Add tests for better handling of symlinks

These tests assume the interpretation of the REUSE specification that:
1. a symlink pointing to a covered file is considered to be the same
   file as the covered file and can therefore be ignored.
2. a symlink pointing to a file that is not a covered file is itself
   considered to be a covered file and should not be ignored, unless the
   symlink itself is ignored by other means.

tests/conftest.py

@@ -1,6 +1,7 @@
 # SPDX-FileCopyrightText: 2017 Free Software Foundation Europe e.V. <https://fsfe.org>
 # SPDX-FileCopyrightText: 2022 Florian Snow <[email protected]>
 # SPDX-FileCopyrightText: 2022 Carmen Bianca Bakker <[email protected]>
+# SPDX-FileCopyrightText: 2023 Matthias Riße
 #
 # SPDX-License-Identifier: GPL-3.0-or-later
 
@@ -137,6 +138,15 @@ def fake_repository(tmpdir_factory) -> Path:
         encoding="utf-8",
     )
 
+    (directory / "symlink-to-covered").symlink_to(directory / "doc/index.rst")
+    (directory / "symlink-to-not-covered").symlink_to(directory)
+    (directory / "symlink-to-not-covered.license").write_text(
+        "# SPDX-FileCopyrightText: 2017 Jane Doe\n"
+        "#\n"
+        "# SPDX-License-Identifier: GPL-3.0-or-later",
+        encoding="utf-8",
+    )
+
     os.chdir(directory)
     return directory
 

tests/test_project.py

@@ -1,6 +1,7 @@
 # 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 <[email protected]>
+# SPDX-FileCopyrightText: 2023 Matthias Riße
 #
 # SPDX-License-Identifier: GPL-3.0-or-later
 
@@ -98,8 +99,8 @@ def test_all_files_ignore_hg(empty_directory):
 
 
 @posix
-def test_all_files_symlinks(empty_directory):
-    """All symlinks must be ignored."""
+def test_all_files_ignore_symlinks_to_covered_file(empty_directory):
+    """All symlinks to covered files must be ignored."""
     (empty_directory / "blob").write_text("foo")
     (empty_directory / "blob.license").write_text(
         cleandoc(
@@ -115,6 +116,15 @@ def test_all_files_symlinks(empty_directory):
     assert Path("symlink").absolute() not in project.all_files()
 
 
+@posix
+@pytest.mark.parametrize("target", ["/outside_file", "non-existent-file"])
+def test_all_files_cover_symlinks_to_uncovered_files(empty_directory, target):
+    """All symlinks to files not covered must be included."""
+    (empty_directory / "symlink").symlink_to(target)
+    project = Project(empty_directory)
+    assert Path("symlink").absolute() in project.all_files()
+
+
 def test_all_files_ignore_zero_sized(empty_directory):
     """Empty files should be skipped."""
     (empty_directory / "foo").touch()
@@ -157,6 +167,33 @@ def test_all_files_git_ignored_contains_newline(git_repository):
     assert Path("hello\nworld.pyc").absolute() not in project.all_files()
 
 
+@posix
+def test_all_files_git_ignore_symlinks_to_covered_file(git_repository):
+    """All symlinks to covered files must be ignored."""
+    (git_repository / "symlink").symlink_to("doc/index.rst")
+    project = Project(git_repository)
+    assert Path("symlink").absolute() not in project.all_files()
+
+
+@posix
+@pytest.mark.parametrize(
+    "target",
+    [
+        "/outside_file",
+        ".git/file-in-dotgit",
+        "build/hello.py",
+        "non-existent-file",
+    ],
+)
+def test_all_files_git_cover_symlinks_to_uncovered_files(
+    git_repository, target
+):
+    """All symlinks to files not covered must be included."""
+    (git_repository / "symlink").symlink_to(target)
+    project = Project(git_repository)
+    assert Path("symlink").absolute() in project.all_files()
+
+
 def test_all_files_submodule_is_ignored(submodule_repository):
     """If a submodule is ignored, all_files should not raise an Exception."""
     (submodule_repository / "submodule/foo.py").write_text("foo")
@@ -202,6 +239,31 @@ def test_all_files_hg_ignored_contains_newline(hg_repository):
     assert Path("hello\nworld.pyc").absolute() not in project.all_files()
 
 
+@posix
+def test_all_files_hg_ignore_symlinks_to_covered_file(hg_repository):
+    """All symlinks to covered files must be ignored."""
+    (hg_repository / "symlink").symlink_to("doc/index.rst")
+    project = Project(hg_repository)
+    assert Path("symlink").absolute() not in project.all_files()
+
+
+@posix
+@pytest.mark.parametrize(
+    "target",
+    [
+        "/outside_file",
+        ".hg/file-in-dothg",
+        "build/hello.py",
+        "non-existent-file",
+    ],
+)
+def test_all_files_hg_cover_symlinks_to_uncovered_files(hg_repository, target):
+    """All symlinks to files not covered must be included."""
+    (hg_repository / "symlink").symlink_to(target)
+    project = Project(hg_repository)
+    assert Path("symlink").absolute() in project.all_files()
+
+
 def test_reuse_info_of_file_does_not_exist(fake_repository):
     """Raise FileNotFoundError when asking for the REUSE info of a file that
     does not exist.