Changeset - 97ed7d05d3d2
[Not reviewed]
default
0 1 0
Mads Kiilerich - 11 years ago 2014-12-15 13:47:36
madski@unity3d.com
diff: eat trailing tabs in ---/+++ lines of diffs

There is usually no date in the diffs we parse but there might still be the tab
that was used as separator between filename and date.

The a_file and b_file entries are usually not used ... but nice to have them
right.
1 file changed with 4 insertions and 4 deletions:
0 comments (0 inline, 0 general)
kallithea/lib/diffs.py
Show inline comments
 
@@ -125,117 +125,117 @@ def get_gitdiff(filenode_old, filenode_n
 

	
 
NEW_FILENODE = 1
 
DEL_FILENODE = 2
 
MOD_FILENODE = 3
 
RENAMED_FILENODE = 4
 
COPIED_FILENODE = 5
 
CHMOD_FILENODE = 6
 
BIN_FILENODE = 7
 

	
 

	
 
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"""
 
        # 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>.+)\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|$))?
 
        (?:^---[ ](a/(?P<a_file>.+?)|/dev/null)\t?(?:\n|$))?
 
        (?:^\+\+\+[ ](b/(?P<b_file>.+?)|/dev/null)\t?(?:\n|$))?
 
    """, re.VERBOSE | re.MULTILINE)
 
    _hg_header_re = re.compile(r"""
 
        # 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>.+)\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|$))?
 
        (?:^---[ ](a/(?P<a_file>.+?)|/dev/null)\t?(?:\n|$))?
 
        (?:^\+\+\+[ ](b/(?P<b_file>.+?)|/dev/null)\t?(?:\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
 
        """
 
        if not isinstance(diff, basestring):
 
            raise Exception('Diff must be a basestring got %s instead' % type(diff))
 

	
 
        self._diff = diff
 
        self._format = format
 
        self.adds = 0
 
        self.removes = 0
 
        # calculate diff size
 
        self.diff_size = len(diff)
 
        self.diff_limit = diff_limit
 
        self.cur_diff_size = 0
 
        self.parsed = False
 
        self.parsed_diff = []
 
        self.vcs = vcs
 

	
 
        if format == 'gitdiff':
 
            self.differ = self._highlight_line_difflib
 
            self._parser = self._parse_gitdiff
 
        else:
 
            self.differ = self._highlight_line_udiff
 
            self._parser = self._parse_udiff
 

	
 
    def _copy_iterator(self):
 
        """
 
        make a fresh copy of generator, we should not iterate thru
 
        an original as it's needed for repeating operations on
 
        this instance of DiffProcessor
 
        """
 
        self.__udiff, iterator_copy = tee(self.__udiff)
 
        return iterator_copy
 

	
0 comments (0 inline, 0 general)