Fix Bug 517064: Make XML well-formed again by properly translating the 5 XML entities

package.sh

@@ -1,5 +1,5 @@
 #!/bin/bash
-SCOURVER="0.23"
+SCOURVER="0.24"
 cd ..
 zip scour/tarballs/scour-$SCOURVER.zip scour/scour.py scour/yocto_css.py scour/svg_regex.py scour/LICENSE scour/NOTICE scour/README.txt scour/release-notes.html
 cd scour

release-notes.html

@@ -9,11 +9,21 @@
 
 <p>Copyright 2009, Jeff Schiller</p>
 
+<section id="0.24">
+	<header>
+		<h2><a href="#0.24">Version 0.24</a></h2>
+	</header>
+	<p>2010-02-04</p>
+	<ul>
+		<li>Fix <a href="https://bugs.launchpad.net/scour/+bug/517064">Bug 517064</a> to make XML well-formed again</li>
+	</ul>
+</section>
+
 <section id="0.23">
 	<header>
 		<h2><a href="#0.23">Version 0.23</a></h2>
 	</header>
-	<p>2009-01-04</p>
+	<p>2010-01-04</p>
 	<ul>
 		<li>Fix <a href="https://bugs.launchpad.net/scour/+bug/482215">Bug 482215</a> by using os.linesep to end lines</li>
 		<li>Fix unittests to run properly in Windows</li>

scour.py

@@ -72,7 +72,7 @@ except ImportError:
 	pass
 
 APP = 'scour'
-VER = '0.23'
+VER = '0.24'
 COPYRIGHT = 'Copyright Jeff Schiller, 2010'
 
 NS = { 	'SVG': 		'http://www.w3.org/2000/svg', 
@@ -2021,19 +2021,13 @@ def remapNamespacePrefix(node, oldprefix, newprefix):
 		remapNamespacePrefix(child, oldprefix, newprefix)	
 
 def makeWellFormed(str):
-	newstr = str
+	newstr = ''
-	
+	xml_ents = { '<':'&lt;', '>':'&gt;', '&':'&amp;', "'":'&apos;', '"':'&quot;'}
-	# encode & as &amp; ( must do this first so that &lt; does not become &amp;lt; )
+	for c in str:
-	if str.find('&') != -1:
+		if c in xml_ents:
-		newstr = str.replace('&', '&amp;')
+			newstr += xml_ents[c]
-			
+		else:
-	# encode < as &lt;
+			newstr += c
-	if str.find("<") != -1:
-		newstr = str.replace('<', '&lt;')
-		
-	# encode > as &gt; (TODO: is this necessary?)
-	if str.find('>') != -1:
-		newstr = str.replace('>', '&gt;')
 			
 	return newstr
 
@@ -2278,7 +2272,7 @@ def scourString(in_string, options=None):
 			embedRasters(elem, options)		
 
 	# properly size the SVG document (ideally width/height should be 100% with a viewBox)
-	if options.viewboxing:
+	if options.enable_viewboxing:
 		properlySizeDoc(doc.documentElement)
 
 	# output the document as a pretty string with a single space for indent
@@ -2354,7 +2348,7 @@ _options_parser.add_option("--strip-xml-prolog",
 	action="store_true", dest="strip_xml_prolog", default=False,
 	help="won't output the <?xml ?> prolog")
 _options_parser.add_option("--enable-viewboxing",
-	action="store_true", dest="viewboxing", default=False,
+	action="store_true", dest="enable_viewboxing", default=False,
 	help="changes document width/height to 100%/100% and creates viewbox coordinates")
 
 # GZ: this is confusing, most people will be thinking in terms of

testscour.py

@@ -47,7 +47,7 @@ class ScourOptions:
 	keep_editor_data = False
 	strip_xml_prolog = False
 	indent_type = "space"
-	viewboxing = False
+	enable_viewboxing = False
 
 class NoInkscapeElements(unittest.TestCase):
 	def runTest(self):
@@ -983,6 +983,11 @@ class EnsureLineEndings(unittest.TestCase):
 		self.assertEquals( len(s.splitlines()), 4, 
 			'Did not output line ending character correctly')
 
+class XmlEntities(unittest.TestCase):
+	def runTest(self):
+		self.assertEquals( scour.makeWellFormed('<>&"\''), '&lt;&gt;&amp;&quot;&apos;',
+			'Incorrectly translated XML entities')
+		
 # TODO: write tests for --enable-viewboxing
 # TODO; write a test for embedding rasters
 # TODO: write a test for --disable-embed-rasters