From 466f9e173285bbdc90379d8878d50a753582505d Mon Sep 17 00:00:00 2001
From: Morgan Courbet <ping@morgan.ploki.fr>
Date: Mon, 26 Jun 2017 23:06:57 +0200
Subject: [PATCH] Add line ending detection

---
 pre_commit_hooks/mixed_line_ending.py | 53 +++++++++++++++++++++++++++
 1 file changed, 53 insertions(+)

diff --git a/pre_commit_hooks/mixed_line_ending.py b/pre_commit_hooks/mixed_line_ending.py
index 0710cf8..7b6a4bf 100644
--- a/pre_commit_hooks/mixed_line_ending.py
+++ b/pre_commit_hooks/mixed_line_ending.py
@@ -1,4 +1,6 @@
 import argparse
+import os
+import re
 import sys
 
 from enum import Enum
@@ -10,6 +12,7 @@ class CLIOption(Enum):
 
 
 class LineEnding(CLIOption):
+    CR = '\r', '\\r', 'cr'
     CRLF = '\r\n', '\\r\\n', 'crlf'
     LF = '\n', '\\n', 'lf'
 
@@ -26,11 +29,33 @@ class MixedLineEndingOption(CLIOption):
     LF = LineEnding.LF.optName
 
 
+class MixedLineDetection(Enum):
+    MIXED_MOSTLY_CRLF = True, LineEnding.CRLF.string
+    MIXED_MOSTLY_LF = True, LineEnding.LF.string
+    NOT_MIXED = False, None
+    UNKNOWN = False, None
+
+    def __init__(self, conversion, line_ending_char):
+        self.conversion = conversion
+        self.line_ending_char = line_ending_char
+
+
+# Matches CRLF
+CRLF_PATTERN = re.compile(LineEnding.CRLF.string, re.DOTALL)
+# Matches LF (without preceding CR)
+LF_PATTERN = re.compile('(?<!' + LineEnding.CR.strPrint + ')' + LineEnding.LF.strPrint, re.DOTALL)
+
+
 def mixed_line_ending(argv=None):
     options = _parse_arguments(argv)
 
     print(options)
 
+    _check_filenames(options['filenames'])
+
+    for filename in options['filenames']:
+        print(_detect_line_ending(filename))
+
     return 0
 
 
@@ -65,5 +90,33 @@ def _parse_arguments(argv=None):
     return options
 
 
+def _check_filenames(filenames):
+    for filename in filenames:
+        if not os.path.isfile(filename):
+            raise IOError('The file "{}" does not exist'.format(filename))
+
+
+def _detect_line_ending(filename):
+    f = open(filename, 'r')
+    buf = f.read()
+
+    crlf_nb = len(CRLF_PATTERN.findall(buf))
+    lf_nb = len(LF_PATTERN.findall(buf))
+
+    crlf_found = crlf_nb > 0
+    lf_found = lf_nb > 0
+
+    if crlf_nb == lf_nb:
+        return MixedLineDetection.UNKNOWN
+
+    if crlf_found ^ lf_found:
+        return MixedLineDetection.NOT_MIXED
+
+    if crlf_nb > lf_nb:
+        return MixedLineDetection.MIXED_MOSTLY_CRLF
+    else:
+        return MixedLineDetection.MIXED_MOSTLY_LF
+
+
 if __name__ == '__main__':
     sys.exit(mixed_line_ending())