Enhance diff parser on handling hunk links, fix #20

cdiff.py

@@ -560,17 +560,14 @@ class DiffParser(object):
             line = decode(line)
 
             if difflet.is_old_path(line):
-                # FIXME: '--- ' breaks here, need to probe next 3 lines,
-                # reproducible with github raw patch
-                #
-                if diff._old_path and diff._new_path and len(diff._hunks) > 0:
-                    # One diff constructed
+                # FIXME: '--- ' breaks here, better to probe next line
+                if diff._old_path and diff._new_path and diff._hunks:
+                    # See a new diff, yield previous diff if exists
                     yield diff
-                    diff = Diff([], None, None, [])
                 diff = Diff(headers, line, None, [])
                 headers = []
 
-            elif difflet.is_new_path(line):
+            elif difflet.is_new_path(line) and diff._old_path:
                 diff._new_path = line
 
             elif difflet.is_hunk_meta(line):
@@ -583,9 +580,9 @@ class DiffParser(object):
                 headers = []
                 diff._hunks.append(hunk)
 
-            elif len(diff._hunks) > 0 and (difflet.is_old(line) or
-                                           difflet.is_new(line) or
-                                           difflet.is_common(line)):
+            elif diff._hunks and not headers and (difflet.is_old(line) or
+                                                  difflet.is_new(line) or
+                                                  difflet.is_common(line)):
                 diff._hunks[-1].append(difflet.parse_hunk_line(line))
 
             elif difflet.is_eof(line):
@@ -597,7 +594,7 @@ class DiffParser(object):
                 # 'Only in foo:' and 'Binary files ... differ' are considered
                 # as separate diffs, so yield current diff, then this line
                 #
-                if diff._old_path and diff._new_path and len(diff._hunks) > 0:
+                if diff._old_path and diff._new_path and diff._hunks:
                     # Current diff is comppletely constructed
                     yield diff
                 headers.append(line)

tests/git-log/in.diff

@@ -0,0 +1,37 @@
+commit 65f33a326fb1d81e9b56cfa9dbe3af887ed91c8d
+Author: Steven Myint
+Date:   Fri Mar 22 14:59:34 2013 -0700
+
+    Remove unused imports
+
+diff --git a/libmodernize/fixes/__init__.py b/libmodernize/fixes/__init__.py
+index f8628f1..2e06052 100644
+--- a/libmodernize/fixes/__init__.py
++++ b/libmodernize/fixes/__init__.py
+@@ -1,7 +1,3 @@
+-import sys
+-from lib2to3 import refactor
+-
+-
+ lib2to3_fix_names = set([
+     'lib2to3.fixes.fix_apply',
+     'lib2to3.fixes.fix_except',
+
+commit 2e80837bf815d0686138beec9a5bb94e3282204d
+Author: Steven Myint
+Date:   Fri Mar 22 14:33:55 2013 -0700
+
+    Format docstring
+
+diff --git a/libmodernize/main.py b/libmodernize/main.py
+index caf847d..b24db29 100644
+--- a/libmodernize/main.py
++++ b/libmodernize/main.py
+@@ -14,6 +14,7 @@ def main(args=None):
+     """Main program.
+ 
+     Returns a suggested exit status (0, 1, 2).
++
+     """
+     # Set up option parser
+     parser = optparse.OptionParser(usage="modernize [options] file|dir ...")

tests/git-log/out.normal

@@ -0,0 +1,38 @@
+commit 65f33a326fb1d81e9b56cfa9dbe3af887ed91c8d
+Author: Steven Myint
+Date:   Fri Mar 22 14:59:34 2013 -0700
+
+    Remove unused imports
+
+diff --git a/libmodernize/fixes/__init__.py b/libmodernize/fixes/__init__.py
+index f8628f1..2e06052 100644
+--- a/libmodernize/fixes/__init__.py
++++ b/libmodernize/fixes/__init__.py
+@@ -1,7 +1,3 @@
+-import sys
+-from lib2to3 import refactor
+-
+-
+ lib2to3_fix_names = set([
+     'lib2to3.fixes.fix_apply',
+     'lib2to3.fixes.fix_except',
+
+commit 2e80837bf815d0686138beec9a5bb94e3282204d
+Author: Steven Myint
+Date:   Fri Mar 22 14:33:55 2013 -0700
+
+    Format docstring
+
+diff --git a/libmodernize/main.py b/libmodernize/main.py
+index caf847d..b24db29 100644
+--- a/libmodernize/main.py
++++ b/libmodernize/main.py
+@@ -14,6 +14,7 @@ def main(args=None):
+     """Main program.
+ 
+     Returns a suggested exit status (0, 1, 2).
++
+     """
+     # Set up option parser
+     parser = optparse.OptionParser(usage="modernize [options] file|dir ...")
+

tests/git-log/out.side-by-side

@@ -0,0 +1,37 @@
+commit 65f33a326fb1d81e9b56cfa9dbe3af887ed91c8d
+Author: Steven Myint
+Date:   Fri Mar 22 14:59:34 2013 -0700
+
+    Remove unused imports
+
+diff --git a/libmodernize/fixes/__init__.py b/libmodernize/fixes/__init__.py
+index f8628f1..2e06052 100644
+--- a/libmodernize/fixes/__init__.py
++++ b/libmodernize/fixes/__init__.py
+@@ -1,7 +1,3 @@
+1 import sys   
+2 from lib2to3 import refactor   
+3    
+4    
+5 lib2to3_fix_names = set([                                                        1 lib2to3_fix_names = set([
+6     'lib2to3.fixes.fix_apply',                                                   2     'lib2to3.fixes.fix_apply',
+7     'lib2to3.fixes.fix_except',                                                  3     'lib2to3.fixes.fix_except',
+
+commit 2e80837bf815d0686138beec9a5bb94e3282204d
+Author: Steven Myint
+Date:   Fri Mar 22 14:33:55 2013 -0700
+
+    Format docstring
+
+diff --git a/libmodernize/main.py b/libmodernize/main.py
+index caf847d..b24db29 100644
+--- a/libmodernize/main.py
++++ b/libmodernize/main.py
+@@ -14,6 +14,7 @@ def main(args=None):
+14     """Main program.                                                             14     """Main program.
+15                                                                                  15 
+16     Returns a suggested exit status (0, 1, 2).                                   16     Returns a suggested exit status (0, 1, 2).
+                                                                                    17 
+17     """                                                                          18     """
+18     # Set up option parser                                                       19     # Set up option parser
+19     parser = optparse.OptionParser(usage="modernize [options] file|dir ...")     20     parser = optparse.OptionParser(usage="modernize [options] file|dir ...")

tests/git-log/out.w70

@@ -0,0 +1,37 @@
+commit 65f33a326fb1d81e9b56cfa9dbe3af887ed91c8d
+Author: Steven Myint
+Date:   Fri Mar 22 14:59:34 2013 -0700
+
+    Remove unused imports
+
+diff --git a/libmodernize/fixes/__init__.py b/libmodernize/fixes/__init__.py
+index f8628f1..2e06052 100644
+--- a/libmodernize/fixes/__init__.py
++++ b/libmodernize/fixes/__init__.py
+@@ -1,7 +1,3 @@
+1 import sys   
+2 from lib2to3 import refactor   
+3    
+4    
+5 lib2to3_fix_names = set([                                              1 lib2to3_fix_names = set([
+6     'lib2to3.fixes.fix_apply',                                         2     'lib2to3.fixes.fix_apply',
+7     'lib2to3.fixes.fix_except',                                        3     'lib2to3.fixes.fix_except',
+
+commit 2e80837bf815d0686138beec9a5bb94e3282204d
+Author: Steven Myint
+Date:   Fri Mar 22 14:33:55 2013 -0700
+
+    Format docstring
+
+diff --git a/libmodernize/main.py b/libmodernize/main.py
+index caf847d..b24db29 100644
+--- a/libmodernize/main.py
++++ b/libmodernize/main.py
+@@ -14,6 +14,7 @@ def main(args=None):
+14     """Main program.                                                   14     """Main program.
+15                                                                        15 
+16     Returns a suggested exit status (0, 1, 2).                         16     Returns a suggested exit status (0, 1, 2).
+                                                                          17 
+17     """                                                                18     """
+18     # Set up option parser                                             19     # Set up option parser
+19     parser = optparse.OptionParser(usage="modernize [options] file|di> 20     parser = optparse.OptionParser(usage="modernize [options] file|di>