def _before(self, *args, **kwargs):

        super(ChangesetController, self)._before(*args, **kwargs)

        c.affected_files_cut_off = 60

    def __load_data(self):

        repo_model = RepoModel()

        c.users_array = repo_model.get_users_js()

        c.user_groups_array = repo_model.get_user_groups_js()

    def _index(self, revision, method):

        c.pull_request = None

        c.anchor_url = anchor_url

        c.ignorews_url = _ignorews_url

        c.context_url = _context_url

        c.fulldiff = request.GET.get('fulldiff') # for reporting number of changed files

        # get ranges of revisions if preset

        rev_range = revision.split('...')[:2]

        enable_comments = True

        c.cs_repo = c.db_repo

        try:

            if len(rev_range) == 2:

                enable_comments = False

                rev_start = rev_range[0]

                rev_end = rev_range[1]

                rev_ranges = c.db_repo_scm_instance.get_changesets(start=rev_start,

                                                             end=rev_end)

            else:

                rev_ranges = [c.db_repo_scm_instance.get_changeset(revision)]

            c.cs_ranges = list(rev_ranges)

            if not c.cs_ranges:

                raise RepositoryError('Changeset range returned empty result')

        except (ChangesetDoesNotExistError, EmptyRepositoryError):

            log.debug(traceback.format_exc())

            msg = _('Such revision does not exist for this repository')

            h.flash(msg, category='error')

            raise HTTPNotFound()

        c.changes = OrderedDict()

        c.lines_added = 0  # count of lines added

        c.lines_deleted = 0  # count of lines removes

        c.changeset_statuses = ChangesetStatus.STATUSES

        comments = dict()

        c.statuses = []

        c.inline_comments = []

        c.inline_cnt = 0

        # Iterate over ranges (default changeset view is always one changeset)

        for changeset in c.cs_ranges:

            if method == 'show':

                c.statuses.extend([ChangesetStatusModel().get_status(

                            c.db_repo.repo_id, changeset.raw_id)])

                # Changeset comments

                comments.update((com.comment_id, com)

                                for com in ChangesetCommentsModel()

                                .get_comments(c.db_repo.repo_id,

                                              revision=changeset.raw_id))

                # Status change comments - mostly from pull requests

                comments.update((st.comment_id, st.comment)

                                for st in ChangesetStatusModel()

                                .get_statuses(c.db_repo.repo_id,

                                              changeset.raw_id, with_revisions=True)

                                if st.comment_id is not None)

                inlines = ChangesetCommentsModel() \

                            .get_inline_comments(c.db_repo.repo_id,

                                                 revision=changeset.raw_id)

                c.inline_comments.extend(inlines)

            cs2 = changeset.raw_id

            cs1 = changeset.parents[0].raw_id if changeset.parents else EmptyChangeset().raw_id

            context_lcl = get_line_ctx('', request.GET)

            ign_whitespace_lcl = get_ignore_ws('', request.GET)

            raw_diff = c.db_repo_scm_instance.get_diff(cs1, cs2,

                ignore_whitespace=ign_whitespace_lcl, context=context_lcl)

            diff_limit = None if c.fulldiff else self.cut_off_limit

            file_diff_data = []

            if method == 'show':

                diff_processor = diffs.DiffProcessor(raw_diff,

                                                     vcs=c.db_repo_scm_instance.alias,

                                                     diff_limit=diff_limit)

                c.limited_diff = diff_processor.limited_diff

                for f in diff_processor.parsed:

                    st = f['stats']

                    c.lines_added += st['added']

                    c.lines_deleted += st['deleted']

                    filename = f['filename']

                    fid = h.FID(changeset.raw_id, filename)

                    url_fid = h.FID('', filename)

            else:

                # downloads/raw we only need RAW diff nothing else

                file_diff_data.append(('', None, None, None, raw_diff, None))

            c.changes[changeset.raw_id] = (cs1, cs2, file_diff_data)

        # sort comments in creation order

        c.comments = [com for com_id, com in sorted(comments.items())]

        # count inline comments

        for __, lines in c.inline_comments:

            for comments in lines.values():

                c.inline_cnt += len(comments)

        if len(c.cs_ranges) == 1:

            c.changeset = c.cs_ranges[0]

            c.parent_tmpl = ''.join(['# Parent  %s\n' % x.raw_id

                                     for x in c.changeset.parents])

        if method == 'download':

            response.content_type = 'text/plain'

            response.content_disposition = 'attachment; filename=%s.diff' \

                                            % revision[:12]

            return raw_diff

        elif method == 'patch':

            response.content_type = 'text/plain'

            c.diff = safe_unicode(raw_diff)

            return render('changeset/patch_changeset.html')

        elif method == 'raw':

            response.content_type = 'text/plain'

            return raw_diff

        elif method == 'show':

            self.__load_data()

            if len(c.cs_ranges) == 1:

                return render('changeset/changeset.html')

            else:

                c.cs_ranges_org = None

                c.cs_comments = {}

                revs = [ctx.revision for ctx in reversed(c.cs_ranges)]

                c.jsdata = graph_data(c.db_repo_scm_instance, revs)

                return render('changeset/changeset_range.html')

    @LoginRequired()

    @HasRepoPermissionLevelDecorator('read')

    def index(self, revision, method='show'):

        return self._index(revision, method=method)

    @LoginRequired()

    @HasRepoPermissionLevelDecorator('read')

    def changeset_raw(self, revision):

        return self._index(revision, method='raw')

    @LoginRequired()

    @HasRepoPermissionLevelDecorator('read')

    def changeset_patch(self, revision):

        return self._index(revision, method='patch')

    @LoginRequired()

    @HasRepoPermissionLevelDecorator('read')

    def changeset_download(self, revision):

        return self._index(revision, method='download')

    @LoginRequired()

    @NotAnonymous()

    @HasRepoPermissionLevelDecorator('read')

    @jsonify

    def comment(self, repo_name, revision):

        assert request.environ.get('HTTP_X_PARTIAL_XHR')

        status = request.POST.get('changeset_status')

        text = request.POST.get('text', '').strip()

        c.comment = create_comment(

            text,

            status,

            revision=revision,

            f_path=request.POST.get('f_path'),

            line_no=request.POST.get('line'),

        )

        # get status if set !

        if status:

            # if latest status was from pull request and it's closed

            # disallow changing status ! RLY?

            try:

                ChangesetStatusModel().set_status(

                    c.db_repo.repo_id,

                    status,

                    request.authuser.user_id,

                    c.comment,

                    revision=revision,

                    dont_allow_on_closed_pull_request=True,

                )

            except StatusChangeOnClosedPullRequestError:

                log.debug('cannot change status on %s with closed pull request', revision)

                raise HTTPBadRequest()

        action_logger(request.authuser,

        #   Changesets in one and not in the other will be ignored

        merge = bool(request.GET.get('merge'))

        # fulldiff disables cut_off_limit

        fulldiff = request.GET.get('fulldiff')

        # partial uses compare_cs.html template directly

        partial = request.environ.get('HTTP_X_PARTIAL_XHR')

        # is_ajax_preview puts hidden input field with changeset revisions

        c.is_ajax_preview = partial and request.GET.get('is_ajax_preview')

        # swap url for compare_diff page - never partial and never is_ajax_preview

        c.swap_url = h.url('compare_url',

            repo_name=c.cs_repo.repo_name,

            org_ref_type=other_ref_type, org_ref_name=other_ref_name,

            other_repo=c.a_repo.repo_name,

            other_ref_type=org_ref_type, other_ref_name=org_ref_name,

            merge=merge or '')

        # set callbacks for generating markup for icons

        c.ignorews_url = _ignorews_url

        c.context_url = _context_url

        ignore_whitespace = request.GET.get('ignorews') == '1'

        line_context = safe_int(request.GET.get('context'), 3)

        c.a_rev = self._get_ref_rev(c.a_repo, org_ref_type, org_ref_name,

            returnempty=True)

        c.cs_rev = self._get_ref_rev(c.cs_repo, other_ref_type, other_ref_name)

        c.compare_home = False

        c.a_ref_name = org_ref_name

        c.a_ref_type = org_ref_type

        c.cs_ref_name = other_ref_name

        c.cs_ref_type = other_ref_type

        c.cs_ranges, c.cs_ranges_org, c.ancestors = self._get_changesets(

            c.a_repo.scm_instance.alias, c.a_repo.scm_instance, c.a_rev,

            c.cs_repo.scm_instance, c.cs_rev)

        raw_ids = [x.raw_id for x in c.cs_ranges]

        c.cs_comments = c.cs_repo.get_comments(raw_ids)

        c.cs_statuses = c.cs_repo.statuses(raw_ids)

        revs = [ctx.revision for ctx in reversed(c.cs_ranges)]

        c.jsdata = graph_data(c.cs_repo.scm_instance, revs)

        if partial:

            return render('compare/compare_cs.html')

        org_repo = c.a_repo

        other_repo = c.cs_repo

        if merge:

            rev1 = msg = None

            if not c.cs_ranges:

                msg = _('Cannot show empty diff')

            elif not c.ancestors:

                msg = _('No ancestor found for merge diff')

            elif len(c.ancestors) == 1:

                rev1 = c.ancestors[0]

            else:

                msg = _('Multiple merge ancestors found for merge compare')

            if rev1 is None:

                h.flash(msg, category='error')

                log.error(msg)

                raise HTTPNotFound

            # case we want a simple diff without incoming changesets,

            # previewing what will be merged.

            # Make the diff on the other repo (which is known to have other_rev)

            log.debug('Using ancestor %s as rev1 instead of %s',

                      rev1, c.a_rev)

            org_repo = other_repo

        else: # comparing tips, not necessarily linearly related

            if org_repo != other_repo:

                # TODO: we could do this by using hg unionrepo

                log.error('cannot compare across repos %s and %s', org_repo, other_repo)

                h.flash(_('Cannot compare repositories without using common ancestor'), category='error')

                raise HTTPBadRequest

            rev1 = c.a_rev

        diff_limit = None if fulldiff else self.cut_off_limit

        log.debug('running diff between %s and %s in %s',

                  rev1, c.cs_rev, org_repo.scm_instance.path)

        raw_diff = org_repo.scm_instance.get_diff(rev1=rev1, rev2=c.cs_rev,

                                      ignore_whitespace=ignore_whitespace,

                                      context=line_context)

        diff_processor = diffs.DiffProcessor(raw_diff or '', diff_limit=diff_limit)

        c.limited_diff = diff_processor.limited_diff

        c.file_diff_data = []

        c.lines_added = 0

        c.lines_deleted = 0

        for f in diff_processor.parsed:

            st = f['stats']

            c.lines_added += st['added']

            c.lines_deleted += st['deleted']

            filename = f['filename']

            fid = h.FID('', filename)

        return render('compare/compare_diff.html')

        c.update_msg_other = ""

        try:

            if not c.cs_ranges:

                c.update_msg = _('Error: changesets not found when displaying pull request from %s.') % c.cs_rev

            elif org_scm_instance.alias == 'hg' and c.a_ref_name != 'ancestor':

                if c.cs_ref_type != 'branch':

                    c.cs_branch_name = org_scm_instance.get_changeset(c.cs_ref_name).branch # use ref_type ?

                c.a_branch_name = c.a_ref_name

                if c.a_ref_type != 'branch':

                    try:

                        c.a_branch_name = other_scm_instance.get_changeset(c.a_ref_name).branch # use ref_type ?

                    except EmptyRepositoryError:

                        c.a_branch_name = 'null' # not a branch name ... but close enough

                # candidates: descendants of old head that are on the right branch

                #             and not are the old head itself ...

                #             and nothing at all if old head is a descendant of target ref name

                if not c.is_range and other_scm_instance._repo.revs('present(%s)::&%s', c.cs_ranges[-1].raw_id, c.a_branch_name):

                    c.update_msg = _('This pull request has already been merged to %s.') % c.a_branch_name

                elif c.pull_request.is_closed():

                    c.update_msg = _('This pull request has been closed and can not be updated.')

                else: # look for descendants of PR head on source branch in org repo

                    avail_revs = org_scm_instance._repo.revs('%s:: & branch(%s)',

                                                             revs[0], c.cs_branch_name)

                    if len(avail_revs) > 1: # more than just revs[0]

                        # also show changesets that not are descendants but would be merged in

                        targethead = other_scm_instance.get_changeset(c.a_branch_name).raw_id

                        if org_scm_instance.path != other_scm_instance.path:

                            # Note: org_scm_instance.path must come first so all

                            # valid revision numbers are 100% org_scm compatible

                            # - both for avail_revs and for revset results

                            hgrepo = unionrepo.unionrepository(org_scm_instance.baseui,

                                                               org_scm_instance.path,

                                                               other_scm_instance.path)

                        else:

                            hgrepo = org_scm_instance._repo

                        show = set(hgrepo.revs('::%ld & !::parents(%s) & !::%s',

                                               avail_revs, revs[0], targethead))

                        c.update_msg = _('The following additional changes are available on %s:') % c.cs_branch_name

                    else:

                        show = set()

                        avail_revs = set() # drop revs[0]

                        c.update_msg = _('No additional changesets found for iterating on this pull request.')

                    # TODO: handle branch heads that not are tip-most

                    brevs = org_scm_instance._repo.revs('%s - %ld - %s', c.cs_branch_name, avail_revs, revs[0])

                    if brevs:

                        # also show changesets that are on branch but neither ancestors nor descendants

                        show.update(org_scm_instance._repo.revs('::%ld - ::%ld - ::%s', brevs, avail_revs, c.a_branch_name))

                        show.add(revs[0]) # make sure graph shows this so we can see how they relate

                        c.update_msg_other = _('Note: Branch %s has another head: %s.') % (c.cs_branch_name,

                            h.short_id(org_scm_instance.get_changeset((max(brevs))).raw_id))

                    avail_show = sorted(show, reverse=True)

            elif org_scm_instance.alias == 'git':

                c.cs_repo.scm_instance.get_changeset(c.cs_rev) # check it exists - raise ChangesetDoesNotExistError if not

                c.update_msg = _("Git pull requests don't support iterating yet.")

        except ChangesetDoesNotExistError:

            c.update_msg = _('Error: some changesets not found when displaying pull request from %s.') % c.cs_rev

        c.avail_revs = avail_revs

        c.avail_cs = [org_scm_instance.get_changeset(r) for r in avail_show]

        c.avail_jsdata = graph_data(org_scm_instance, avail_show)

        raw_ids = [x.raw_id for x in c.cs_ranges]

        c.cs_comments = c.cs_repo.get_comments(raw_ids)

        c.cs_statuses = c.cs_repo.statuses(raw_ids)

        ignore_whitespace = request.GET.get('ignorews') == '1'

        line_context = safe_int(request.GET.get('context'), 3)

        c.ignorews_url = _ignorews_url

        c.context_url = _context_url

        fulldiff = request.GET.get('fulldiff')

        diff_limit = None if fulldiff else self.cut_off_limit

        # we swap org/other ref since we run a simple diff on one repo

        log.debug('running diff between %s and %s in %s',

                  c.a_rev, c.cs_rev, org_scm_instance.path)

        try:

            raw_diff = org_scm_instance.get_diff(rev1=safe_str(c.a_rev), rev2=safe_str(c.cs_rev),

                                                ignore_whitespace=ignore_whitespace,

                                                context=line_context)

        except ChangesetDoesNotExistError:

            raw_diff = _("The diff can't be shown - the PR revisions could not be found.")

        diff_processor = diffs.DiffProcessor(raw_diff or '', diff_limit=diff_limit)

        c.limited_diff = diff_processor.limited_diff

        c.file_diff_data = []

        c.lines_added = 0

        c.lines_deleted = 0

        for f in diff_processor.parsed:

            st = f['stats']

            c.lines_added += st['added']

            c.lines_deleted += st['deleted']

            filename = f['filename']

            fid = h.FID('', filename)

        # inline comments

        c.inline_cnt = 0

        c.inline_comments = cc_model.get_inline_comments(

                                c.db_repo.repo_id,

                                pull_request=pull_request_id)

        # count inline comments

        for __, lines in c.inline_comments:

            for comments in lines.values():

                c.inline_cnt += len(comments)

        # comments

        c.comments = cc_model.get_comments(c.db_repo.repo_id, pull_request=pull_request_id)

        # (badly named) pull-request status calculation based on reviewer votes

        (c.pull_request_reviewers,

         c.pull_request_pending_reviewers,

         c.current_voting_result,

         ) = cs_model.calculate_pull_request_result(c.pull_request)

        c.changeset_statuses = ChangesetStatus.STATUSES

        c.is_ajax_preview = False

        c.ancestors = None # [c.a_rev] ... but that is shown in an other way

        return render('/pullrequests/pullrequest_show.html')

    @LoginRequired()

    @NotAnonymous()

    @HasRepoPermissionLevelDecorator('read')

    @jsonify

    def comment(self, repo_name, pull_request_id):

        pull_request = PullRequest.get_or_404(pull_request_id)

        status = request.POST.get('changeset_status')

        close_pr = request.POST.get('save_close')

        delete = request.POST.get('save_delete')

        f_path = request.POST.get('f_path')

        line_no = request.POST.get('line')

        if (status or close_pr or delete) and (f_path or line_no):

            # status votes and closing is only possible in general comments

            raise HTTPBadRequest()

        allowed_to_change_status = self._get_is_allowed_change_status(pull_request)

        if not allowed_to_change_status:

            if status or close_pr:

                h.flash(_('No permission to change pull request status'), 'error')

                raise HTTPForbidden()

        if delete == "delete":

            if (pull_request.owner_id == request.authuser.user_id or

                h.HasPermissionAny('hg.admin')() or

                h.HasRepoPermissionLevel('admin')(pull_request.org_repo.repo_name) or

                h.HasRepoPermissionLevel('admin')(pull_request.other_repo.repo_name)

                ) and not pull_request.is_closed():

                PullRequestModel().delete(pull_request)

                Session().commit()

                h.flash(_('Successfully deleted pull request %s') % pull_request_id,

                        category='success')

                return {

                   'location': url('my_pullrequests'), # or repo pr list?

                }

                raise HTTPFound(location=url('my_pullrequests')) # or repo pr list?

            raise HTTPForbidden()

        text = request.POST.get('text', '').strip()

        comment = create_comment(

            text,

            status,

            pull_request_id=pull_request_id,

            f_path=f_path,

            line_no=line_no,

            closing_pr=close_pr,

        )

        action_logger(request.authuser,

                      'user_commented_pull_request:%s' % pull_request_id,

                      c.db_repo, request.ip_addr)

        if status:

            ChangesetStatusModel().set_status(

                c.db_repo.repo_id,

                status,

                request.authuser.user_id,

                comment,

                pull_request=pull_request_id

            )

        if close_pr:

            PullRequestModel().close_pull_request(pull_request_id)

            action_logger(request.authuser,

                          'user_closed_pull_request:%s' % pull_request_id,

                          c.db_repo, request.ip_addr)

        Session().commit()

        if not request.environ.get('HTTP_X_PARTIAL_XHR'):

# -*- coding: utf-8 -*-

# This program is free software: you can redistribute it and/or modify

# it under the terms of the GNU General Public License as published by

# the Free Software Foundation, either version 3 of the License, or

# (at your option) any later version.

#

# This program is distributed in the hope that it will be useful,

# but WITHOUT ANY WARRANTY; without even the implied warranty of

# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the

# GNU General Public License for more details.

#

# You should have received a copy of the GNU General Public License

# along with this program.  If not, see <http://www.gnu.org/licenses/>.

"""

kallithea.lib.diffs

~~~~~~~~~~~~~~~~~~~

Set of diffing helpers, previously part of vcs

This file was forked by the Kallithea project in July 2014.

Original author and date, and relevant copyright and licensing information is below:

:created_on: Dec 4, 2011

:author: marcink

:copyright: (c) 2013 RhodeCode GmbH, and others.

:license: GPLv3, see LICENSE.md for more details.

"""

import re

import difflib

import logging

from tg.i18n import ugettext as _

from kallithea.lib.vcs.exceptions import VCSError

from kallithea.lib.vcs.nodes import FileNode, SubModuleNode

from kallithea.lib.vcs.backends.base import EmptyChangeset

from kallithea.lib.helpers import escape

from kallithea.lib.utils2 import safe_unicode

log = logging.getLogger(__name__)

def wrap_to_table(html):

    """Given a string with html, return it wrapped in a table, similar to what

    DiffProcessor returns."""

    return '''\

              <table class="code-difftable">

                <tr class="line no-comment">

                <td class="lineno new"></td>

                <td class="code no-comment"><pre>%s</pre></td>

                </tr>

              </table>''' % html

def wrapped_diff(filenode_old, filenode_new, diff_limit=None,

                ignore_whitespace=True, line_context=3,

                enable_comments=False):

    """

    Returns a file diff wrapped into a table.

    Checks for diff_limit and presents a message if the diff is too big.

    """

    if filenode_old is None:

        filenode_old = FileNode(filenode_new.path, '', EmptyChangeset())

    op = None

    a_path = filenode_old.path # default, might be overriden by actual rename in diff

    if filenode_old.is_binary or filenode_new.is_binary:

        stats = (0, 0)

    elif diff_limit != -1 and (

            diff_limit is None or

            (filenode_old.size < diff_limit and filenode_new.size < diff_limit)):

        raw_diff = get_gitdiff(filenode_old, filenode_new,

                                ignore_whitespace=ignore_whitespace,

                                context=line_context)

        diff_processor = DiffProcessor(raw_diff)

        if diff_processor.parsed: # there should be exactly one element, for the specified file

            f = diff_processor.parsed[0]

            op = f['operation']

            a_path = f['old_filename']

        stats = diff_processor.stat()

    else:

                               'diff menu to display this diff'))

        stats = (0, 0)

        submodules = filter(lambda o: isinstance(o, SubModuleNode),

                            [filenode_new, filenode_old])

        if submodules:

        else:

    cs1 = filenode_old.changeset.raw_id

    cs2 = filenode_new.changeset.raw_id

def get_gitdiff(filenode_old, filenode_new, ignore_whitespace=True, context=3):

    """

    Returns git style diff between given ``filenode_old`` and ``filenode_new``.

    """

    # make sure we pass in default context

    context = context or 3

    submodules = filter(lambda o: isinstance(o, SubModuleNode),

                        [filenode_new, filenode_old])

    if submodules:

        return ''

    for filenode in (filenode_old, filenode_new):

        if not isinstance(filenode, FileNode):

            raise VCSError("Given object should be FileNode object, not %s"

                % filenode.__class__)

    repo = filenode_new.changeset.repository

    old_raw_id = getattr(filenode_old.changeset, 'raw_id', repo.EMPTY_CHANGESET)

    new_raw_id = getattr(filenode_new.changeset, 'raw_id', repo.EMPTY_CHANGESET)

    vcs_gitdiff = repo.get_diff(old_raw_id, new_raw_id, filenode_new.path,

                                ignore_whitespace, context)

    return vcs_gitdiff

NEW_FILENODE = 1

DEL_FILENODE = 2

MOD_FILENODE = 3

RENAMED_FILENODE = 4

COPIED_FILENODE = 5

CHMOD_FILENODE = 6

BIN_FILENODE = 7

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.

    """

    _diff_git_re = re.compile('^diff --git', re.MULTILINE)

    _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

            [ ]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|$))?

        (?:^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)\t?(?:\n|$))?

        (?:^\+\+\+[ ](b/(?P<b_file>.+?)|/dev/null)\t?(?:\n|$))?

    """, re.VERBOSE | re.MULTILINE)

    _hg_header_re = re.compile(r"""

        ^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)\t?(?:\n|$))?

        (?:^\+\+\+[ ](b/(?P<b_file>.+?)|/dev/null)\t?(?:\n|$))?

    """, re.VERBOSE | re.MULTILINE)

    # Used for inline highlighter word split, must match the substitutions in _escaper

    _token_re = re.compile(r'()(&amp;|&lt;|&gt;|<u>\t</u>|<u class="cr"></u>| <i></i>|\W+?)')

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

    def __init__(self, diff, vcs='hg', diff_limit=None, inline_diff=True):

        """

        :param diff:   a text in diff format

        :param vcs: type of version control hg or git

        :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.adds = 0

        self.removes = 0

        self.diff_limit = diff_limit

        self.limited_diff = False

        self.vcs = vcs

        self.parsed = self._parse_gitdiff(inline_diff=inline_diff)

    def _escaper(self, string):

        """

        Do HTML escaping/markup

        """

        def substitute(m):

            groups = m.groups()

            if groups[0]:

                return '&'

            if groups[1]:

                return '<'

            if groups[2]:

                return '>'

            if groups[3]:

                return '<u>\t</u>'

            if groups[4]:

                return '<u class="cr"></u>'

            if groups[5]:

                return ' <i></i>'

            assert False

        """

        Given an iterator of diff body lines, parse them and return a dict per

        line and added/removed totals.

        """

        added = deleted = 0