health-check: Add copyright header check

New files will require a copyright header, which is checked by CI
and locally.

BUG=b:242605601
TEST=touch foo.rs && ./tools/health-check; ./tools/health-check --fix

Change-Id: I31bf299bd636a5da4f806c32ca8bdf9cfd4c01f3
Reviewed-on: https://chromium-review.googlesource.com/c/crosvm/crosvm/+/3832787
Reviewed-by: Daniel Verkamp <dverkamp@chromium.org>
Tested-by: Dennis Kempin <denniskempin@google.com>
Commit-Queue: Dennis Kempin <denniskempin@google.com>

tools/health-check

@@ -3,11 +3,22 @@
 # Use of this source code is governed by a BSD-style license that can be
 # found in the LICENSE file.
 
+import re
+import sys
+from datetime import datetime
 from pathlib import Path
 from typing import List
-import sys
+
-from impl.common import CROSVM_ROOT, parallel, run_main, cmd, chdir, argh, cwd_context
 from impl.check_code_hygiene import has_crlf_line_endings
+from impl.common import (
+    CROSVM_ROOT,
+    argh,
+    chdir,
+    cmd,
+    cwd_context,
+    parallel,
+    run_main,
+)
 from impl.health_check import Check, CheckContext, run_checks
 
 
@@ -74,8 +85,55 @@ def check_rust_lockfiles(_: CheckContext):
                 raise Exception("Cargo.lock out of date")
 
 
+LICENSE_HEADER_RE = (
+    # Line 1 - copyright.
+    r".*Copyright(?P<copyright> \(c\))? (?P<year>20[0-9]{2})(?:-20[0-9]{2})? "
+    r"The Chromium(?P<chromium_space_os> )?OS Authors\."
+    r"(?P<rights_reserved> All rights reserved\.)?\n"
+    # Line 2 - License.
+    r".*Use of this source code is governed by a BSD-style license that can "
+    r"be\n"
+    # Line 3 - License continuation.
+    r".*found in the LICENSE file\.\n"
+)
+
+NEW_LICENSE_HEADER = [
+    f"Copyright {datetime.now().year} The ChromiumOS Authors.",
+    "Use of this source code is governed by a BSD-style license that can be",
+    "found in the LICENSE file.",
+]
+
+
+def new_licence_header(file_suffix: str):
+    if file_suffix == ".py" or file_suffix == "":
+        prefix = "#"
+    else:
+        prefix = "//"
+    return "\n".join(f"{prefix} {line}" for line in NEW_LICENSE_HEADER) + "\n\n"
+
+
+def check_copyright_header(context: CheckContext):
+    "Checks copyright header on new files only. Can 'fix' them if needed."
+    license_re = re.compile(LICENSE_HEADER_RE, re.MULTILINE)
+    for file in context.new_files:
+        header = file.open("r").read(256)
+        license_match = license_re.search(header)
+        if license_match:
+            continue
+        if context.fix:
+            contents = file.read_text()
+            file.write_text(new_licence_header(file.suffix) + contents)
+        else:
+            raise Exception(f"Bad copyright header: {file}")
+
+
 # List of all checks and on which files they should run.
 CHECKS: List[Check] = [
+    Check(
+        check_copyright_header,
+        files=["**.rs", "**.py"],
+        python_tools=True,
+    ),
     Check(
         check_rust_format,
         files=["**.rs"],

tools/impl/health_check.py

@@ -3,13 +3,13 @@
 # Use of this source code is governed by a BSD-style license that can be
 # found in the LICENSE file.
 
+from dataclasses import dataclass
 from fnmatch import fnmatch
+from pathlib import Path
 from time import time
 from typing import Callable, List, NamedTuple
-from pathlib import Path
-from impl.common import all_tracked_files, cmd, verbose
 
-from dataclasses import dataclass
+from impl.common import all_tracked_files, cmd, verbose
 
 git = cmd("git")
 
@@ -30,6 +30,9 @@ class CheckContext(object):
     # Those files of all_files that were modified locally.
     modified_files: List[Path]
 
+    # Files that do not exist upstream and have been added locally.
+    new_files: List[Path]
+
 
 class Check(NamedTuple):
     "Metadata for each check, definining on which files it should run."
@@ -53,7 +56,7 @@ class Check(NamedTuple):
         return name
 
 
-def list_modified_files():
+def list_file_diff():
     """
     Lists files there were modified compared to the upstream branch.
 
@@ -61,10 +64,13 @@ def list_modified_files():
     """
     upstream = git("rev-parse @{u}").stdout(check=False)
     if upstream:
-        return (Path(f) for f in git("diff --name-only", upstream).lines())
+        for line in git("diff --name-status", upstream).lines():
+            parts = line.split("\t", 1)
+            yield (parts[0].strip(), Path(parts[1].strip()))
     else:
         print("WARNING: Not tracking a branch. Checking all files.")
-        return all_tracked_files()
+        for file in all_tracked_files():
+            yield ("M", file)
 
 
 def should_run_check_on_file(check: Check, file: Path):
@@ -129,10 +135,12 @@ def run_checks(
         nightly: Use nightly version of rust tooling.
     """
     all_files = [*all_tracked_files()]
+    file_diff = [*list_file_diff()]
+    new_files = [f for (s, f) in file_diff if s == "A"]
     if run_on_all_files:
         modified_files = all_files
     else:
-        modified_files = [*list_modified_files()]
+        modified_files = [f for (s, f) in file_diff if s in ("M", "A")]
 
     success = True
     for check in checks_list:
@@ -141,6 +149,7 @@ def run_checks(
             nightly=nightly,
             all_files=[f for f in all_files if should_run_check_on_file(check, f)],
             modified_files=[f for f in modified_files if should_run_check_on_file(check, f)],
+            new_files=[f for f in new_files if should_run_check_on_file(check, f)],
         )
         if context.modified_files:
             if not run_check(check, context):