parsing.html#numeric-character-reference-end-state
const CCR_REPLACEMENTS: { [key: number]: number | undefined } = {
0x80: 0x20ac,
0x82: 0x201a,
0x83: 0x0192,
0x84: 0x201e,
0x85: 0x2026,
0x86: 0x2020,
0x87: 0x2021,
0x88: 0x02c6,
0x89: 0x2030,
0x8a: 0x0160,
0x8b: 0x2039,
0x8c: 0x0152,
0x8e: 0x017d,
0x91: 0x2018,
0x92: 0x2019,
0x93: 0x201c,
0x94: 0x201d,
0x95: 0x2022,
0x96: 0x2013,
0x97: 0x2014,
0x98: 0x02dc,
0x99: 0x2122,
0x9a: 0x0161,
0x9b: 0x203a,
0x9c: 0x0153,
0x9e: 0x017e,
0x9f: 0x0178
}

View File

@@ -0,0 +1,6 @@
let decoder: HTMLDivElement
export function decodeHtmlBrowser(raw: string): string {
;(decoder || (decoder = document.createElement('div'))).innerHTML = raw
return decoder.textContent as string
}

View File

@@ -9,8 +9,7 @@ import {
NodeTransform,
DirectiveTransform
} from '@vue/compiler-core'
import { parserOptionsMinimal } from './parserOptionsMinimal'
import { parserOptionsStandard } from './parserOptionsStandard'
import { parserOptions } from './parserOptions'
import { transformStyle } from './transforms/transformStyle'
import { transformVHtml } from './transforms/vHtml'
import { transformVText } from './transforms/vText'
@@ -20,9 +19,7 @@ import { transformShow } from './transforms/vShow'
import { warnTransitionChildren } from './transforms/warnTransitionChildren'
import { stringifyStatic } from './transforms/stringifyStatic'
export const parserOptions = __BROWSER__
? parserOptionsMinimal
: parserOptionsStandard
export { parserOptions }
export const DOMNodeTransforms: NodeTransform[] = [
transformStyle,

View File

@@ -8,6 +8,8 @@ import {
} from '@vue/compiler-core'
import { makeMap, isVoidTag, isHTMLTag, isSVGTag } from '@vue/shared'
import { TRANSITION, TRANSITION_GROUP } from './runtimeHelpers'
import { decodeHtml } from './decodeHtml'
import { decodeHtmlBrowser } from './decodeHtmlBrowser'
const isRawTextContainer = /*#__PURE__*/ makeMap(
'style,iframe,script,noscript',
@@ -20,10 +22,11 @@ export const enum DOMNamespaces {
MATH_ML
}
export const parserOptionsMinimal: ParserOptions = {
export const parserOptions: ParserOptions = {
isVoidTag,
isNativeTag: tag => isHTMLTag(tag) || isSVGTag(tag),
isPreTag: tag => tag === 'pre',
decodeEntities: __BROWSER__ ? decodeHtmlBrowser : decodeHtml,
isBuiltInComponent: (tag: string): symbol | undefined => {
if (isBuiltInType(tag, `Transition`)) {

View File

@@ -1,17 +0,0 @@
import { ParserOptions } from '@vue/compiler-core'
import { parserOptionsMinimal } from './parserOptionsMinimal'
import namedCharacterReferences from './namedChars.json'
export { DOMNamespaces } from './parserOptionsMinimal'
export const parserOptionsStandard: ParserOptions = {
// extends the minimal options with more spec-compliant overrides
...parserOptionsMinimal,
// https://html.spec.whatwg.org/multipage/named-characters.html#named-character-references
namedCharacterReferences,
maxCRNameLength: /*#__PURE__*/ Object.keys(namedCharacterReferences).reduce(
(max, name) => Math.max(max, name.length),
0
)
}