Handle unicode characters. Specify utf-8 encoding on python scripts. Added unit test for non-ASCII characters.

package.sh

@@ -1,5 +1,5 @@
 #!/bin/bash
-SCOURVER="0.10"
+SCOURVER="0.11"
 cd ..
 tar cvf scour/tarballs/scour-$SCOURVER.tar scour/scour.py scour/svg_regex.py scour/LICENSE scour/NOTICE scour/README.txt scour/release-notes.html
 gzip scour/tarballs/scour-$SCOURVER.tar

scour.py

@@ -1,4 +1,5 @@
 #!/usr/bin/env python
+# -*- coding: utf-8 -*-
 
 #  Scour
 #
@@ -71,7 +72,7 @@ from decimal import *
 getcontext().prec = 6
 
 APP = 'scour'
-VER = '0.10'
+VER = '0.11'
 COPYRIGHT = 'Copyright Jeff Schiller, 2009'
 
 NS = { 	'SVG': 		'http://www.w3.org/2000/svg', 
@@ -1046,7 +1047,7 @@ def scourXmlFile(filename, options=[]):
 #	print 'IN=',in_string
 	out_string = scourString(in_string, options)
 #	print 'OUT=',out_string
-	return xml.dom.minidom.parseString(out_string)
+	return xml.dom.minidom.parseString(out_string.encode('utf-8'))
 
 def printHeader():
 	print APP , VER
@@ -1119,7 +1120,7 @@ if __name__ == '__main__':
 	# do the work
 	in_string = input.read()
 	out_string = scourString(in_string, options)
-	output.write(out_string)
+	output.write(out_string.encode("utf-8"))
 
 	# Close input and output files
 	input.close()

testscour.py

@@ -1,4 +1,6 @@
-#!/usr/local/bin/python
+#!/usr/bin/env python
+# -*- coding: utf-8 -*-
+
 #  Test Harness for Scour
 #
 #  Copyright 2009 Jeff Schiller
@@ -492,6 +494,12 @@ class ChangeLineToVerticalLineSegmentInPath(unittest.TestCase):
 		self.assertEquals(path[2][1][0], 100.0,
 			'Did not calculate vertical line segment in path correctly' )
 
+class HandleNonAsciiUtf8(unittest.TestCase):
+	def runTest(self):
+		doc = scour.scourXmlFile('unittests/utf8.svg')
+		desc = unicode(doc.getElementsByTagNameNS(SVGNS, 'desc')[0].firstChild.wholeText).strip()
+		self.assertEquals( desc, u'ú',
+			'Did not handle non-ASCII characters' )
 
 if __name__ == '__main__':
     unittest.main()

unittests/utf8.svg

@@ -0,0 +1,4 @@
+<svg xmlns:xlink="http://www.w3.org/1999/xlink" 
+   xmlns="http://www.w3.org/2000/svg">
+  <desc>ú</desc>
+</svg>