Merge pull request #88 from sirosen/alphabetize-codeowners-gitlab-syntax

Support a gitlab dialect for codeowners

README.md

@@ -179,6 +179,8 @@ following sample config:
 ### Unreleased
 
 <!-- bumpversion-changelog -->
+- Support GitLab section headers in alphabetize-codeowners when
+  `--dialect=gitlab` is passed.
 - Casefold codeowner names for better unicode sorting. Thanks @adam-moss for
   the PR!
 

src/texthooks/alphabetize_codeowners.py

@@ -3,32 +3,41 @@
 Alphabetize the list of owners for each path in .github/CODEOWNERS
 
 Ignores empty lines and comments, but normalizes whitespace on semantically significant
-lines
+lines.
+
+Use '--dialect=gitlab' in order to support GitLab's extended CODEOWNERS syntax.
 """
 
+import re
 import sys
+import typing as t
 
 from ._common import parse_cli_args
 from ._recorders import DiffRecorder
 
+GITLAB_SECTION_TITLE_PATTERN = re.compile(r"\^?\[[^\]]+\]")
+GITLAB_SECTION_N_APPROVALS_PATTERN = re.compile(r"\[\d+\]")
+
 
 def main(*, argv=None) -> int:
     args = parse_cli_args(
         __doc__,
         fixer=True,
         argv=argv,
         disable_args=["files"],
-        modify_parser=_add_files_arg,
+        modify_parser=_add_args,
     )
     filenames = args.files
     if not filenames:
         filenames = [".github/CODEOWNERS"]
 
+    line_fixer = make_line_fixer(args.dialect)
+
     recorder = DiffRecorder(args.verbosity)
     missing_file = False
     for fn in filenames:
         try:
-            recorder.run_line_fixer(sort_line, fn)
+            recorder.run_line_fixer(line_fixer, fn)
         except FileNotFoundError:
             missing_file = True
     if recorder or missing_file:
@@ -38,11 +47,32 @@ def main(*, argv=None) -> int:
     return 0
 
 
-def _add_files_arg(parser):
+def _add_args(parser):
     parser.add_argument("files", nargs="*", help="default: .github/CODEOWNERS")
+    parser.add_argument(
+        "--dialect",
+        default="standard",
+        choices=(
+            "standard",
+            "gitlab",
+        ),
+        help=(
+            "A dialect of codeowners parsing to use. "
+            "Defaults to the common syntax ('standard')."
+        ),
+    )
+
+
+def make_line_fixer(dialect: str) -> t.Callable[[str], str]:
+    if dialect == "standard":
+        return sort_line_standard
+    elif dialect == "gitlab":
+        return sort_line_gitlab
+    else:
+        raise NotImplementedError(f"Unrecognized dialect: {dialect}")
 
 
-def sort_line(line: str) -> str:
+def sort_line_standard(line: str) -> str:
     if line.strip() == "" or line.strip().startswith("#"):
         return line
     # also normalizes whitespace
@@ -52,5 +82,29 @@ def sort_line(line: str) -> str:
     return " ".join([path] + sorted(owners, key=str.casefold))
 
 
+def sort_line_gitlab(line: str) -> str:
+    if line.strip() == "" or line.strip().startswith("#"):
+        return line
+
+    title_match = GITLAB_SECTION_TITLE_PATTERN.match(line)
+    if title_match:
+        after_section = line[title_match.end() :]
+        n_approvals_match = GITLAB_SECTION_N_APPROVALS_PATTERN.match(after_section)
+        if n_approvals_match:
+            cut_point = title_match.end() + n_approvals_match.end()
+        else:
+            cut_point = title_match.end()
+        section = line[:cut_point]
+        default_owners = line[cut_point:].split()
+        if not default_owners:
+            return line
+        return " ".join([section] + sorted(default_owners, key=str.casefold))
+    else:
+        path, *owners = line.split()
+        if not owners:
+            return line
+        return " ".join([path] + sorted(owners, key=str.casefold))
+
+
 if __name__ == "__main__":
     sys.exit(main())

tests/acceptance/test_alphabetize_codeowners.py

@@ -1,8 +1,11 @@
+import pytest
+
 from texthooks.alphabetize_codeowners import main as alphabetize_codeowners_main
 
 
-def test_alphabetize_codeowners_no_changes(runner):
-    result = runner(alphabetize_codeowners_main, "foo")
+@pytest.mark.parametrize("dialect", ("standard", "gitlab"))
+def test_alphabetize_codeowners_no_changes(runner, dialect):
+    result = runner(alphabetize_codeowners_main, "foo", add_args=["--dialect", dialect])
     assert result.exit_code == 0
     assert result.file_data == "foo"
 
@@ -11,22 +14,34 @@ def test_alphabetize_codeowners_no_changes(runner):
     assert result.file_data == "/foo/bar.txt @alice @bob"
 
 
-def test_alphabetize_codeowners_normalizes_spaces(runner):
-    result = runner(alphabetize_codeowners_main, " /foo/bar.txt @alice\t@bob ")
+@pytest.mark.parametrize("dialect", ("standard", "gitlab"))
+def test_alphabetize_codeowners_normalizes_spaces(runner, dialect):
+    result = runner(
+        alphabetize_codeowners_main,
+        " /foo/bar.txt @alice\t@bob ",
+        add_args=["--dialect", dialect],
+    )
     assert result.exit_code == 1
     assert result.file_data == "/foo/bar.txt @alice @bob"
 
 
-def test_alphabetize_codeowners_sorts(runner):
-    result = runner(alphabetize_codeowners_main, "/foo/bar.txt @Bob @alice @charlie")
+@pytest.mark.parametrize("dialect", ("standard", "gitlab"))
+def test_alphabetize_codeowners_sorts(runner, dialect):
+    result = runner(
+        alphabetize_codeowners_main,
+        "/foo/bar.txt @Bob @alice @charlie",
+        add_args=["--dialect", dialect],
+    )
     assert result.exit_code == 1
     assert result.file_data == "/foo/bar.txt @alice @Bob @charlie"
 
 
-def test_alphabetize_codeowners_sorts_other(runner):
+@pytest.mark.parametrize("dialect", ("standard", "gitlab"))
+def test_alphabetize_codeowners_sorts_other(runner, dialect):
     result = runner(
         alphabetize_codeowners_main,
         "/foo/bar.txt @Andy @adam @Bob @alice @charlie @groß @grost @grose",
+        add_args=["--dialect", dialect],
     )
     assert result.exit_code == 1
     assert (
@@ -35,7 +50,8 @@ def test_alphabetize_codeowners_sorts_other(runner):
     )
 
 
-def test_alphabetize_codeowners_ignores_non_semantic_lines(runner):
+@pytest.mark.parametrize("dialect", ("standard", "gitlab"))
+def test_alphabetize_codeowners_ignores_non_semantic_lines(runner, dialect):
     result = runner(
         alphabetize_codeowners_main,
         """
@@ -44,5 +60,56 @@ def test_alphabetize_codeowners_ignores_non_semantic_lines(runner):
 # comment 2: some non-alphabetized strings
 # d c b a
 /foo/bar.txt @alice @charlie""",
+        add_args=["--dialect", dialect],
     )
     assert result.exit_code == 0
+
+
+def test_gitlab_alphabetize_codeowners_alphabetizes_default_owners(runner):
+    result = runner(
+        alphabetize_codeowners_main,
+        """\
+# section
+[D A C B]
+# optional section
+^[D A C B E]
+# section with owners
+[D A C B] @mallory @alice
+/foo/bar.txt
+/foo/baz.txt""",
+        add_args=["--dialect", "gitlab"],
+    )
+    assert result.exit_code == 1
+    assert (
+        result.file_data
+        == """\
+# section
+[D A C B]
+# optional section
+^[D A C B E]
+# section with owners
+[D A C B] @alice @mallory
+/foo/bar.txt
+/foo/baz.txt"""
+    )
+
+
+def test_gitlab_alphabetize_codeowners_alphabetizes_default_owners_with_min_reviewers(
+    runner,
+):
+    result = runner(
+        alphabetize_codeowners_main,
+        """\
+[D A C B][2] @bob @mallory @alice
+/foo/bar.txt
+/foo/baz.txt""",
+        add_args=["--dialect", "gitlab"],
+    )
+    assert result.exit_code == 1
+    assert (
+        result.file_data
+        == """\
+[D A C B][2] @alice @bob @mallory
+/foo/bar.txt
+/foo/baz.txt"""
+    )