rhodecode/tests/fixtures/diff_with_diff_data.diff

diff --git a/vcs/backends/base.py b/vcs/backends/base.py
index 212267ca23949807b8d89fa8ca495827dcfab3b1..ad17f16634da602503ed4ddd7cdd2e1ccdf4bed4 100644
--- a/vcs/backends/base.py
+++ b/vcs/backends/base.py
@@ -54,6 +54,7 @@ class BaseRepository(object):
     """
     scm = None
     DEFAULT_BRANCH_NAME = None
+    EMPTY_CHANGESET = '0' * 40
 
     def __init__(self, repo_path, create=False, **kwargs):
         """
@@ -204,6 +205,23 @@ class BaseRepository(object):
         """
         raise NotImplementedError
 
+    def get_diff(self, rev1, rev2, path=None, ignore_whitespace=False,
+            context=3):
+        """
+        Returns (git like) *diff*, as plain text. Shows changes introduced by
+        ``rev2`` since ``rev1``.
+
+        :param rev1: Entry point from which diff is shown. Can be
+          ``self.EMPTY_CHANGESET`` - in this case, patch showing all
+          the changes since empty state of the repository until ``rev2``
+        :param rev2: Until which revision changes should be shown.
+        :param ignore_whitespace: If set to ``True``, would not show whitespace
+          changes. Defaults to ``False``.
+        :param context: How many lines before/after changed lines should be
+          shown. Defaults to ``3``.
+        """
+        raise NotImplementedError
+
     # ========== #
     # COMMIT API #
     # ========== #
@@ -341,7 +359,6 @@ class BaseChangeset(object):
             otherwise; trying to access this attribute while there is no
             changesets would raise ``EmptyRepositoryError``
     """
-
     def __str__(self):
         return '<%s at %s:%s>' % (self.__class__.__name__, self.revision,
             self.short_id)
@@ -591,7 +608,6 @@ class BaseChangeset(object):
         return data
 
 
-
 class BaseWorkdir(object):
     """
     Working directory representation of single repository.
diff --git a/vcs/backends/git/repository.py b/vcs/backends/git/repository.py
index 8b9d1247fdee44e7a021b80e4965d8609cfd5720..e9f04e74dedd2f57417eb91dd2f4f7c61ec7e097 100644
--- a/vcs/backends/git/repository.py
+++ b/vcs/backends/git/repository.py
@@ -12,6 +12,7 @@
 import os
 import re
 import time
+import inspect
 import posixpath
 from dulwich.repo import Repo, NotGitRepository
 #from dulwich.config import ConfigFile
@@ -101,21 +102,6 @@ class GitRepository(BaseRepository):
                 "stderr:\n%s" % (cmd, se))
         return so, se
 
-    def _get_diff(self, rev1, rev2, path=None, ignore_whitespace=False,
-            context=3):
-        rev1 = self._get_revision(rev1)
-        rev2 = self._get_revision(rev2)
-        
-        if ignore_whitespace:
-            cmd = 'diff -U%s -w %s %s' % (context, rev1, rev2)
-        else:
-            cmd = 'diff -U%s %s %s' % (context, rev1, rev2)
-        if path:
-            cmd += ' -- "%s"' % path
-        so, se = self.run_git_command(cmd)
-
-        return so
-
     def _check_url(self, url):
         """
         Functon will check given url and try to verify if it's a valid
@@ -322,6 +308,8 @@ class GitRepository(BaseRepository):
         Returns ``GitChangeset`` object representing commit from git repository
         at the given revision or head (most recent commit) if None given.
         """
+        if isinstance(revision, GitChangeset):
+            return revision
         revision = self._get_revision(revision)
         changeset = GitChangeset(repository=self, revision=revision)
         return changeset
@@ -398,6 +386,49 @@ class GitRepository(BaseRepository):
         for rev in revs:
             yield self.get_changeset(rev)
 
+    def get_diff(self, rev1, rev2, path=None, ignore_whitespace=False,
+            context=3):
+        """
+        Returns (git like) *diff*, as plain text. Shows changes introduced by
+        ``rev2`` since ``rev1``.
+
+        :param rev1: Entry point from which diff is shown. Can be
+          ``self.EMPTY_CHANGESET`` - in this case, patch showing all
+          the changes since empty state of the repository until ``rev2``
+        :param rev2: Until which revision changes should be shown.
+        :param ignore_whitespace: If set to ``True``, would not show whitespace
+          changes. Defaults to ``False``.
+        :param context: How many lines before/after changed lines should be
+          shown. Defaults to ``3``.
+        """
+        flags = ['-U%s' % context]
+        if ignore_whitespace:
+            flags.append('-w')
+
+        if rev1 == self.EMPTY_CHANGESET:
+            rev2 = self.get_changeset(rev2).raw_id
+            cmd = ' '.join(['show'] + flags + [rev2])
+        else:
+            rev1 = self.get_changeset(rev1).raw_id
+            rev2 = self.get_changeset(rev2).raw_id
+            cmd = ' '.join(['diff'] + flags + [rev1, rev2])
+
+        if path:
+            cmd += ' -- "%s"' % path
+        stdout, stderr = self.run_git_command(cmd)
+        # If we used 'show' command, strip first few lines (until actual diff
+        # starts)
+        if rev1 == self.EMPTY_CHANGESET:
+            lines = stdout.splitlines()
+            x = 0
+            for line in lines:
+                if line.startswith('diff'):
+                    break
+                x += 1
+            # Append new line just like 'diff' command do
+            stdout = '\n'.join(lines[x:]) + '\n'
+        return stdout
+
     @LazyProperty
     def in_memory_changeset(self):
         """
diff --git a/vcs/backends/hg.py b/vcs/backends/hg.py
index f1f9f95e4d476ab01d8e7b02a8b59034c0740a3b..b7d63c552c39b2f8aaec17ef46551369c8b8e793 100644
--- a/vcs/backends/hg.py
+++ b/vcs/backends/hg.py
@@ -256,13 +256,32 @@ class MercurialRepository(BaseRepository):
 
         return map(lambda x: hex(x[7]), self._repo.changelog.index)[:-1]
 
-    def _get_diff(self, rev1, rev2, path=None, ignore_whitespace=False,
+    def get_diff(self, rev1, rev2, path='', ignore_whitespace=False,
                   context=3):
+        """
+        Returns (git like) *diff*, as plain text. Shows changes introduced by
+        ``rev2`` since ``rev1``.
+
+        :param rev1: Entry point from which diff is shown. Can be
+          ``self.EMPTY_CHANGESET`` - in this case, patch showing all
+          the changes since empty state of the repository until ``rev2``
+        :param rev2: Until which revision changes should be shown.
+        :param ignore_whitespace: If set to ``True``, would not show whitespace
+          changes. Defaults to ``False``.
+        :param context: How many lines before/after changed lines should be
+          shown. Defaults to ``3``.
+        """
+        # Check if given revisions are present at repository (may raise
+        # ChangesetDoesNotExistError)
+        if rev1 != self.EMPTY_CHANGESET:
+            self.get_changeset(rev1)
+        self.get_changeset(rev2)
+
         file_filter = match(self.path, '', [path])
-        return patch.diff(self._repo, rev1, rev2, match=file_filter,
+        return ''.join(patch.diff(self._repo, rev1, rev2, match=file_filter,
                           opts=diffopts(git=True,
                                         ignorews=ignore_whitespace,
-                                        context=context))
+                                        context=context)))
 
     def _check_url(self, url):
         """
diff --git a/vcs/tests/test_git.py b/vcs/tests/test_git.py
index 30da035a2a35c3dca14064778e97188b6d4ce5d6..d4b82b9e612af8bb5bf490a827377c7c2567735a 100644
--- a/vcs/tests/test_git.py
+++ b/vcs/tests/test_git.py
@@ -639,19 +639,19 @@ class GitSpecificWithRepoTest(BackendTestMixin, unittest.TestCase):
 
     def test_get_diff_runs_git_command_with_hashes(self):
         self.repo.run_git_command = mock.Mock(return_value=['', ''])
-        self.repo._get_diff(0, 1)
+        self.repo.get_diff(0, 1)
         self.repo.run_git_command.assert_called_once_with('diff -U%s %s %s' %
             (3, self.repo._get_revision(0), self.repo._get_revision(1)))
 
     def test_get_diff_runs_git_command_with_str_hashes(self):
         self.repo.run_git_command = mock.Mock(return_value=['', ''])
-        self.repo._get_diff('0' * 40, 1)
-        self.repo.run_git_command.assert_called_once_with('diff -U%s %s %s' %
-            (3, self.repo._get_revision(0), self.repo._get_revision(1)))
+        self.repo.get_diff(self.repo.EMPTY_CHANGESET, 1)
+        self.repo.run_git_command.assert_called_once_with('show -U%s %s' %
+            (3, self.repo._get_revision(1)))
 
     def test_get_diff_runs_git_command_with_path_if_its_given(self):
         self.repo.run_git_command = mock.Mock(return_value=['', ''])
-        self.repo._get_diff(0, 1, 'foo')
+        self.repo.get_diff(0, 1, 'foo')
         self.repo.run_git_command.assert_called_once_with('diff -U%s %s %s -- "foo"'
             % (3, self.repo._get_revision(0), self.repo._get_revision(1)))
 
diff --git a/vcs/tests/test_repository.py b/vcs/tests/test_repository.py
index e34033e29fa9b3d3366b723beab129cee73869b9..b6e3f419778d6009229e9108824acaf83eea1784 100644
--- a/vcs/tests/test_repository.py
+++ b/vcs/tests/test_repository.py
@@ -1,9 +1,12 @@
 from __future__ import with_statement
+import datetime
 from base import BackendTestMixin
 from conf import SCM_TESTS
+from conf import TEST_USER_CONFIG_FILE
+from vcs.nodes import FileNode
 from vcs.utils.compat import unittest
+from vcs.exceptions import ChangesetDoesNotExistError
 
-from conf import TEST_USER_CONFIG_FILE
 
 class RepositoryBaseTest(BackendTestMixin):
     recreate_repo_per_test = False
@@ -29,6 +32,176 @@ class RepositoryBaseTest(BackendTestMixin):
             '[e-mail]')
 
 
+
+class RepositoryGetDiffTest(BackendTestMixin):
+
+    @classmethod
+    def _get_commits(cls):
+        commits = [
+            {
+                'message': 'Initial commit',
+                'author': 'Joe Doe <[e-mail]>',
+                'date': datetime.datetime(2010, 1, 1, 20),
+                'added': [
+                    FileNode(b'foobar', content='foobar'),
+                    FileNode(b'foobar2', content='foobar2'),
+                ],
+            },
+            {
+                'message': 'Changed foobar, added foobar3',
+                'author': 'Jane Doe <[e-mail]>',
+                'date': datetime.datetime(2010, 1, 1, 21),
+                'added': [
+                    FileNode(b'foobar3', content='foobar3'),
+                ],
+                'changed': [
+                    FileNode(b'foobar', 'FOOBAR'),
+                ],
+            },
+            {
+                'message': 'Removed foobar, changed foobar3',
+                'author': 'Jane Doe <[e-mail]>',
+                'date': datetime.datetime(2010, 1, 1, 22),
+                'changed': [
+                    FileNode(b'foobar3', content='FOOBAR\nFOOBAR\nFOOBAR\n'),
+                ],
+                'removed': [FileNode(b'foobar')],
+            },
+        ]
+        return commits
+
+    def test_raise_for_wrong(self):
+        with self.assertRaises(ChangesetDoesNotExistError):
+            self.repo.get_diff('a' * 40, 'b' * 40)
+
+class GitRepositoryGetDiffTest(RepositoryGetDiffTest, unittest.TestCase):
+    backend_alias = 'git'
+
+    def test_initial_commit_diff(self):
+        initial_rev = self.repo.revisions[0]
+        self.assertEqual(self.repo.get_diff(self.repo.EMPTY_CHANGESET, initial_rev), '''diff --git a/foobar b/foobar
+new file mode 100644
+index 0000000..f6ea049
+--- /dev/null
++++ b/foobar
+@@ -0,0 +1 @@
++foobar
+\ No newline at end of file
+diff --git a/foobar2 b/foobar2
+new file mode 100644
+index 0000000..e8c9d6b
+--- /dev/null
++++ b/foobar2
+@@ -0,0 +1 @@
++foobar2
+\ No newline at end of file
+''')
+
+    def test_second_changeset_diff(self):
+        revs = self.repo.revisions
+        self.assertEqual(self.repo.get_diff(revs[0], revs[1]), '''diff --git a/foobar b/foobar
+index f6ea049..389865b 100644
+--- a/foobar
++++ b/foobar
+@@ -1 +1 @@
+-foobar
+\ No newline at end of file
++FOOBAR
+\ No newline at end of file
+diff --git a/foobar3 b/foobar3
+new file mode 100644
+index 0000000..c11c37d
+--- /dev/null
++++ b/foobar3
+@@ -0,0 +1 @@
++foobar3
+\ No newline at end of file
+''')
+
+    def test_third_changeset_diff(self):
+        revs = self.repo.revisions
+        self.assertEqual(self.repo.get_diff(revs[1], revs[2]), '''diff --git a/foobar b/foobar
+deleted file mode 100644
+index 389865b..0000000
+--- a/foobar
++++ /dev/null
+@@ -1 +0,0 @@
+-FOOBAR
+\ No newline at end of file
+diff --git a/foobar3 b/foobar3
+index c11c37d..f932447 100644
+--- a/foobar3
++++ b/foobar3
+@@ -1 +1,3 @@
+-foobar3
+\ No newline at end of file
++FOOBAR
++FOOBAR
++FOOBAR
+''')
+
+
+class HgRepositoryGetDiffTest(RepositoryGetDiffTest, unittest.TestCase):
+    backend_alias = 'hg'
+
+    def test_initial_commit_diff(self):
+        initial_rev = self.repo.revisions[0]
+        self.assertEqual(self.repo.get_diff(self.repo.EMPTY_CHANGESET, initial_rev), '''diff --git a/foobar b/foobar
+new file mode 100755
+--- /dev/null
++++ b/foobar
+@@ -0,0 +1,1 @@
++foobar
+\ No newline at end of file
+diff --git a/foobar2 b/foobar2
+new file mode 100755
+--- /dev/null
++++ b/foobar2
+@@ -0,0 +1,1 @@
++foobar2
+\ No newline at end of file
+''')
+
+    def test_second_changeset_diff(self):
+        revs = self.repo.revisions
+        self.assertEqual(self.repo.get_diff(revs[0], revs[1]), '''diff --git a/foobar b/foobar
+--- a/foobar
++++ b/foobar
+@@ -1,1 +1,1 @@
+-foobar
+\ No newline at end of file
++FOOBAR
+\ No newline at end of file
+diff --git a/foobar3 b/foobar3
+new file mode 100755
+--- /dev/null
++++ b/foobar3
+@@ -0,0 +1,1 @@
++foobar3
+\ No newline at end of file
+''')
+
+    def test_third_changeset_diff(self):
+        revs = self.repo.revisions
+        self.assertEqual(self.repo.get_diff(revs[1], revs[2]), '''diff --git a/foobar b/foobar
+deleted file mode 100755
+--- a/foobar
++++ /dev/null
+@@ -1,1 +0,0 @@
+-FOOBAR
+\ No newline at end of file
+diff --git a/foobar3 b/foobar3
+--- a/foobar3
++++ b/foobar3
+@@ -1,1 +1,3 @@
+-foobar3
+\ No newline at end of file
++FOOBAR
++FOOBAR
++FOOBAR
+''')
+
+
 # For each backend create test case class
 for alias in SCM_TESTS:
     attrs = {
@@ -38,7 +211,6 @@ for alias in SCM_TESTS:
     bases = (RepositoryBaseTest, unittest.TestCase)
     globals()[cls_name] = type(cls_name, bases, attrs)
 
-
 if __name__ == '__main__':
     unittest.main()