Support Unicode in files when generating commits

pedroasad · jimporter · commit 127a1cc49a28 · 2021-09-06T16:15:16.000-07:00
Previously, there was an issue where `FileInfo` objects could be constructed
with the data as a `str`, confusing `git fast-import`, since `len(my_str)`
returns the number of code points, not the size in bytes.
diff --git a/CHANGES.md b/CHANGES.md
@@ -1,5 +1,13 @@
 # Changes
 
+## v1.1.1 (in progress)
+
+### Bug fixes
+
+- Fix support for Unicode in redirection templates
+
+---
+
 ## v1.1.0 (2021-09-01)
 
 ### New features
diff --git a/mike/git_utils.py b/mike/git_utils.py
@@ -130,6 +130,8 @@ def try_rebase_branch(remote, branch, force=False):
 
 class FileInfo:
     def __init__(self, path, data, mode=0o100644):
+        if isinstance(data, str):
+            data = data.encode('utf-8')
         self.path = path
         self.data = data
         self.mode = mode
@@ -171,10 +173,17 @@ def __exit__(self, exc_type, exc_value, traceback):
                 self.finish()
 
     def _write(self, data):
-        if isinstance(data, str):  # pragma: no branch
+        if isinstance(data, str):
             data = data.encode('utf-8')
         return self._pipe.stdin.write(data)
 
+    def _write_data(self, data):
+        if isinstance(data, str):
+            data = data.encode('utf-8')
+        self._write('data {}\n'.format(len(data)))
+        self._write(data)
+        self._write('\n')
+
     def _start_commit(self, branch, message):
         name = get_config('user.name')
         if re.search(r'[<>\n]', name):
@@ -190,9 +199,7 @@ def _start_commit(self, branch, message):
         self._write('committer {name}<{email}> {time}\n'.format(
             name=name + ' ' if name else '', email=email, time=make_when()
         ))
-        self._write('data {length}\n{message}\n'.format(
-            length=len(message), message=message
-        ))
+        self._write_data(message)
         try:
             head = get_latest_commit(branch)
             self._write('from {}\n'.format(head))
@@ -210,9 +217,7 @@ def add_file(self, file_info):
         self._write('M {mode:06o} inline {path}\n'.format(
             path=git_path(file_info.path), mode=file_info.mode
         ))
-        self._write('data {}\n'.format(len(file_info.data)))
-        self._write(file_info.data)
-        self._write('\n')
+        self._write_data(file_info.data)
 
     def finish(self):
         if self._finished:
diff --git a/test/unit/test_git_utils.py b/test/unit/test_git_utils.py
@@ -243,9 +243,9 @@ def setUp(self):
         self.stage = stage_dir('commit')
         git_init()
 
-    def _add_file(self, name, branch='master'):
+    def _add_file(self, name, branch='master', data='this is some text'):
         with git_utils.Commit(branch, 'add file') as commit:
-            commit.add_file(git_utils.FileInfo(name, 'this is some text'))
+            commit.add_file(git_utils.FileInfo(name, data))
 
     def test_add_file(self):
         self._add_file('file.txt')
@@ -254,6 +254,13 @@ def test_add_file(self):
         with open('file.txt') as f:
             self.assertEqual(f.read(), 'this is some text')
 
+    def test_add_file_unicode(self):
+        self._add_file('file.txt', data='レッサーパンダ')
+        check_call_silent(['git', 'checkout', 'master'])
+        assertDirectory('.', {'file.txt'})
+        with open('file.txt', encoding='utf-8') as f:
+            self.assertEqual(f.read(), 'レッサーパンダ')
+
     def test_add_file_to_dir(self):
         self._add_file(os.path.join('dir', 'file.txt'))
         check_call_silent(['git', 'checkout', 'master'])
