feat: no strict dep5 parsing

src/reuse/project.py

@@ -306,7 +306,7 @@ def _copyright(self) -> Optional[Copyright]:
             copyright_path = self.root / ".reuse/dep5"
             try:
                 with copyright_path.open(encoding="utf-8") as fp:
-                    self._copyright_val = Copyright(fp)
+                    self._copyright_val = Copyright(fp, strict=False)
             except OSError:
                 _LOGGER.debug("no .reuse/dep5 file, or could not read it")
             except DebianError:

src/reuse/report.py

@@ -19,9 +19,9 @@
 from typing import Any, Dict, Iterable, List, NamedTuple, Optional, Set, cast
 from uuid import uuid4
 
-from . import __REUSE_version__, __version__
+from . import ReuseInfo, __REUSE_version__, __version__
 from ._util import _LICENSING, StrPath, _checksum
-from .project import Project, ReuseInfo
+from .project import Project
 
 _LOGGER = logging.getLogger(__name__)
 

tests/conftest.py

@@ -298,7 +298,7 @@ def dep5_copyright():
     with (RESOURCES_DIRECTORY / "fake_repository/.reuse/dep5").open(
         encoding="utf-8"
     ) as fp:
-        return Copyright(fp)
+        return Copyright(fp, strict=False)
 
 
 @pytest.fixture()

tests/test_report.py

@@ -15,6 +15,7 @@
 import pytest
 
 from reuse import SourceType
+from reuse._util import _LICENSING
 from reuse.project import Project
 from reuse.report import FileReport, ProjectReport
 
@@ -267,6 +268,92 @@ def test_generate_file_report_to_dict_lint_source_information(fake_repository):
             assert expression["value"] == "MIT OR 0BSD"
 
 
+def test_strict_dep5_in_file_report(fake_repository):
+    """All copyright information of a strictly formatted dep5 file should be
+    taken into account in the file report. 'Strictly formatted' here means no
+    syntax errors. Specifically, there shouldn't be any duplicate Copyright
+    tags/fields.
+    """
+    # pylint: disable=line-too-long
+    (fake_repository / ".reuse/dep5").write_text(
+        dedent(
+            """
+            Format: https://www.debian.org/doc/packaging-manuals/copyright-format/1.0/
+            Upstream-Name: Some project
+            Upstream-Contact: Jane Doe
+            Source: https://example.com/
+
+            Files: doc/*
+            Copyright: 2017 Jane Doe
+                       2018 John Doe
+                       2019 Joey Doe
+            License: CC0-1.0
+            """
+        )
+    )
+    project = Project(fake_repository)
+    report = FileReport.generate(
+        project,
+        "doc/index.rst",
+    )
+    assert len(report.reuse_infos) == 1
+    reuse_info = report.reuse_infos[0]
+    assert reuse_info.spdx_expressions == {_LICENSING.parse("CC0-1.0")}
+    assert reuse_info.copyright_lines == {
+        "2017 Jane Doe",
+        "2018 John Doe",
+        "2019 Joey Doe",
+    }
+    assert reuse_info.source_path == ".reuse/dep5"
+    assert reuse_info.source_type == SourceType.DEP5
+
+
+def test_dep5_duplicate_copyright_in_file_report(fake_repository):
+    """Duplicate 'Copyright:' fields (tags) in a DEP5 file is a syntax error.
+    Since python-debian 0.1.47, the parser raises an error on this class of
+    syntax error.
+
+    We suppress that error (by parsing the file 'non-strictly'), because this
+    error is undocumented breaking behaviour.
+
+    Towards that end, verify that no error is raised. As a side effect of
+    parsing non-strictly, only the first Copyright tag is parsed. This is
+    undesireable, but reflects pre-0.1.47 behaviour.
+
+    See also <https://github.com/fsfe/reuse-tool/issues/803>.
+
+    TODO: Expect at least a user-facing warning when this happens.
+    """
+    # pylint: disable=line-too-long
+    (fake_repository / ".reuse/dep5").write_text(
+        dedent(
+            """
+            Format: https://www.debian.org/doc/packaging-manuals/copyright-format/1.0/
+            Upstream-Name: Some project
+            Upstream-Contact: Jane Doe
+            Source: https://example.com/
+
+            Files: doc/*
+            Copyright: 2017 Jane Doe
+            Copyright: 2018 John Doe
+            Copyright: 2019 Joey Doe
+            License: CC0-1.0
+            """
+        )
+    )
+    project = Project(fake_repository)
+    report = FileReport.generate(
+        project,
+        "doc/index.rst",
+    )
+    assert len(report.reuse_infos) == 1
+    reuse_info = report.reuse_infos[0]
+    assert reuse_info.spdx_expressions == {_LICENSING.parse("CC0-1.0")}
+    assert reuse_info.copyright_lines == {"2017 Jane Doe"}
+    assert reuse_info.source_path == ".reuse/dep5"
+    assert reuse_info.source_type == SourceType.DEP5
+
+
 def test_generate_project_report_simple(fake_repository, multiprocessing):
     """Simple generate test, just to see if it sort of works."""
     project = Project(fake_repository)