New hook 'destroyed-symlinks' to detect symlinks which are changed to regular files with a content of a path which that symlink was pointing to; move zsplit to util

pre_commit_hooks/check_executables_have_shebangs.py

@@ -8,18 +8,11 @@ from typing import Sequence
 from typing import Set
 
 from pre_commit_hooks.util import cmd_output
+from pre_commit_hooks.util import zsplit
 
 EXECUTABLE_VALUES = frozenset(('1', '3', '5', '7'))
 
 
-def zsplit(s: str) -> List[str]:
-    s = s.strip('\0')
-    if s:
-        return s.split('\0')
-    else:
-        return []
-
-
 def check_executables(paths: List[str]) -> int:
     if sys.platform == 'win32':  # pragma: win32 cover
         return _check_git_filemode(paths)

pre_commit_hooks/destroyed_symlinks.py

@@ -0,0 +1,96 @@
+import argparse
+import shlex
+import subprocess
+from typing import List
+from typing import Optional
+from typing import Sequence
+
+from pre_commit_hooks.util import cmd_output
+from pre_commit_hooks.util import zsplit
+
+ORDINARY_CHANGED_ENTRIES_MARKER = '1'
+PERMS_LINK = '120000'
+PERMS_NONEXIST = '000000'
+
+
+def find_destroyed_symlinks(files: Sequence[str]) -> List[str]:
+    destroyed_links: List[str] = []
+    if not files:
+        return destroyed_links
+    for line in zsplit(
+        cmd_output('git', 'status', '--porcelain=v2', '-z', '--', *files),
+    ):
+        splitted = line.split(' ')
+        if splitted and splitted[0] == ORDINARY_CHANGED_ENTRIES_MARKER:
+            # https://git-scm.com/docs/git-status#_changed_tracked_entries
+            (
+                _, _, _,
+                mode_HEAD,
+                mode_index,
+                _,
+                hash_HEAD,
+                hash_index,
+                *path_splitted,
+            ) = splitted
+            path = ' '.join(path_splitted)
+            if (
+                    mode_HEAD == PERMS_LINK and
+                    mode_index != PERMS_LINK and
+                    mode_index != PERMS_NONEXIST
+            ):
+                if hash_HEAD == hash_index:
+                    # if old and new hashes are equal, it's not needed to check
+                    # anything more, we've found a destroyed symlink for sure
+                    destroyed_links.append(path)
+                else:
+                    # if old and new hashes are *not* equal, it doesn't mean
+                    # that everything is OK - new file may be altered
+                    # by something like trailing-whitespace and/or
+                    # mixed-line-ending hooks so we need to go deeper
+                    SIZE_CMD = ('git', 'cat-file', '-s')
+                    size_index = int(cmd_output(*SIZE_CMD, hash_index).strip())
+                    size_HEAD = int(cmd_output(*SIZE_CMD, hash_HEAD).strip())
+
+                    # in the worst case new file may have CRLF added
+                    # so check content only if new file is bigger
+                    # not more than 2 bytes compared to the old one
+                    if size_index <= size_HEAD + 2:
+                        head_content = subprocess.check_output(
+                            ('git', 'cat-file', '-p', hash_HEAD),
+                        ).rstrip()
+                        index_content = subprocess.check_output(
+                            ('git', 'cat-file', '-p', hash_index),
+                        ).rstrip()
+                        if head_content == index_content:
+                            destroyed_links.append(path)
+    return destroyed_links
+
+
+def main(argv: Optional[Sequence[str]] = None) -> int:
+    parser = argparse.ArgumentParser()
+    parser.add_argument('filenames', nargs='*', help='Filenames to check.')
+    args = parser.parse_args(argv)
+    destroyed_links = find_destroyed_symlinks(files=args.filenames)
+    if destroyed_links:
+        print('Destroyed symlinks:')
+        for destroyed_link in destroyed_links:
+            print(f'- {destroyed_link}')
+        print('You should unstage affected files:')
+        print(
+            '\tgit reset HEAD -- {}'.format(
+                ' '.join(shlex.quote(link) for link in destroyed_links),
+            ),
+        )
+        print(
+            'And retry commit. As a long term solution '
+            'you may try to explicitly tell git that your '
+            'environment does not support symlinks:',
+        )
+        print('\tgit config core.symlinks false')
+        return 1
+    else:
+        return 0
+
+
+if __name__ == '__main__':
+    exit(main())

pre_commit_hooks/util.py

@@ -1,5 +1,6 @@
 import subprocess
 from typing import Any
+from typing import List
 from typing import Optional
 from typing import Set
 
@@ -22,3 +23,11 @@ def cmd_output(*cmd: str, retcode: Optional[int] = 0, **kwargs: Any) -> str:
     if retcode is not None and proc.returncode != retcode:
         raise CalledProcessError(cmd, retcode, proc.returncode, stdout, stderr)
     return stdout
+
+
+def zsplit(s: str) -> List[str]:
+    s = s.strip('\0')
+    if s:
+        return s.split('\0')
+    else:
+        return []