Remove line_sep option in favor of document line ending

Inspired by #23

DataConverter.py

@@ -21,6 +21,12 @@
 
 Freely adapted from Mr. Data Converter: http://shancarter.com/data_converter/
 """
+LINEENDINGS = {
+    'Unix': '\n',
+    'Windows': '\r\n',
+    'CR': '\r',
+    'LF': '\n',
+}
 
 # Borrowed from Apply Syntax
 
@@ -278,7 +284,7 @@ def get_settings(self, kwargs):
         settings['typed'] = kwargs['format'] not in self.untyped_formats
 
         # New lines
-        settings['newline'] = user_settings.get('line_sep') or os.linesep
+        settings['newline'] = LINEENDINGS.get(self.view.line_endings(), os.linesep)
 
         user_quoting = user_settings.get('quoting', 'QUOTE_MINIMAL')
         settings['quoting'] = getattr(csv, user_quoting, csv.QUOTE_MINIMAL)

DataConverter.sublime-settings

@@ -5,10 +5,6 @@
   // When false, DataConverter will assume there are no headers, and use default headers ([val1, val2, ...])
   "headers": "sniff",
 
-  // Newline character for output. Set to either a character or null.
-  // When null, DataConverter uses the OS line separator, which is "\r" in Windows.
-  "line_sep": null,
-
   // Define a dialect for the csv reader
   // Check the python docs for a description of how to define a dialect
   // https://docs.python.org/3.3/library/csv.html#dialects-and-formatting-parameters

README.md

@@ -94,13 +94,6 @@ Possible values: `"sniff"`, `true`, or `false`.
 ````
 When true, the first row is always treated as the header row. When "sniff", DataConverter will sniff for headers (sniffing isn't perfect). When false, DataConverter will assume there are no headers, and use default headers (`[val1, val2, ...]`).
 
-#### line_sep
-Character or null
-````
-"line_sep": null
-````
-Newline character for output. Set to either a character or `null`. When `null`, DataConverter uses the OS line separator, which is `"\r"` in Windows.
-
 #### dialects
 Object
 ````