node-html-parser 6.1.0 → 6.1.2

package/CHANGELOG.md

@@ -2,6 +2,20 @@
 
 All notable changes to this project will be documented in this file. See [standard-version](https://github.com/conventional-changelog/standard-version) for commit guidelines.
 
+### [6.1.2](https://github.com/taoqf/node-fast-html-parser/compare/v6.1.1...v6.1.2) (2022-11-14)
+
+
+### Bug Fixes
+
+* [#224](https://github.com/taoqf/node-fast-html-parser/issues/224) ([fc367fa](https://github.com/taoqf/node-fast-html-parser/commit/fc367fa294d72794a0dea49edbd986d527a6314b))
+
+### [6.1.1](https://github.com/taoqf/node-fast-html-parser/compare/v6.1.0...v6.1.1) (2022-09-24)
+
+
+### Bug Fixes
+
+* parse comments ([82b68ff](https://github.com/taoqf/node-fast-html-parser/commit/82b68ff9eb944e0c55ca2e0ea13fb714e2004803))
+
 ## [6.1.0](https://github.com/taoqf/node-fast-html-parser/compare/v6.0.0...v6.1.0) (2022-09-19)
 
 

package/dist/index.js

@@ -18,10 +18,7 @@ var parse_1 = __importDefault(require("./parse"));
 var valid_1 = __importDefault(require("./valid"));
 exports.valid = valid_1.default;
 function parse(data, options) {
-    if (options === void 0) { options = {
-        lowerCaseTagName: false,
-        comment: false
-    }; }
+    if (options === void 0) { options = {}; }
     return (0, parse_1.default)(data, options);
 }
 exports.default = parse;

package/dist/main.js

@@ -171,7 +171,7 @@ define("matcher", ["require", "exports", "nodes/type"], function (require, expor
     }
     function findOne(test, elems) {
         var elem = null;
-        for (var i = 0, l = elems.length; i < l && !elem; i++) {
+        for (var i = 0, l = elems === null || elems === void 0 ? void 0 : elems.length; i < l && !elem; i++) {
             var el = elems[i];
             if (test(el)) {
                 elem = el;
@@ -487,9 +487,10 @@ define("nodes/html", ["require", "exports", "css-select", "he", "back", "matcher
          *
          * @memberof HTMLElement
          */
-        function HTMLElement(tagName, keyAttrs, rawAttrs, parentNode, range, voidTag) {
+        function HTMLElement(tagName, keyAttrs, rawAttrs, parentNode, range, voidTag, _parseOptions) {
             if (rawAttrs === void 0) { rawAttrs = ''; }
             if (voidTag === void 0) { voidTag = new void_tag_1.default(); }
+            if (_parseOptions === void 0) { _parseOptions = {}; }
             var _this = _super.call(this, parentNode, range) || this;
             _this.rawAttrs = rawAttrs;
             _this.voidTag = voidTag;
@@ -501,6 +502,7 @@ define("nodes/html", ["require", "exports", "css-select", "he", "back", "matcher
             _this.rawAttrs = rawAttrs || '';
             _this.id = keyAttrs.id || '';
             _this.childNodes = [];
+            _this._parseOptions = _parseOptions;
             _this.classList = new DOMTokenList(keyAttrs.class ? keyAttrs.class.split(/\s+/) : [], function (classList) { return _this.setAttribute('class', classList.toString()); } // eslint-disable-line @typescript-eslint/no-unsafe-member-access, @typescript-eslint/no-unsafe-call
             );
             if (keyAttrs.id) {
@@ -682,8 +684,7 @@ define("nodes/html", ["require", "exports", "css-select", "he", "back", "matcher
                     .join('');
             },
             set: function (content) {
-                //const r = parse(content, global.options); // TODO global.options ?
-                var r = parse(content);
+                var r = parse(content, this._parseOptions);
                 var nodes = r.childNodes.length ? r.childNodes : [new text_1.default(content, this)];
                 resetParent(nodes, this);
                 resetParent(this.childNodes, null);
@@ -698,8 +699,9 @@ define("nodes/html", ["require", "exports", "css-select", "he", "back", "matcher
                 content = [content];
             }
             else if (typeof content == 'string') {
+                options = __assign(__assign({}, this._parseOptions), options);
                 var r = parse(content, options);
-                content = r.childNodes.length ? r.childNodes : [new text_1.default(content, this)];
+                content = r.childNodes.length ? r.childNodes : [new text_1.default(r.innerHTML, this)];
             }
             resetParent(this.childNodes, null);
             resetParent(content, this);
@@ -719,8 +721,7 @@ define("nodes/html", ["require", "exports", "css-select", "he", "back", "matcher
                     return [node];
                 }
                 else if (typeof node == 'string') {
-                    // const r = parse(content, global.options); // TODO global.options ?
-                    var r = parse(node);
+                    var r = parse(node, _this._parseOptions);
                     return r.childNodes.length ? r.childNodes : [new text_1.default(node, _this)];
                 }
                 return [];
@@ -1161,7 +1162,7 @@ define("nodes/html", ["require", "exports", "css-select", "he", "back", "matcher
             if (arguments.length < 2) {
                 throw new Error('2 arguments required');
             }
-            var p = parse(html);
+            var p = parse(html, this._parseOptions);
             if (where === 'afterend') {
                 var idx = this.parentNode.childNodes.findIndex(function (child) {
                     return child === _this;
@@ -1282,7 +1283,7 @@ define("nodes/html", ["require", "exports", "css-select", "he", "back", "matcher
          * Clone this Node
          */
         HTMLElement.prototype.clone = function () {
-            return parse(this.toString()).firstChild;
+            return parse(this.toString(), this._parseOptions).firstChild;
         };
         return HTMLElement;
     }(node_2.default));
@@ -1369,7 +1370,7 @@ define("nodes/html", ["require", "exports", "css-select", "he", "back", "matcher
      */
     function base_parse(data, options) {
         var _a, _b;
-        if (options === void 0) { options = { lowerCaseTagName: false, comment: false }; }
+        if (options === void 0) { options = {}; }
         var voidTag = new void_tag_1.default((_a = options === null || options === void 0 ? void 0 : options.voidTag) === null || _a === void 0 ? void 0 : _a.closingSlash, (_b = options === null || options === void 0 ? void 0 : options.voidTag) === null || _b === void 0 ? void 0 : _b.tags);
         var elements = options.blockTextElements || {
             script: true,
@@ -1387,7 +1388,7 @@ define("nodes/html", ["require", "exports", "css-select", "he", "back", "matcher
             return kBlockTextElements.some(function (it) { return it.test(tag); });
         }
         var createRange = function (startPos, endPos) { return [startPos - frameFlagOffset, endPos - frameFlagOffset]; };
-        var root = new HTMLElement(null, {}, '', null, [0, data.length], voidTag);
+        var root = new HTMLElement(null, {}, '', null, [0, data.length], voidTag, options);
         var currentParent = root;
         var stack = [root];
         var lastTextPos = -1;
@@ -1458,7 +1459,7 @@ define("nodes/html", ["require", "exports", "css-select", "he", "back", "matcher
                 var tagStartPos_1 = tagEndPos_1 - matchLength;
                 currentParent = currentParent.appendChild(
                 // Initialize range (end position updated later for closed tags)
-                new HTMLElement(tagName, attrs, attributes.slice(1), null, createRange(tagStartPos_1, tagEndPos_1), voidTag));
+                new HTMLElement(tagName, attrs, attributes.slice(1), null, createRange(tagStartPos_1, tagEndPos_1), voidTag, options));
                 stack.push(currentParent);
                 if (is_block_text_element(tagName)) {
                     // Find closing tag
@@ -1519,7 +1520,7 @@ define("nodes/html", ["require", "exports", "css-select", "he", "back", "matcher
      * Parse a chuck of HTML source.
      */
     function parse(data, options) {
-        if (options === void 0) { options = { lowerCaseTagName: false, comment: false }; }
+        if (options === void 0) { options = {}; }
         var stack = base_parse(data, options);
         var root = stack[0];
         var _loop_1 = function () {
@@ -1624,7 +1625,7 @@ define("valid", ["require", "exports", "nodes/html"], function (require, exports
      * Parse a chuck of HTML source.
      */
     function valid(data, options) {
-        if (options === void 0) { options = { lowerCaseTagName: false, comment: false }; }
+        if (options === void 0) { options = {}; }
         var stack = (0, html_2.base_parse)(data, options);
         return Boolean(stack.length === 1);
     }
@@ -1648,10 +1649,7 @@ define("index", ["require", "exports", "nodes/comment", "nodes/html", "nodes/nod
     exports.NodeType = type_5.default;
     exports.valid = valid_1.default;
     function parse(data, options) {
-        if (options === void 0) { options = {
-            lowerCaseTagName: false,
-            comment: false
-        }; }
+        if (options === void 0) { options = {}; }
         return (0, parse_1.default)(data, options);
     }
     exports.default = parse;

package/dist/matcher.js

@@ -63,7 +63,7 @@ function hasAttrib(elem, name) {
 }
 function findOne(test, elems) {
     var elem = null;
-    for (var i = 0, l = elems.length; i < l && !elem; i++) {
+    for (var i = 0, l = elems === null || elems === void 0 ? void 0 : elems.length; i < l && !elem; i++) {
         var el = elems[i];
         if (test(el)) {
             elem = el;

package/dist/nodes/html.d.ts

@@ -41,6 +41,7 @@ export default class HTMLElement extends Node {
     private voidTag;
     private _attrs;
     private _rawAttrs;
+    private _parseOptions;
     rawTagName: string;
     id: string;
     classList: DOMTokenList;
@@ -61,7 +62,7 @@ export default class HTMLElement extends Node {
      *
      * @memberof HTMLElement
      */
-    constructor(tagName: string, keyAttrs: KeyAttributes, rawAttrs: string, parentNode: HTMLElement | null, range: [number, number], voidTag?: VoidTag);
+    constructor(tagName: string, keyAttrs: KeyAttributes, rawAttrs: string, parentNode: HTMLElement | null, range: [number, number], voidTag?: VoidTag, _parseOptions?: Partial<Options>);
     /**
      * Remove Child element from childNodes array
      * @param {HTMLElement} node     node to remove
@@ -142,7 +143,7 @@ export default class HTMLElement extends Node {
      * traverses the Element and its parents (heading toward the document root) until it finds a node that matches the provided selector string. Will return itself or the matching ancestor. If no such element exists, it returns null.
      * @param selector a DOMString containing a selector list
      */
-    closest(selector: string): Node;
+    closest(selector: string): HTMLElement;
     /**
      * Append a child node to childNodes
      * @param  {Node} node node to append
@@ -201,8 +202,8 @@ export default class HTMLElement extends Node {
     clone(): Node;
 }
 export interface Options {
-    lowerCaseTagName: boolean;
-    comment: boolean;
+    lowerCaseTagName?: boolean;
+    comment?: boolean;
     /**
      * @see PR #215 for explanation
      */

package/dist/nodes/html.js

@@ -154,9 +154,10 @@ var HTMLElement = /** @class */ (function (_super) {
      *
      * @memberof HTMLElement
      */
-    function HTMLElement(tagName, keyAttrs, rawAttrs, parentNode, range, voidTag) {
+    function HTMLElement(tagName, keyAttrs, rawAttrs, parentNode, range, voidTag, _parseOptions) {
         if (rawAttrs === void 0) { rawAttrs = ''; }
         if (voidTag === void 0) { voidTag = new void_tag_1.default(); }
+        if (_parseOptions === void 0) { _parseOptions = {}; }
         var _this = _super.call(this, parentNode, range) || this;
         _this.rawAttrs = rawAttrs;
         _this.voidTag = voidTag;
@@ -168,6 +169,7 @@ var HTMLElement = /** @class */ (function (_super) {
         _this.rawAttrs = rawAttrs || '';
         _this.id = keyAttrs.id || '';
         _this.childNodes = [];
+        _this._parseOptions = _parseOptions;
         _this.classList = new DOMTokenList(keyAttrs.class ? keyAttrs.class.split(/\s+/) : [], function (classList) { return _this.setAttribute('class', classList.toString()); } // eslint-disable-line @typescript-eslint/no-unsafe-member-access, @typescript-eslint/no-unsafe-call
         );
         if (keyAttrs.id) {
@@ -349,8 +351,7 @@ var HTMLElement = /** @class */ (function (_super) {
                 .join('');
         },
         set: function (content) {
-            //const r = parse(content, global.options); // TODO global.options ?
-            var r = parse(content);
+            var r = parse(content, this._parseOptions);
             var nodes = r.childNodes.length ? r.childNodes : [new text_1.default(content, this)];
             resetParent(nodes, this);
             resetParent(this.childNodes, null);
@@ -365,8 +366,9 @@ var HTMLElement = /** @class */ (function (_super) {
             content = [content];
         }
         else if (typeof content == 'string') {
+            options = __assign(__assign({}, this._parseOptions), options);
             var r = parse(content, options);
-            content = r.childNodes.length ? r.childNodes : [new text_1.default(content, this)];
+            content = r.childNodes.length ? r.childNodes : [new text_1.default(r.innerHTML, this)];
         }
         resetParent(this.childNodes, null);
         resetParent(content, this);
@@ -386,8 +388,7 @@ var HTMLElement = /** @class */ (function (_super) {
                 return [node];
             }
             else if (typeof node == 'string') {
-                // const r = parse(content, global.options); // TODO global.options ?
-                var r = parse(node);
+                var r = parse(node, _this._parseOptions);
                 return r.childNodes.length ? r.childNodes : [new text_1.default(node, _this)];
             }
             return [];
@@ -828,7 +829,7 @@ var HTMLElement = /** @class */ (function (_super) {
         if (arguments.length < 2) {
             throw new Error('2 arguments required');
         }
-        var p = parse(html);
+        var p = parse(html, this._parseOptions);
         if (where === 'afterend') {
             var idx = this.parentNode.childNodes.findIndex(function (child) {
                 return child === _this;
@@ -949,7 +950,7 @@ var HTMLElement = /** @class */ (function (_super) {
      * Clone this Node
      */
     HTMLElement.prototype.clone = function () {
-        return parse(this.toString()).firstChild;
+        return parse(this.toString(), this._parseOptions).firstChild;
     };
     return HTMLElement;
 }(node_1.default));
@@ -1036,7 +1037,7 @@ var frameflag = 'documentfragmentcontainer';
  */
 function base_parse(data, options) {
     var _a, _b;
-    if (options === void 0) { options = { lowerCaseTagName: false, comment: false }; }
+    if (options === void 0) { options = {}; }
     var voidTag = new void_tag_1.default((_a = options === null || options === void 0 ? void 0 : options.voidTag) === null || _a === void 0 ? void 0 : _a.closingSlash, (_b = options === null || options === void 0 ? void 0 : options.voidTag) === null || _b === void 0 ? void 0 : _b.tags);
     var elements = options.blockTextElements || {
         script: true,
@@ -1054,7 +1055,7 @@ function base_parse(data, options) {
         return kBlockTextElements.some(function (it) { return it.test(tag); });
     }
     var createRange = function (startPos, endPos) { return [startPos - frameFlagOffset, endPos - frameFlagOffset]; };
-    var root = new HTMLElement(null, {}, '', null, [0, data.length], voidTag);
+    var root = new HTMLElement(null, {}, '', null, [0, data.length], voidTag, options);
     var currentParent = root;
     var stack = [root];
     var lastTextPos = -1;
@@ -1125,7 +1126,7 @@ function base_parse(data, options) {
             var tagStartPos_1 = tagEndPos_1 - matchLength;
             currentParent = currentParent.appendChild(
             // Initialize range (end position updated later for closed tags)
-            new HTMLElement(tagName, attrs, attributes.slice(1), null, createRange(tagStartPos_1, tagEndPos_1), voidTag));
+            new HTMLElement(tagName, attrs, attributes.slice(1), null, createRange(tagStartPos_1, tagEndPos_1), voidTag, options));
             stack.push(currentParent);
             if (is_block_text_element(tagName)) {
                 // Find closing tag
@@ -1186,7 +1187,7 @@ exports.base_parse = base_parse;
  * Parse a chuck of HTML source.
  */
 function parse(data, options) {
-    if (options === void 0) { options = { lowerCaseTagName: false, comment: false }; }
+    if (options === void 0) { options = {}; }
     var stack = base_parse(data, options);
     var root = stack[0];
     var _loop_1 = function () {

package/dist/valid.js

@@ -6,7 +6,7 @@ var html_1 = require("./nodes/html");
  * Parse a chuck of HTML source.
  */
 function valid(data, options) {
-    if (options === void 0) { options = { lowerCaseTagName: false, comment: false }; }
+    if (options === void 0) { options = {}; }
     var stack = (0, html_1.base_parse)(data, options);
     return Boolean(stack.length === 1);
 }

package/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "node-html-parser",
-	"version": "6.1.0",
+	"version": "6.1.2",
 	"description": "A very fast HTML parser, generating a simplified DOM, with basic element query support.",
 	"main": "dist/index.js",
 	"types": "dist/index.d.ts",
@@ -8,6 +8,7 @@
 		"compile": "tsc",
 		"build": "npm run lint && npm run clean && npm run compile:cjs && npm run compile:amd",
 		"compile:cjs": "tsc -m commonjs",
+		"watch": "npx tsc -m commonjs --watch --preserveWatchOutput",
 		"compile:amd": "tsc -t es5 -m amd -d false --outFile ./dist/main.js",
 		"lint": "eslint ./src/*.ts ./src/**/*.ts",
 		"---------------": "",
@@ -23,7 +24,8 @@
 		"test:target": "mocha --recursive \"./test/tests\"",
 		"test:ci": "cross-env TEST_TARGET=dist yarn run test:target",
 		"posttest": "yarn run benchmark",
-		"prepare": "cd test && yarn install"
+		"prepare": "cd test && yarn install",
+		"release": "standard-version && git push --follow-tags origin main"
 	},
 	"keywords": [
 		"html",
@@ -47,7 +49,7 @@
 		"registry": "https://registry.npmjs.org"
 	},
 	"dependencies": {
-		"css-select": "^4.2.1",
+		"css-select": "^5.1.0",
 		"he": "1.2.0"
 	},
 	"devDependencies": {
@@ -58,31 +60,31 @@
 		"@typescript-eslint/eslint-plugin-tslint": "latest",
 		"@typescript-eslint/parser": "latest",
 		"blanket": "latest",
-		"cheerio": "^1.0.0-rc.5",
+		"cheerio": "^1.0.0-rc.12",
 		"cross-env": "^7.0.3",
-		"eslint": "^7.32.0",
+		"eslint": "^8.23.1",
 		"eslint-config-prettier": "latest",
 		"eslint-plugin-import": "latest",
 		"high5": "^1.0.0",
-		"html-dom-parser": "^1.0.4",
+		"html-dom-parser": "^3.1.2",
 		"html-parser": "^0.11.0",
 		"html5parser": "^2.0.2",
-		"htmljs-parser": "^2.11.1",
+		"htmljs-parser": "^5.1.4",
 		"htmlparser": "^1.7.7",
 		"htmlparser-benchmark": "^1.1.3",
-		"htmlparser2": "^6.0.0",
+		"htmlparser2": "^8.0.1",
 		"mocha": "latest",
 		"mocha-each": "^2.0.1",
 		"neutron-html5parser": "^0.2.0",
 		"np": "latest",
-		"parse5": "^6.0.1",
+		"parse5": "^7.1.1",
 		"rimraf": "^3.0.2",
 		"saxes": "^6.0.0",
 		"should": "latest",
 		"spec": "latest",
-		"standard-version": "^9.3.1",
+		"standard-version": "^9.5.0",
 		"travis-cov": "latest",
-		"ts-node": "^10.2.1",
+		"ts-node": "^10.9.1",
 		"typescript": "latest"
 	},
 	"config": {