[PATCH pywin/scintilla/document.py] PY3 File save failed when not ascii or latin-1; And PY3 utf-8 default encoding

[kxrob]

The patch below solves:

• PY3 File save failed when not ascii or latin-1 : a regex split "[\r\n]*" for initial lines produced single chars on PY3 ! Try:

# PY2
>>> re.split("[\r\n]*", "abc\ndef", 3)
['abc', 'def']
# PY3 
>>> re.split("[\r\n]*", "abc\ndef", 3)
['', 'a', 'b', 'c\ndef']

• PY3 utf-8 default encoding

--- pywin/scintilla/document.py
+++ pywin/scintilla/document.py
@@ -113,10 +113,9 @@
 
 		# Translate from source encoding to UTF-8 bytes for Scintilla
 		source_encoding = self.source_encoding
-		# If we don't know an encoding, just use latin-1 to treat
-		# it as bytes...
+		# If we don't know an encoding, try utf-8 (PY3 default) first
 		if source_encoding is None:
-			source_encoding = 'latin1'
+			source_encoding = 'utf-8'
 		# we could optimize this by avoiding utf8 to-ing and from-ing,
 		# but then we would lose the ability to handle invalid utf8
 		# (and even then, the use of encoding aliases makes this tricky)
@@ -154,7 +153,7 @@
 				source_encoding = self.source_encoding
 			else:
 				# no BOM - look for an encoding.
-				bits = re.split("[\r\n]*", s, 3)
+				bits = re.split("[\r\n]+", s, 3)
 				for look in bits[:-1]:
 					match = re_encoding_text.search(look)
 					if match is not None:
@@ -163,7 +162,7 @@
 						break
 	
 			if source_encoding is None:
-				source_encoding = 'latin1'
+				source_encoding = 'utf-8'
 
 		## encode data before opening file so script is not lost if encoding fails
 		file_contents = s.encode(source_encoding)

• build 228