Add support for parsing CDATA to SaxLtx (#109)

xmppjs · Feb 12, 2017 · ff770ee · ff770ee
1 parent 6d351f3
commit ff770ee
Show file tree

Hide file tree

Showing 2 changed files with 43 additions and 3 deletions.
diff --git a/lib/parsers/ltx.js b/lib/parsers/ltx.js
@@ -13,6 +13,7 @@ var STATE_ATTR_NAME = 5
 var STATE_ATTR_EQ = 6
 var STATE_ATTR_QUOT = 7
 var STATE_ATTR_VALUE = 8
+var STATE_CDATA = 9
 
 var SaxLtx = module.exports = function SaxLtx () {
   EventEmitter.call(this)
@@ -73,13 +74,27 @@ var SaxLtx = module.exports = function SaxLtx () {
             attrs = {}
           }
           break
+        case STATE_CDATA:
+          if (c === 93 /* ] */ && data.substr(pos + 1, 2) === ']>') {
+            var cData = endRecording()
+            if (cData) {
+              this.emit('text', unescapeXML(cData))
+            }
+            state = STATE_IGNORE_COMMENT
+          }
+          break
         case STATE_TAG_NAME:
           if (c === 47 /* / */ && recordStart === pos) {
             recordStart = pos + 1
             endTag = true
           } else if (c === 33 /* ! */) {
-            recordStart = undefined
-            state = STATE_IGNORE_COMMENT
+            if (data.substr(pos + 1, 7) === '[CDATA[') {
+              recordStart = pos + 8
+              state = STATE_CDATA
+            } else {
+              recordStart = undefined
+              state = STATE_IGNORE_COMMENT
+            }
           } else if (c === 63 /* ? */) {
             recordStart = undefined
             state = STATE_IGNORE_INSTRUCTION
@@ -93,7 +108,8 @@ var SaxLtx = module.exports = function SaxLtx () {
           if (c === 62 /* > */) {
             var prevFirst = data.charCodeAt(pos - 1)
             var prevSecond = data.charCodeAt(pos - 2)
-            if (prevFirst === 45 /* - */ && prevSecond === 45 /* - */) {
+            if ((prevFirst === 45 /* - */ && prevSecond === 45 /* - */) ||
+                (prevFirst === 93 /* ] */ && prevSecond === 93 /* ] */)) {
               state = STATE_TEXT
             }
           }

diff --git a/test/cdata-test.js b/test/cdata-test.js
@@ -0,0 +1,24 @@
+'use strict'
+
+var vows = require('vows')
+var assert = require('assert')
+var ltx = require('..')
+var parsers = require('../lib/parsers')
+
+var Parser = parsers.find(function (parser) {
+  return (parser.name === 'SaxLtx')
+})
+
+var parse = function (s) {
+  return ltx.parse(s, {Parser: Parser})
+}
+
+vows.describe('sax_ltx').addBatch({
+  'CDATA parsing': {
+    'issue-19: parse CDATA content as text': function () {
+      var el = parse('<root><![CDATA[Content]]></root>')
+      assert.equal(el.name, 'root')
+      assert.equal(el.getText(), 'Content')
+    }
+  }
+}).export(module)