dan/node-html-parser
1
0
Fork 0
mirror of https://github.com/danbulant/node-html-parser synced 2026-05-19 12:29:07 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

🐛 fix: Issue #17

Browse source
This commit is contained in:
taoqf 2020-02-02 13:28:04 +08:00
parent e9560b2d0e
commit 5ae3604472
4 changed files with 126 additions and 32 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

10
package.json
View file

@ -6,10 +6,11 @@
"types": "dist/index.d.ts",
"scripts": {
"test": "mocha",
"lint": "eslint ./src/*.ts",
"clean": "del-cli ./dist/",
"ts:cjs": "tsc -m commonjs",
"ts:umd": "tsc -t es5 -m umd -d false --outDir ./dist/umd/",
"build": "npm run clean && npm run ts:cjs && npm run ts:umd",
"build": "npm run lint && npm run clean && npm run ts:cjs && npm run ts:umd",
"dev": "tsc -w",
"pretest": "tsc -m commonjs"
},
@ -25,15 +26,20 @@
"he": "1.1.1"
},
"devDependencies": {
"@typescript-eslint/eslint-plugin": "latest",
"@typescript-eslint/eslint-plugin-tslint": "latest",
"@typescript-eslint/parser": "latest",
"@types/entities": "latest",
"@types/he": "latest",
"@types/node": "latest",
"blanket": "latest",
"del-cli": "latest",
"eslint": "latest",
"mocha": "latest",
"should": "latest",
"spec": "latest",
"travis-cov": "latest",
"tslint": "latest",
"typescript": "next"
},
"config": {
@ -58,4 +64,4 @@
"url": "https://github.com/taoqf/node-fast-html-parser/issues"
},
"homepage": "https://github.com/taoqf/node-fast-html-parser"
}
}

76
src/index.ts
View file

@ -1,4 +1,4 @@
import { decode, encode } from 'he';
import { decode } from 'he';
export enum NodeType {
ELEMENT_NODE = 1,
@ -481,7 +481,8 @@ export class HTMLElement extends Node {
this._attrs = {};
const attrs = this.rawAttributes;
for (const key in attrs) {
this._attrs[key] = decode(attrs[key]);
const val = attrs[key] || '';
this._attrs[key] = decode(val.replace(/^['"]/, '').replace(/['"]$/, ''));
}
return this._attrs;
}
@ -495,10 +496,10 @@ export class HTMLElement extends Node {
return this._rawAttrs;
const attrs = {} as RawAttributes;
if (this.rawAttrs) {
const re = /\b([a-z][a-z0-9\-]*)(?:\s*=\s*(?:"([^"]*)"|'([^']*)'|(\S+)))?/ig;
const re = /\b([a-z][a-z0-9\-]*)(?:\s*=\s*("(?:[^"]*)"|'(?:[^']*)'|(?:\S+)))?/ig;
let match: RegExpExecArray;
while (match = re.exec(this.rawAttrs)) {
attrs[match[1]] = match[2] || match[3] || match[4] || "";
attrs[match[1]] = match[2] || null;
}
}
this._rawAttrs = attrs;
@ -508,20 +509,33 @@ export class HTMLElement extends Node {
/**
* Set an attribute value to the HTMLElement
* @param {string} key The attribute name
* @param {string} value The value to set, or null / undefined to remove an attribute
* @param {string|number} value The value to set, or null / undefined to remove an attribute
*/
setAttribute(key: string, value: string) {
//Update the attributes map
const attrs = this.attributes;
if(value===undefined || value===null) delete attrs[key];
else attrs[key] = value+'';
//Update the raw attributes
if(this._rawAttrs) {
if(value===undefined || value===null) delete this._rawAttrs[key];
else this._rawAttrs[key] = encode(value+'');
setAttribute(key: string, value: string | number) {
// Update the this.attributes
if (this._attrs) {
delete this._attrs;
}
//Update rawString
this.rawAttrs = Object.keys(attrs).map(attr => attr+'='+encode(attrs[attr])).join(' ');
const attrs = this.rawAttributes; // ref this._rawAttrs
if (value === undefined || value === null) {
delete attrs[key];
} else {
attrs[key] = JSON.stringify(value);
// if (typeof value === 'string') {
// attrs[key] = JSON.stringify(encode(value));//??? should we encode value here?
// } else {
// attrs[key] = JSON.stringify(value);
// }
}
// Update rawString
this.rawAttrs = Object.keys(attrs).map((name) => {
const val = attrs[name];
if (val === undefined || val === null) {
return name;
} else {
return name + '=' + val;
}
}).join(' ');
}
/**
@ -529,18 +543,28 @@ export class HTMLElement extends Node {
* @param {Attributes} attributes the new attribute set
*/
setAttributes(attributes: Attributes) {
//Update the attributes map
if(this.attributes) {
Object.keys(this.attributes).forEach(key => delete this.attributes[key]);
Object.keys(attributes).forEach(key => this.attributes[key] = attributes[key]+'');
// Update the this.attributes
if (this._attrs) {
delete this._attrs;
}
//Update the raw attributes map
if(this.rawAttributes) {
Object.keys(this.rawAttributes).forEach(key => delete this.rawAttributes[key]);
Object.keys(attributes).forEach(key => this.rawAttributes[key] = encode(attributes[key]+''));
// Update the raw attributes map
if (this._rawAttrs) {
delete this._rawAttrs;
}
//Update rawString
this.rawAttrs = Object.keys(attributes).map(attr => attr+'='+encode(attributes[attr]+'')).join(' ');
// Update rawString
this.rawAttrs = Object.keys(attributes).map((name) => {
const val = attributes[name];
if (val === undefined || val === null) {
return name;
} else {
return name + '=' + JSON.stringify(val);
// if (typeof val === 'string') {
// return name + '=' + JSON.stringify(encode(val)); //??? should we encode value here?
// } else {
// return name + '=' + JSON.stringify(val);
// }
}
}).join(' ');
}
}

49
t.js Normal file
View file

@ -0,0 +1,49 @@
function rawAttributes(rawAttrs) {
const attrs = {};
if (rawAttrs) {
// const re = /\b([a-z][a-z0-9\-]*)(?:\s*=\s*(?:("[^"]*")|('[^']*')|(\S+)))?/ig;
const re = /\b([a-z][a-z0-9\-]*)(?:\s*=\s*("(?:[^"]*)"|'(?:[^']*)'|(?:\S+)))?/ig;
let match;
console.debug('0000', rawAttrs);
while (match = re.exec(rawAttrs)) {
console.debug('1111', match[1]);
const v = match[2] || '';
console.debug('2222', v.replace(/^['"]/, '').replace(/['"]$/, ''));
attrs[match[1]] = v.replace(/^['"]/, '').replace(/['"]$/, '');
}
}
return attrs;
}
function attr2str(attrs) {
return Object.keys(attrs).map((name) => {
const val = attrs[name];
if (val === undefined || val === null) {
return name;
} else {
return name + '=' + val
}
}).join(' ')
}
function main() {
let r;
// r = rawAttributes('a="1"');
// r = rawAttributes('a=\'1\'');
// r = rawAttributes('a=');
// r = rawAttributes('a');
// r = rawAttributes('a=1');
// r = rawAttributes('a=aa b="bb" c= \'cc\' d="\'dd\'" e=e\'e\"e f');
r = attr2str({
a: 'aa',
b: '"bb"',
c: "'cc'",
d: "'dd'",
e: `e'e"e`,
f: null
});
console.debug(r);
}
main();

23
test/html.js
View file

@ -302,8 +302,8 @@ describe('HTML Parser', function () {
var root = parseHTML('<p a=12 data-id="!$$&amp;" yAz=\'1\'></p>');
root.firstChild.rawAttributes.should.eql({
'a': '12',
'data-id': '!$$&amp;',
'yAz': '1'
'data-id': '"!$$&amp;"',
'yAz': '\'1\''
});
});
});
@ -348,16 +348,31 @@ describe('HTML Parser', function () {
});
root.firstChild.toString().should.eql('<p a=12></p>');
});
it('should keep quotes arount value', function () {
var root = parseHTML('<p a="12"></p>');
root.firstChild.setAttribute('b', 13);
root.firstChild.setAttribute('c', '2');
root.firstChild.attributes.should.eql({
'a': '12',
'b': '13',
'c': '2'
});
root.firstChild.toString().should.eql('<p a="12" b=13 c="2"></p>');
});
});
describe('#setAttributes', function () {
it('should return attributes of the element', function () {
var root = parseHTML('<p a=12 data-id="!$$&amp;" yAz=\'1\' class="" disabled></p>');
root.firstChild.setAttributes({c: 12});
root.firstChild.setAttributes({
c: 12,
d: '&&<>foo'
});
root.firstChild.attributes.should.eql({
'c': '12',
d: '&&<>foo'
});
root.firstChild.toString().should.eql('<p c=12></p>');
root.firstChild.toString().should.eql('<p c=12 d="&#x26;&#x26;&#x3C;&#x3E;foo"></p>');
});
});