Changeset - 11d6d1cec667
[Not reviewed]
default
0 3 0
Mads Kiilerich - 11 years ago 2014-12-15 13:47:36
madski@unity3d.com
diff: fix diff of renamed files with whitespace in their names

Before, header parsing would fail.
3 files changed with 18 insertions and 16 deletions:
0 comments (0 inline, 0 general)
kallithea/lib/diffs.py
Show inline comments
 
@@ -136,73 +136,75 @@ class DiffLimitExceeded(Exception):
 
    pass
 

	
 

	
 
class LimitedDiffContainer(object):
 

	
 
    def __init__(self, diff_limit, cur_diff_size, diff):
 
        self.diff = diff
 
        self.diff_limit = diff_limit
 
        self.cur_diff_size = cur_diff_size
 

	
 
    def __iter__(self):
 
        for l in self.diff:
 
            yield l
 

	
 

	
 
class DiffProcessor(object):
 
    """
 
    Give it a unified or git diff and it returns a list of the files that were
 
    mentioned in the diff together with a dict of meta information that
 
    can be used to render it in a HTML template.
 
    """
 
    _chunk_re = re.compile(r'^@@ -(\d+)(?:,(\d+))? \+(\d+)(?:,(\d+))? @@(.*)')
 
    _newline_marker = re.compile(r'^\\ No newline at end of file')
 
    _git_header_re = re.compile(r"""
 
        #^diff[ ]--git
 
        # has already been split on this:
 
        # ^diff[ ]--git
 
            [ ]a/(?P<a_path>.+?)[ ]b/(?P<b_path>.+?)\n
 
        (?:^similarity[ ]index[ ](?P<similarity_index>\d+)%\n
 
           ^rename[ ]from[ ](?P<rename_from>\S+)\n
 
           ^rename[ ]to[ ](?P<rename_to>\S+)(?:\n|$))?
 
           ^rename[ ]from[ ](?P<rename_from>.+)\n
 
           ^rename[ ]to[ ](?P<rename_to>.+)(?:\n|$))?
 
        (?:^old[ ]mode[ ](?P<old_mode>\d+)\n
 
           ^new[ ]mode[ ](?P<new_mode>\d+)(?:\n|$))?
 
        (?:^new[ ]file[ ]mode[ ](?P<new_file_mode>.+)(?:\n|$))?
 
        (?:^deleted[ ]file[ ]mode[ ](?P<deleted_file_mode>.+)(?:\n|$))?
 
        (?:^index[ ](?P<a_blob_id>[0-9A-Fa-f]+)
 
            \.\.(?P<b_blob_id>[0-9A-Fa-f]+)[ ]?(?P<b_mode>.+)?(?:\n|$))?
 
        (?:^(?P<bin_patch>GIT[ ]binary[ ]patch)(?:\n|$))?
 
        (?:^---[ ](a/(?P<a_file>.+)|/dev/null)(?:\n|$))?
 
        (?:^\+\+\+[ ](b/(?P<b_file>.+)|/dev/null)(?:\n|$))?
 
    """, re.VERBOSE | re.MULTILINE)
 
    _hg_header_re = re.compile(r"""
 
        #^diff[ ]--git
 
        # has already been split on this:
 
        # ^diff[ ]--git
 
            [ ]a/(?P<a_path>.+?)[ ]b/(?P<b_path>.+?)\n
 
        (?:^old[ ]mode[ ](?P<old_mode>\d+)\n
 
           ^new[ ]mode[ ](?P<new_mode>\d+)(?:\n|$))?
 
        (?:^similarity[ ]index[ ](?P<similarity_index>\d+)%(?:\n|$))?
 
        (?:^rename[ ]from[ ](?P<rename_from>\S+)\n
 
           ^rename[ ]to[ ](?P<rename_to>\S+)(?:\n|$))?
 
        (?:^copy[ ]from[ ](?P<copy_from>\S+)\n
 
           ^copy[ ]to[ ](?P<copy_to>\S+)(?:\n|$))?
 
        (?:^rename[ ]from[ ](?P<rename_from>.+)\n
 
           ^rename[ ]to[ ](?P<rename_to>.+)(?:\n|$))?
 
        (?:^copy[ ]from[ ](?P<copy_from>.+)\n
 
           ^copy[ ]to[ ](?P<copy_to>.+)(?:\n|$))?
 
        (?:^new[ ]file[ ]mode[ ](?P<new_file_mode>.+)(?:\n|$))?
 
        (?:^deleted[ ]file[ ]mode[ ](?P<deleted_file_mode>.+)(?:\n|$))?
 
        (?:^index[ ](?P<a_blob_id>[0-9A-Fa-f]+)
 
            \.\.(?P<b_blob_id>[0-9A-Fa-f]+)[ ]?(?P<b_mode>.+)?(?:\n|$))?
 
        (?:^(?P<bin_patch>GIT[ ]binary[ ]patch)(?:\n|$))?
 
        (?:^---[ ](a/(?P<a_file>.+)|/dev/null)(?:\n|$))?
 
        (?:^\+\+\+[ ](b/(?P<b_file>.+)|/dev/null)(?:\n|$))?
 
    """, re.VERBOSE | re.MULTILINE)
 

	
 
    #used for inline highlighter word split
 
    _token_re = re.compile(r'()(&gt;|&lt;|&amp;|<u>\t</u>| <i></i>|\W+?)')
 

	
 
    _escape_re = re.compile(r'(&)|(<)|(>)|(\t)|(?<=.)( \n| $)')
 

	
 

	
 
    def __init__(self, diff, vcs='hg', format='gitdiff', diff_limit=None):
 
        """
 
        :param diff:   a text in diff format
 
        :param vcs: type of version controll hg or git
 
        :param format: format of diff passed, `udiff` or `gitdiff`
 
        :param diff_limit: define the size of diff that is considered "big"
 
            based on that parameter cut off will be triggered, set to None
 
            to show full diff
 
        """
kallithea/tests/fixtures/hg_diff_rename_space_cr.diff
Show inline comments
 
diff --git a/ohno b/ohyes
 
rename from ohno
 
rename to ohyes
 
--- a/ohno	
 
+++ b/ohyes	
 
diff --git a/oh no b/oh yes
 
rename from oh no
 
rename to oh yes
 
--- a/oh no	
 
+++ b/oh yes	
 
@@ -1,5 +1,6 @@
 
 1
 
-2+ 2
 
-3 
 
-3 
 
+2
 
+3
 
 4
 
@@ -5,3 +6,4 @@
 
 5
 
 6
 
 7
 
+8
kallithea/tests/models/test_diff_parsers.py
Show inline comments
 
@@ -227,45 +227,45 @@ DIFF_FIXTURES = {
 
         {'added': 1,
 
          'deleted': 0,
 
          'binary': False,
 
          'ops': {COPIED_FILENODE: 'file copied from file2 to file3',
 
                  MOD_FILENODE: 'modified file'}}),
 
    ],
 
    'hg_diff_copy_and_chmod_file.diff': [
 
        ('file4', 'M',
 
         {'added': 0,
 
          'deleted': 0,
 
          'binary': True,
 
          'ops': {COPIED_FILENODE: 'file copied from file3 to file4',
 
                  CHMOD_FILENODE: 'modified file chmod 100644 => 100755'}}),
 
    ],
 
    'hg_diff_copy_chmod_and_edit_file.diff': [
 
        ('file5', 'M',
 
         {'added': 2,
 
          'deleted': 1,
 
          'binary': False,
 
          'ops': {COPIED_FILENODE: 'file copied from file4 to file5',
 
                  CHMOD_FILENODE: 'modified file chmod 100755 => 100644',
 
                  MOD_FILENODE: 'modified file'}}),
 
    ],
 
    'hg_diff_rename_space_cr.diff': [
 
        ('ohyes', 'R',
 
        ('oh yes', 'R',
 
         {'added': 4,
 
          'deleted': 2,
 
          'binary': False,
 
          'ops': {RENAMED_FILENODE: 'file renamed from ohno to ohyes'}}),
 
          'ops': {RENAMED_FILENODE: 'file renamed from oh no to oh yes'}}),
 
    ],
 
}
 

	
 

	
 
class DiffLibTest(BaseTestCase):
 

	
 
    @parameterized.expand([(x,) for x in DIFF_FIXTURES])
 
    def test_diff(self, diff_fixture):
 

	
 
        diff = fixture.load_resource(diff_fixture, strip=False)
 

	
 
        diff_proc = DiffProcessor(diff)
 
        diff_proc_d = diff_proc.prepare()
 
        data = [(x['filename'], x['operation'], x['stats']) for x in diff_proc_d]
 
        expected_data = DIFF_FIXTURES[diff_fixture]
 
        self.assertListEqual(expected_data, data)
0 comments (0 inline, 0 general)