Attempt to handle line endings in a cross-platform manner

package.sh

@@ -1,5 +1,5 @@
 #!/bin/bash
-SCOURVER="0.22"
+SCOURVER="0.23"
 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

@@ -9,6 +9,16 @@
 
 <p>Copyright 2009, Jeff Schiller</p>
 
+<section id="0.23">
+	<header>
+		<h2><a href="#0.23">Version 0.23</a></h2>
+	</header>
+	<p>TBD</p>
+	<ul>
+		<li>...</li>
+	</ul>
+</section>
+
 <section id="0.22">
 	<header>
 		<h2><a href="#0.22">Version 0.22</a></h2>

scour.py

@@ -2092,7 +2092,7 @@ def serializeXML(element, options, ind = 0, preserveWhitespace = False):
 				if preserveWhitespace:
 					outString += serializeXML(child, options, 0, preserveWhitespace)
 				else:
-					outString += '\n' + serializeXML(child, options, indent + 1, preserveWhitespace)
+					outString += os.linesep + serializeXML(child, options, indent + 1, preserveWhitespace)
 					onNewLine = True
 			# text node
 			elif child.nodeType == 3:
@@ -2114,10 +2114,10 @@ def serializeXML(element, options, ind = 0, preserveWhitespace = False):
 				
 		if onNewLine: outString += (I * ind)
 		outString += '</' + element.nodeName + '>'
-		if indent > 0: outString += '\n'
+		if indent > 0: outString += os.linesep
 	else:
 		outString += '/>'
-		if indent > 0: outString += '\n'
+		if indent > 0: outString += os.linesep
 		
 	return outString
 	
@@ -2273,7 +2273,7 @@ def scourString(in_string, options=None):
 	# http://ronrothman.com/public/leftbraned/xml-dom-minidom-toprettyxml-and-silly-whitespace/
 	# rolled our own serialize function here to save on space, put id first, customize indentation, etc
 #	out_string = doc.documentElement.toprettyxml(' ')
-	out_string = serializeXML(doc.documentElement, options)
+	out_string = serializeXML(doc.documentElement, options) + os.linesep
 	
 	# now strip out empty lines
 	lines = []
@@ -2284,7 +2284,7 @@ def scourString(in_string, options=None):
 
 	# return the string stripped of empty lines
 	if options.strip_xml_prolog == False:
-		xmlprolog = '<?xml version="1.0" encoding="UTF-8" standalone="no"?>\n'
+		xmlprolog = '<?xml version="1.0" encoding="UTF-8" standalone="no"?>' + os.linesep
 	else:
 		xmlprolog = ""
 		
@@ -2385,15 +2385,15 @@ def parse_args(args=None):
 	return options, [infile, outfile]
 
 def getReport():
-	return ' Number of elements removed: ' + str(numElemsRemoved) + \
-		'\n Number of attributes removed: ' + str(numAttrsRemoved) + \
-		'\n Number of unreferenced id attributes removed: ' + str(numIDsRemoved) + \
-		'\n Number of style properties fixed: ' + str(numStylePropsFixed) + \
-		'\n Number of raster images embedded inline: ' + str(numRastersEmbedded) + \
-		'\n Number of path segments reduced/removed: ' + str(numPathSegmentsReduced) + \
-		'\n Number of bytes saved in path data: ' + str(numBytesSavedInPathData) + \
-		'\n Number of bytes saved in colors: ' + str(numBytesSavedInColors) + \
-		'\n Number of points removed from polygons: ' + str(numPointsRemovedFromPolygon)
+	return ' Number of elements removed: ' + str(numElemsRemoved) + os.linesep + \
+		' Number of attributes removed: ' + str(numAttrsRemoved) + os.linesep + \
+		' Number of unreferenced id attributes removed: ' + str(numIDsRemoved) + os.linesep + \
+		' Number of style properties fixed: ' + str(numStylePropsFixed) + os.linesep + \
+		' Number of raster images embedded inline: ' + str(numRastersEmbedded) + os.linesep + \
+		' Number of path segments reduced/removed: ' + str(numPathSegmentsReduced) + os.linesep + \
+		' Number of bytes saved in path data: ' + str(numBytesSavedInPathData) + os.linesep + \
+		' Number of bytes saved in colors: ' + str(numBytesSavedInColors) + os.linesep + \
+		' Number of points removed from polygons: ' + str(numPointsRemovedFromPolygon)
 
 if __name__ == '__main__':
 	if sys.platform == "win32":
@@ -2423,7 +2423,7 @@ if __name__ == '__main__':
 	# GZ: unless silenced by -q or something?
 	# GZ: not using globals would be good too
 	print >>sys.stderr, ' File:', input.name, \
-		'\n Time taken:', str(end-start) + 's\n', \
+		os.linesep + ' Time taken:', str(end-start) + 's' + os.linesep, \
 		getReport()
 	
 	oldsize = len(in_string)

testscour.py

@@ -834,7 +834,8 @@ class CDATAInXml(unittest.TestCase):
  <script type="application/ecmascript"><![CDATA[
   	alert('pb&j');
  ]]></script>
-</svg>''',
+</svg>
+''',
 			'Improperly serialized the cdata unit tests')
 
 class WellFormedXMLLesserThanInAttrValue(unittest.TestCase):
@@ -933,7 +934,8 @@ class DoNotPrettyPrintWhenWhitespacePreserved(unittest.TestCase):
 			'''<?xml version="1.0" encoding="UTF-8" standalone="no"?>
 <svg xmlns="http://www.w3.org/2000/svg">
  <text xml:space="preserve">This is some <tspan font-style="italic">messed-up</tspan> markup</text>
-</svg>''',
+</svg>
+''',
 			'Whitespace not preserved')
 
 class DoNotPrettyPrintWhenNestedWhitespacePreserved(unittest.TestCase):
@@ -942,7 +944,8 @@ class DoNotPrettyPrintWhenNestedWhitespacePreserved(unittest.TestCase):
 			'''<?xml version="1.0" encoding="UTF-8" standalone="no"?>
 <svg xmlns="http://www.w3.org/2000/svg">
  <text xml:space="preserve"><tspan font-style="italic">Use <tspan font-style="bold">bold</tspan> text</tspan></text>
-</svg>''',
+</svg>
+''',
 			'Whitespace not preserved when nested')
 	
 class GetAttrPrefixRight(unittest.TestCase):
@@ -954,7 +957,7 @@ class GetAttrPrefixRight(unittest.TestCase):
 class EnsurePreserveWhitespaceOnNonTextElements(unittest.TestCase):
 	def runTest(self):
 		s = scour.scourString(open('unittests/no-collapse-lines.svg').read())
-		self.assertEquals( s.count('\n'), 5,
+		self.assertEquals( len(s.splitlines()), 6,
 			'Did not properly preserve whitespace on elements even if they were not textual')
 
 class HandleEmptyStyleElement(unittest.TestCase):
@@ -967,6 +970,12 @@ class HandleEmptyStyleElement(unittest.TestCase):
 		self.assertEquals( fail, False,
 			'Could not handle an empty style element')
 
+class EnsureLineEndings(unittest.TestCase):
+	def runTest(self):
+		s = scour.scourString(open('unittests/whitespace-important.svg').read())
+		self.assertEquals( len(s.splitlines()), 4, 
+			'Did not output line ending character correctly')
+		
 # TODO; write a test for embedding rasters
 # TODO: write a test for --disable-embed-rasters
 # TODO: write tests for --keep-editor-data

unittests/whitespace-important.svg

@@ -1,4 +1,4 @@
 <?xml version="1.0" encoding="UTF-8" standalone="no"?>
 <svg xmlns="http://www.w3.org/2000/svg">
  <text xml:space="preserve">This is some <tspan font-style="italic">messed-up</tspan> markup</text>
-</svg>
+</svg>