Revert "Decode patch from UTF-8 while parsing from stdin"

[patchwork] / apps / patchwork / parser.py

diff --git a/apps/patchwork/parser.py b/apps/patchwork/parser.py
index 021dbcdc36ae59bc382ad4cdba458be310f3495e..2b5e9a093c478756d65ef001cdb59e48eb0af0c5 100644 (file)
--- a/apps/patchwork/parser.py
+++ b/apps/patchwork/parser.py
@@ -21,6 +21,7 @@
 
 
 import re
+
 try:
     import hashlib
     sha1_hash = hashlib.sha1
@@ -132,6 +133,9 @@ def parse_patch(text):
                 lc[0] -= 1
             elif line.startswith('+'):
                 lc[1] -= 1
+            elif line.startswith('\ No newline at end of file'):
+                # Special case: Not included as part of the hunk's line count
+                pass
             else:
                 lc[0] -= 1
                 lc[1] -= 1
@@ -157,10 +161,10 @@ def parse_patch(text):
 
     return (patchbuf, commentbuf)
 
-def patch_hash(str):
+def hash_patch(str):
+    # normalise spaces
     str = str.replace('\r', '')
     str = str.strip() + '\n'
-    lines = str.split('\n')
 
     prefixes = ['-', '+', ' ']
     hash = sha1_hash()
@@ -183,9 +187,8 @@ def patch_hash(str):
 
             line = filename_match.group(1) + ' ' + filename
 
-            
         elif hunk_match:
-            # remove line numbers
+            # remove line numbers, but leave line counts
             def fn(x):
                 if not x:
                     return 1
@@ -194,18 +197,45 @@ def patch_hash(str):
             line = '@@ -%d +%d @@' % tuple(line_nos)
 
         elif line[0] in prefixes:
+            # if we have a +, - or context line, leave as-is
             pass
 
         else:
+            # other lines are ignored
             continue
 
-        hash.update(line + '\n')
+        hash.update(line.encode('utf-8') + '\n')
+
+    return hash
+
+
+def main(args):
+    from optparse import OptionParser
+
+    parser = OptionParser()
+    parser.add_option('-p', '--patch', action = 'store_true',
+            dest = 'print_patch', help = 'print parsed patch')
+    parser.add_option('-c', '--comment', action = 'store_true',
+            dest = 'print_comment', help = 'print parsed comment')
+    parser.add_option('-#', '--hash', action = 'store_true',
+            dest = 'print_hash', help = 'print patch hash')
+
+    (options, args) = parser.parse_args()
+
+    # decode from (assumed) UTF-8
+    content = sys.stdin.read().decode('utf-8')
+
+    (patch, comment) = parse_patch(content)
+
+    if options.print_hash and patch:
+        print hash_patch(patch).hexdigest()
+
+    if options.print_patch and patch:
+        print "Patch: ------\n" + patch
+
+    if options.print_comment and comment:
+        print "Comment: ----\n" + comment
 
 if __name__ == '__main__':
     import sys
-#    (patch, comment) = parse_patch(sys.stdin.read())
-#    if patch:
-#        print "Patch: ------\n" + patch
-#    if comment:
-#        print "Comment: ----\n" + comment
-    normalise_patch_content(sys.stdin.read())
+    sys.exit(main(sys.argv))