Better error handling for DEP5 parse errors

CHANGELOG.md

@@ -1,7 +1,7 @@
 <!--
 SPDX-FileCopyrightText: 2017 Free Software Foundation Europe e.V. <https://fsfe.org>
-SPDX-FileCopyrightText: © 2020 Liferay, Inc. <https://liferay.com>
 SPDX-FileCopyrightText: 2023 DB Systel GmbH
+SPDX-FileCopyrightText: © 2020 Liferay, Inc. <https://liferay.com>
 
 SPDX-License-Identifier: CC-BY-SA-4.0
 -->
@@ -80,6 +80,7 @@ CLI command and its behaviour. There are no guarantees of stability for the
 
 ### Fixed
 
+- Syntax errors in .reuse/dep5 now have better error handling. (#841)
 - Reduced python-debian minimum version to 0.1.34. (#808)
 - Fix issue in `annotate` where `--single-line` and `--multi-line` would not
   correctly raise an error with an incompatible comment style. (#853)

src/reuse/_main.py

@@ -1,6 +1,6 @@
 # SPDX-FileCopyrightText: 2017 Free Software Foundation Europe e.V. <https://fsfe.org>
-# SPDX-FileCopyrightText: © 2020 Liferay, Inc. <https://liferay.com>
 # SPDX-FileCopyrightText: 2022 Florian Snow <florian@familysnow.net>
+# SPDX-FileCopyrightText: © 2020 Liferay, Inc. <https://liferay.com>
 #
 # SPDX-License-Identifier: GPL-3.0-or-later
 
@@ -11,8 +11,11 @@
 import sys
 import warnings
 from gettext import gettext as _
+from pathlib import Path
 from typing import IO, Callable, List, Optional, Type, cast
 
+from debian.copyright import Error as DebianError
+
 from . import (
     __REUSE_version__,
     __version__,
@@ -25,7 +28,8 @@
 )
 from ._format import INDENT, fill_all, fill_paragraph
 from ._util import PathType, setup_logging
-from .project import Project, create_project
+from .project import Project
+from .vcs import find_root
 
 _LOGGER = logging.getLogger(__name__)
 
@@ -277,10 +281,29 @@ def main(args: Optional[List[str]] = None, out: IO[str] = sys.stdout) -> int:
         out.write(f"reuse {__version__}\n")
         return 0
 
-    if parsed_args.root:
-        project = Project(parsed_args.root)
-    else:
-        project = create_project()
+    root = parsed_args.root
+    if root is None:
+        root = find_root()
+    if root is None:
+        root = Path.cwd()
+    try:
+        project = Project.from_directory(root)
+    # FileNotFoundError and NotADirectoryError don't need to be caught because
+    # argparse already made sure of these things.
+    except UnicodeDecodeError:
+        main_parser.error(
+            _("'{dep5}' could not be decoded as UTF-8.").format(
+                dep5=root / ".reuse/dep5"
+            )
+        )
+    except DebianError as error:
+        main_parser.error(
+            _(
+                "'{dep5}' could not be parsed. We received the following error"
+                " message: {message}"
+            ).format(dep5=root / ".reuse/dep5", message=str(error))
+        )
+
     project.include_submodules = parsed_args.include_submodules
     project.include_meson_subprojects = parsed_args.include_meson_subprojects
 

src/reuse/_util.py

@@ -1,12 +1,12 @@
 # SPDX-FileCopyrightText: 2017 Free Software Foundation Europe e.V. <https://fsfe.org>
-# SPDX-FileCopyrightText: © 2020 Liferay, Inc. <https://liferay.com>
 # SPDX-FileCopyrightText: 2020 Tuomas Siipola <tuomas@zpl.fi>
-# SPDX-FileCopyrightText: 2022 Nico Rikken <nico.rikken@fsfe.org>
-# SPDX-FileCopyrightText: 2022 Florian Snow <florian@familysnow.net>
 # SPDX-FileCopyrightText: 2022 Carmen Bianca Bakker <carmenbianca@fsfe.org>
+# SPDX-FileCopyrightText: 2022 Florian Snow <florian@familysnow.net>
+# SPDX-FileCopyrightText: 2022 Nico Rikken <nico.rikken@fsfe.org>
 # SPDX-FileCopyrightText: 2022 Pietro Albini <pietro.albini@ferrous-systems.com>
 # SPDX-FileCopyrightText: 2023 DB Systel GmbH
 # SPDX-FileCopyrightText: 2023 Johannes Zarl-Zierl <johannes@zarl-zierl.at>
+# SPDX-FileCopyrightText: © 2020 Liferay, Inc. <https://liferay.com>
 #
 # SPDX-License-Identifier: GPL-3.0-or-later
 
@@ -44,6 +44,7 @@
 from binaryornot.check import is_binary
 from boolean.boolean import Expression, ParseError
 from debian.copyright import Copyright
+from debian.copyright import Error as DebianError
 from license_expression import ExpressionError, Licensing
 
 from . import ReuseInfo, SourceType
@@ -244,6 +245,30 @@ def _determine_license_suffix_path(path: StrPath) -> Path:
     return Path(f"{path}.license")
 
 
+def _parse_dep5(path: StrPath) -> Copyright:
+    """Parse the dep5 file and create a dep5 Copyright object.
+
+    Raises:
+        FileNotFoundError: file doesn't exist.
+        DebianError: file could not be parsed.
+        UnicodeDecodeError: could not decode file as UTF-8.
+    """
+    path = Path(path)
+    try:
+        with path.open(encoding="utf-8") as fp:
+            return Copyright(fp)
+    except FileNotFoundError:
+        _LOGGER.debug(_("no '{}' file, or could not read it").format(path))
+        raise
+    # TODO: Remove ValueError once
+    # <https://salsa.debian.org/python-debian-team/python-debian/-/merge_requests/123>
+    # is closed
+    except (DebianError, ValueError) as error:
+        if error.__class__ == ValueError:
+            raise DebianError(str(error)) from error
+        raise
+
+
 def _copyright_from_dep5(path: StrPath, dep5_copyright: Copyright) -> ReuseInfo:
     """Find the reuse information of *path* in the dep5 Copyright object."""
     path = PurePath(path).as_posix()