exchange sax to txml

TobiasNickel · Sep 13, 2020 · 5cefca6 · 5cefca6
1 parent 07ca976
commit 5cefca6
Show file tree

Hide file tree

Showing 3 changed files with 145 additions and 172 deletions.
diff --git a/lib/svgo/svg2js.js b/lib/svgo/svg2js.js
@@ -1,19 +1,13 @@
 'use strict';
 
-var SAX = require('sax'),
+const { property } = require('css-tree');
+
+var txml = require('txml'),
     JSAPI = require('./jsAPI.js'),
     CSSClassList = require('./css-class-list'),
     CSSStyleDeclaration = require('./css-style-declaration'),
     entityDeclaration = /<!ENTITY\s+(\S+)\s+(?:'([^\']+)'|"([^\"]+)")\s*>/g;
 
-var config = {
-    strict: true,
-    trim: false,
-    normalize: true,
-    lowercase: true,
-    xmlns: true,
-    position: true
-};
 
 /**
  * Convert SVG (XML) string to SVG-as-JS object.
@@ -22,179 +16,112 @@ var config = {
  * @param {Function} callback
  */
 module.exports = function(data, callback) {
+    try {
+        var DOM = txml(data, {noChildNodes:['?xml','!DOCTYPE', '!ENTITY'], keepComments: true });
 
-    var sax = SAX.parser(config.strict, config),
-        root = new JSAPI({ elem: '#document', content: [] }),
-        current = root,
-        stack = [root],
-        textContext = null,
-        parsingError = false;
-
-    function pushToContent(content) {
-
-        content = new JSAPI(content, current);
-
-        (current.content = current.content || []).push(content);
-
-        return content;
-
+        const root = dom2js(DOM);
+
+        callback(root);
+    } catch(err) {
+        callback({error: 'Error in parsing SVG: ' + err.message });
     }
 
-    sax.ondoctype = function(doctype) {
-
-        pushToContent({
-            doctype: doctype
-        });
-
-        var subsetStart = doctype.indexOf('['),
-            entityMatch;
+};
 
-        if (subsetStart >= 0) {
-            entityDeclaration.lastIndex = subsetStart;
+/**
+ * 
+ * @param {(txml.INode | string)[]} dom 
+ */
+function dom2js(dom, root, ENTITIES){
+    if(!root) root = new JSAPI({ elem: '#document', content: [] });
+    if(!ENTITIES) ENTITIES={};
+    if(!Array.isArray(dom)) return dom
+    dom.forEach(n => {
+        if (typeof n === 'object') {
+            let child;
+            if(n.tagName === '?xml'){
+                child = pushToContent({ processinginstruction: { 
+                    name: 'xml',
+                    body: Object.keys(n.attributes).map(name=>name+'='+JSON.stringify(n.attributes[name])).join(' '),
+                } }, root)
+            } else {
+                var elem = {
+                    name: n.tagName,
+                    elem: n.tagName,
+                    prefix: '',
+                    local: n.tagName,
+                    attrs: n.attributes,
+                    content: []
+                };
 
-            while ((entityMatch = entityDeclaration.exec(data)) != null) {
-                sax.ENTITIES[entityMatch[1]] = entityMatch[2] || entityMatch[3];
+                Object.keys(elem.attrs).forEach(name=>{
+                    if(name.startsWith('xmlns:')){
+
+                        elem.attrs[name] = ENTITIES[elem.attrs[name].substr(1,elem.attrs[name].length-2)] || elem.attrs[name];
+                    }
+                    //xmlns:x="&ns_extend;"
+                    elem.attrs[name] = {
+                        name,
+                        value: elem.attrs[name],
+                        prefix: '',
+                        local: name
+                    };
+                });
+
+                elem.class = new CSSClassList(elem);
+                elem.style = new CSSStyleDeclaration(elem);
+
+                child = pushToContent(elem, root);
             }
-        }
-    };
-
-    sax.onprocessinginstruction = function(data) {
-
-        pushToContent({
-            processinginstruction: data
-        });
-
-    };
-
-    sax.oncomment = function(comment) {
-
-        pushToContent({
-            comment: comment.trim()
-        });
-
-    };
-
-    sax.oncdata = function(cdata) {
-
-        pushToContent({
-            cdata: cdata
-        });
-
-    };
-
-    sax.onopentag = function(data) {
-
-        var elem = {
-            elem: data.name,
-            prefix: data.prefix,
-            local: data.local,
-            attrs: {}
-        };
-
-        elem.class = new CSSClassList(elem);
-        elem.style = new CSSStyleDeclaration(elem);
-
-        if (Object.keys(data.attributes).length) {
-            for (var name in data.attributes) {
-
-                if (name === 'class') { // has class attribute
-                    elem.class.hasClass();
-                }
-
-                if (name === 'style') { // has style attribute
-                    elem.style.hasStyle();
+
+            if (n.children) {
+                dom2js(n.children, child, ENTITIES)
+            }
+        } else if (typeof n === 'string') {
+            if(n.startsWith('<!--') && n.endsWith('-->')){
+                pushToContent({
+                    comment: n.substring(4,n.length-4).trim()
+                }, root);
+            } else if(n.startsWith('!DOCTYPE')) {
+                const data = n.substr(8);
+                pushToContent({
+                    doctype: data
+                }, root);
+
+                var subsetStart = data.indexOf('[');
+
+                if (subsetStart >= 0) {
+                    entityDeclaration.lastIndex = subsetStart;
+                    //console.log(data)
+                    var entryLines = data.split('<!ENTITY ');
+                    entryLines.shift();
+                    entryLines = entryLines
+                        .map(el=>el.split('>')[0])
+                        .map(el=>el.split(' '));
+
+                    entryLines.forEach(([name, rest]) => {
+                        ENTITIES[name] = JSON.parse(rest)
+                    });
+                    //console.log({ENTITIES});
                 }
-
-                elem.attrs[name] = {
-                    name: name,
-                    value: data.attributes[name].value,
-                    prefix: data.attributes[name].prefix,
-                    local: data.attributes[name].local
-                };
+            } else {
+                pushToContent(n, root);
             }
+        }else{
+            pushToContent(n, root);
         }
+    });
 
-        elem = pushToContent(elem);
-        current = elem;
-
-        // Save info about <text> tag to prevent trimming of meaningful whitespace
-        if (data.name == 'text' && !data.prefix) {
-            textContext = current;
-        }
-
-        stack.push(elem);
-
-    };
-
-    sax.ontext = function(text) {
-
-        if (/\S/.test(text) || textContext) {
-
-            if (!textContext)
-                text = text.trim();
-
-            pushToContent({
-                text: text
-            });
-
-        }
-
-    };
-
-    sax.onclosetag = function() {
-
-        var last = stack.pop();
-
-        // Trim text inside <text> tag.
-        if (last == textContext) {
-            trim(textContext);
-            textContext = null;
-        }
-        current = stack[stack.length - 1];
-
-    };
-
-    sax.onerror = function(e) {
-
-        e.message = 'Error in parsing SVG: ' + e.message;
-        if (e.message.indexOf('Unexpected end') < 0) {
-            throw e;
-        }
-
-    };
-
-    sax.onend = function() {
-
-        if (!this.error) {
-            callback(root);
-        } else {
-            callback({ error: this.error.message });
-        }
-
-    };
-
-    try {
-        sax.write(data);
-    } catch (e) {
-        callback({ error: e.message });
-        parsingError = true;
-    }
-    if (!parsingError) sax.close();
-
-    function trim(elem) {
-        if (!elem.content) return elem;
+    return root;
+}
 
-        var start = elem.content[0],
-            end = elem.content[elem.content.length - 1];
+function pushToContent(content, parent) {
 
-        while (start && start.content && !start.text) start = start.content[0];
-        if (start && start.text) start.text = start.text.replace(/^\s+/, '');
+    const jsapi = new JSAPI(content, parent);
 
-        while (end && end.content && !end.text) end = end.content[end.content.length - 1];
-        if (end && end.text) end.text = end.text.replace(/\s+$/, '');
+    (parent.content = parent.content || []).push(jsapi);
 
-        return elem;
+    return jsapi;
 
-    }
+}
 
-};
diff --git a/package-lock.json b/package-lock.json
diff --git a/package.json b/package.json
@@ -57,8 +57,8 @@
     "js-yaml": "^3.13.1",
     "mkdirp": "~0.5.1",
     "object.values": "^1.1.0",
-    "sax": "~1.2.4",
     "stable": "^0.1.8",
+    "txml": "^3.2.5",
     "unquote": "~1.1.1",
     "util.promisify": "~1.0.0"
   },