- thanks to Martin R. Hufsky for the patch for git_diff_find_similar

- I updated the unittest (they failed with the patch), and added a short unittest which emulates --find-copies-harder

src/pygit2.c

@@ -264,6 +264,18 @@ moduleinit(PyObject* m)
     PyModule_AddIntConstant(m, "GIT_DIFF_RECURSE_UNTRACKED_DIRS",
                             GIT_DIFF_RECURSE_UNTRACKED_DIRS);
 
+    /* Flags for diff find similar */
+    PyModule_AddIntConstant(m, "GIT_DIFF_FIND_RENAMES",                // --find-renames
+                            GIT_DIFF_FIND_RENAMES);
+    PyModule_AddIntConstant(m, "GIT_DIFF_FIND_RENAMES_FROM_REWRITES",  // --break-rewrites=N
+                            GIT_DIFF_FIND_RENAMES_FROM_REWRITES);
+    PyModule_AddIntConstant(m, "GIT_DIFF_FIND_COPIES",                 // --find-copies
+                            GIT_DIFF_FIND_COPIES);
+    PyModule_AddIntConstant(m, "GIT_DIFF_FIND_COPIES_FROM_UNMODIFIED", // --find-copies-harder
+                            GIT_DIFF_FIND_COPIES_FROM_UNMODIFIED);
+    PyModule_AddIntConstant(m, "GIT_DIFF_FIND_AND_BREAK_REWRITES",     // --break-rewrites=/M
+                            GIT_DIFF_FIND_AND_BREAK_REWRITES);
+
     /* Flags for diffed files */
     PyModule_AddIntConstant(m, "GIT_DIFF_FILE_VALID_OID",
                             GIT_DIFF_FILE_VALID_OID);

src/pygit2/diff.c

@@ -177,10 +177,11 @@ static int diff_file_cb(void *cb_data, const git_diff_delta *delta,
             Py_DECREF(files);
         }
 
-        file = Py_BuildValue("(s,s,i)",
+        file = Py_BuildValue("(s,s,i,i)",
             delta->old_file.path,
             delta->new_file.path,
-            delta->status
+            delta->status,
+            delta->similarity
         );
 
         if (PyList_Append(files, file) == 0) {
@@ -355,6 +356,24 @@ Diff_merge(Diff *self, PyObject *args)
     Py_RETURN_NONE;
 }
 
+PyObject *
+Diff_find_similar(Diff *self, PyObject *args)
+{
+    int err;
+    git_diff_options opts = {0};
+
+    if (!PyArg_ParseTuple(args, "|i", &opts.flags))
+        return NULL;
+
+    err = git_diff_find_similar(self->diff, &opts);
+    if (err < 0)
+        return Error_set(err);
+
+    Py_XDECREF(self->diff_changes);
+    self->diff_changes = NULL;
+    Py_RETURN_NONE;
+}
+
 static void
 Diff_dealloc(Diff *self)
 {
@@ -373,6 +392,8 @@ PyGetSetDef Diff_getseters[] = {
 static PyMethodDef Diff_methods[] = {
     {"merge", (PyCFunction)Diff_merge, METH_VARARGS,
      "Merge one diff into another."},
+     {"find_similar", (PyCFunction)Diff_find_similar, METH_VARARGS,
+     "Find renamed files in diff."},
     {NULL, NULL, 0, NULL}
 };
 

test/data/testrepo.git/refs/heads/master

@@ -1 +1 @@
-ccca47fbb26183e71a7a46d165299b84e2e6c0b3
+056e626e51b1fc1ee2182800e399ed8d84c8f082

test/test_config.py

@@ -78,7 +78,7 @@ class ConfigTest(utils.RepoTestCase):
         self.assertFalse(config_read['core.bare'])
         self.assertTrue('core.editor' in config_read)
         self.assertEqual(config_read['core.editor'], 'ed')
-
+        
     def test_add(self):
         config = pygit2.Config()
 

test/test_diff.py

@@ -38,6 +38,7 @@ COMMIT_SHA1_1 = '5fe808e8953c12735680c257f56600cb0de44b10'
 COMMIT_SHA1_2 = 'c2792cfa289ae6321ecf2cd5806c2194b0fd070c'
 COMMIT_SHA1_3 = '2cdae28389c059815e951d0bb9eed6533f61a46b'
 COMMIT_SHA1_4 = 'ccca47fbb26183e71a7a46d165299b84e2e6c0b3'
+COMMIT_SHA1_5 = '056e626e51b1fc1ee2182800e399ed8d84c8f082'
 
 PATCH = b"""diff --git a/a b/a
 index 7f129fd..af431f2 100644
@@ -119,7 +120,7 @@ class DiffTest(utils.BareRepoTestCase):
         # self.assertIsNotNone is 2.7 only
         self.assertTrue(diff is not None)
         # self.assertIn is 2.7 only
-        self.assertTrue(('a','a', 3) in diff.changes['files'])
+        self.assertTrue(('a','a', 3, 0) in diff.changes['files'])
         self.assertEqual(2, len(diff.changes['hunks']))
 
         hunk = diff.changes['hunks'][0]
@@ -162,13 +163,13 @@ class DiffTest(utils.BareRepoTestCase):
         self.assertTrue(diff_c is not None)
 
         # assertIn / assertNotIn are 2.7 only
-        self.assertTrue(('b','b', 3) not in diff_b.changes['files'])
-        self.assertTrue(('b','b', 3) in diff_c.changes['files'])
+        self.assertTrue(('b','b', 3, 0) not in diff_b.changes['files'])
+        self.assertTrue(('b','b', 3, 0) in diff_c.changes['files'])
 
         diff_b.merge(diff_c)
 
         # assertIn is 2.7 only
-        self.assertTrue(('b','b', 3) in diff_b.changes['files'])
+        self.assertTrue(('b','b', 3, 0) in diff_b.changes['files'])
 
         hunk = diff_b.changes['hunks'][1]
         self.assertEqual(hunk.old_start, 1)
@@ -203,5 +204,16 @@ class DiffTest(utils.BareRepoTestCase):
         self.assertEqual(diff.changes['hunks'][0].old_oid, '7f129fd57e31e935c6d60a0c794efe4e6927664b')
         self.assertEqual(diff.changes['hunks'][0].new_oid, 'af431f20fc541ed6d5afede3e2dc7160f6f01f16')
 
+    def test_find_similar(self):
+        commit_a = self.repo[COMMIT_SHA1_4]
+        commit_b = self.repo[COMMIT_SHA1_5]
+        
+        #~ Must pass GIT_DIFF_INCLUDE_UNMODIFIED if you expect to emulate
+        #~ --find-copies-harder during rename transformion...
+        diff = commit_a.tree.diff(commit_b.tree, pygit2.GIT_DIFF_INCLUDE_UNMODIFIED)
+        self.assertFalse(('a', 'a.copy', 5, 100) in diff.changes['files'])
+        diff.find_similar(pygit2.GIT_DIFF_FIND_COPIES_FROM_UNMODIFIED)
+        self.assertTrue(('a', 'a.copy', 5, 100) in diff.changes['files'])
+
 if __name__ == '__main__':
     unittest.main()

test/test_repository.py

@@ -39,8 +39,8 @@ from pygit2 import GIT_OBJ_ANY, GIT_OBJ_BLOB, GIT_OBJ_COMMIT, init_repository, \
 from . import utils
 
 
-HEAD_SHA  = 'ccca47fbb26183e71a7a46d165299b84e2e6c0b3'
-PARENT_SHA = '2cdae28389c059815e951d0bb9eed6533f61a46b'  # HEAD^
+HEAD_SHA  = '056e626e51b1fc1ee2182800e399ed8d84c8f082'
+PARENT_SHA = 'ccca47fbb26183e71a7a46d165299b84e2e6c0b3'  # HEAD^
 A_HEX_SHA = 'af431f20fc541ed6d5afede3e2dc7160f6f01f16'
 A_BIN_SHA = binascii.unhexlify(A_HEX_SHA.encode('ascii'))