[beman-tidy] Implement LICENSE.APPROVED

tools/beman-tidy/beman_tidy/lib/checks/beman_standard/license.py

@@ -1,7 +1,12 @@
 #!/usr/bin/env python3
 # SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
 
+import re
+import textwrap
+
 from ..base.file_base_check import FileBaseCheck
+from ..system.registry import register_beman_standard_check
+
 
 # [LICENSE.*] checks category.
 # All checks in this file extend the LicenseBaseCheck class.
@@ -14,7 +19,55 @@ def __init__(self, repo_info, beman_standard_check_config):
         super().__init__(repo_info, beman_standard_check_config, "LICENSE")
 
 
-# TODO LICENSE.APPROVED
+@register_beman_standard_check("LICENSE.APPROVED")
+class LicenseApprovedCheck(LicenseBaseCheck):
+    def __init__(self, repo_info, beman_standard_check_config):
+        super().__init__(repo_info, beman_standard_check_config)
+
+    def check(self):
+        content = self.read()
+
+        # Regex for LICENSE check
+        # - fixed header matching
+        # - non empty body matching
+        # - fixed footer matching
+        regex = re.compile(
+            textwrap.dedent(r"""
+                ^={78}
+                The Beman Project is under the (Apache License v2\.0 with LLVM Exceptions|Boost Software License 1\.0|MIT License):
+                ={78}
+
+                (.+?)
+
+                ={78}
+                Software from third parties included in the Beman Project:
+                ={78}
+                The Beman Project contains third party software which is under different license
+                terms\. All such code will be identified clearly using at least one of two
+                mechanisms:
+                1\) It will be in a separate directory tree with its own `LICENSE\.txt` or
+                   `LICENSE` file at the top containing the specific license and restrictions
+                   which apply to that software, or
+                2\) It will contain specific license and restriction terms at the top of every
+                   file\.
+                """).strip(),
+            re.DOTALL,
+        )
+
+        if not regex.match(content):
+            self.log(
+                "LICENSE file does not match the required format. "
+                "See https://github.com/bemanproject/beman/blob/main/docs/BEMAN_STANDARD.md#licenseapproved for more information."
+            )
+            return False
+
+        return True
+
+    def fix(self):
+        self.log(
+            "Please update the LICENSE file to include an approved license. "
+            "See https://github.com/bemanproject/beman/blob/main/docs/BEMAN_STANDARD.md#licenseapproved for more information."
+        )
 
 
 # TODO LICENSE.APACHE_LLVM

tools/beman-tidy/tests/lib/checks/beman_standard/license/conftest.py

@@ -0,0 +1,16 @@
+#!/usr/bin/env python3
+# SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+
+import pytest
+
+from tests.utils.conftest import mock_repo_info, mock_beman_standard_check_config  # noqa: F401
+
+
+@pytest.fixture(autouse=True)
+def repo_info(mock_repo_info):  # noqa: F811
+    return mock_repo_info
+
+
+@pytest.fixture
+def beman_standard_check_config(mock_beman_standard_check_config):  # noqa: F811
+    return mock_beman_standard_check_config

tools/beman-tidy/tests/lib/checks/beman_standard/license/data/invalid/invalid-LICENSE-v1

@@ -0,0 +1,14 @@
+# Dummy LICENSE lines.
+# beman-tidy does not actually check the content between the header and the footer.
+
+==============================================================================
+Software from third parties included in the Beman Project:
+==============================================================================
+The Beman Project contains third party software which is under different license
+terms. All such code will be identified clearly using at least one of two
+mechanisms:
+1) It will be in a separate directory tree with its own `LICENSE.txt` or
+   `LICENSE` file at the top containing the specific license and restrictions
+   which apply to that software, or
+2) It will contain specific license and restriction terms at the top of every
+   file.

tools/beman-tidy/tests/lib/checks/beman_standard/license/data/invalid/invalid-LICENSE-v2

@@ -0,0 +1,6 @@
+==============================================================================
+The Beman Project is under the Apache License v2.0 with LLVM Exceptions:
+==============================================================================
+
+# Dummy LICENSE lines.
+# beman-tidy does not actually check the content between the header and the footer.

tools/beman-tidy/tests/lib/checks/beman_standard/license/data/invalid/invalid-LICENSE-v3

@@ -0,0 +1,15 @@
+==============================================================================
+The Beman Project is under the Apache License v2.0 with LLVM Exceptions:
+==============================================================================
+
+==============================================================================
+Software from third parties included in the Beman Project:
+==============================================================================
+The Beman Project contains third party software which is under different license
+terms. All such code will be identified clearly using at least one of two
+mechanisms:
+1) It will be in a separate directory tree with its own `LICENSE.txt` or
+   `LICENSE` file at the top containing the specific license and restrictions
+   which apply to that software, or
+2) It will contain specific license and restriction terms at the top of every
+   file.

tools/beman-tidy/tests/lib/checks/beman_standard/license/data/invalid/invalid-LICENSE-v4

@@ -0,0 +1,18 @@
+==============================================================================
+The "Extra" Beman Project is under the Apache License v2.0 with LLVM Exceptions:
+==============================================================================
+
+# Dummy LICENSE lines.
+# beman-tidy does not actually check the content between the header and the footer.
+
+==============================================================================
+Software from third parties included in the Beman Project:
+==============================================================================
+The Beman Project contains third party software which is under different license
+terms. All such code will be identified clearly using at least one of two
+mechanisms:
+1) It will be in a separate directory tree with its own `LICENSE.txt` or
+   `LICENSE` file at the top containing the specific license and restrictions
+   which apply to that software, or
+2) It will contain specific license and restriction terms at the top of every
+   file.

tools/beman-tidy/tests/lib/checks/beman_standard/license/data/invalid/invalid-LICENSE-v5

@@ -0,0 +1,18 @@
+==============================================================================
+The Beman Project is under the Apache License v2.0 with LLVM Exceptions:
+==============================================================================
+
+# Dummy LICENSE lines.
+# beman-tidy does not actually check the content between the header and the footer.
+
+==============================================================================
+"Extra" Software from third parties included in the Beman Project:
+==============================================================================
+The Beman Project contains third party software which is under different license
+terms. All such code will be identified clearly using at least one of two
+mechanisms:
+1) It will be in a separate directory tree with its own `LICENSE.txt` or
+   `LICENSE` file at the top containing the specific license and restrictions
+   which apply to that software, or
+2) It will contain specific license and restriction terms at the top of every
+   file.

tools/beman-tidy/tests/lib/checks/beman_standard/license/data/valid/valid-LICENSE-v1

@@ -0,0 +1,18 @@
+==============================================================================
+The Beman Project is under the Apache License v2.0 with LLVM Exceptions:
+==============================================================================
+
+# Dummy LICENSE lines.
+# beman-tidy does not actually check the content between the header and the footer.
+
+==============================================================================
+Software from third parties included in the Beman Project:
+==============================================================================
+The Beman Project contains third party software which is under different license
+terms. All such code will be identified clearly using at least one of two
+mechanisms:
+1) It will be in a separate directory tree with its own `LICENSE.txt` or
+   `LICENSE` file at the top containing the specific license and restrictions
+   which apply to that software, or
+2) It will contain specific license and restriction terms at the top of every
+   file.

tools/beman-tidy/tests/lib/checks/beman_standard/license/data/valid/valid-LICENSE-v2

@@ -0,0 +1,18 @@
+==============================================================================
+The Beman Project is under the Boost Software License 1.0:
+==============================================================================
+
+# Dummy LICENSE lines.
+# beman-tidy does not actually check the content between the header and the footer.
+
+==============================================================================
+Software from third parties included in the Beman Project:
+==============================================================================
+The Beman Project contains third party software which is under different license
+terms. All such code will be identified clearly using at least one of two
+mechanisms:
+1) It will be in a separate directory tree with its own `LICENSE.txt` or
+   `LICENSE` file at the top containing the specific license and restrictions
+   which apply to that software, or
+2) It will contain specific license and restriction terms at the top of every
+   file.

tools/beman-tidy/tests/lib/checks/beman_standard/license/data/valid/valid-LICENSE-v3

@@ -0,0 +1,18 @@
+==============================================================================
+The Beman Project is under the MIT License:
+==============================================================================
+
+# Dummy LICENSE lines.
+# beman-tidy does not actually check the content between the header and the footer.
+
+==============================================================================
+Software from third parties included in the Beman Project:
+==============================================================================
+The Beman Project contains third party software which is under different license
+terms. All such code will be identified clearly using at least one of two
+mechanisms:
+1) It will be in a separate directory tree with its own `LICENSE.txt` or
+   `LICENSE` file at the top containing the specific license and restrictions
+   which apply to that software, or
+2) It will contain specific license and restriction terms at the top of every
+   file.

tools/beman-tidy/tests/lib/checks/beman_standard/license/test_license.py

@@ -0,0 +1,71 @@
+#!/usr/bin/env python3
+# SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+
+import pytest
+from pathlib import Path
+
+from tests.utils.path_runners import (
+    run_check_for_each_path,
+)
+
+# Actual tested checks.
+from beman_tidy.lib.checks.beman_standard.license import (
+    LicenseApprovedCheck,
+)
+
+test_data_prefix = "tests/lib/checks/beman_standard/license/data"
+valid_prefix = f"{test_data_prefix}/valid"
+invalid_prefix = f"{test_data_prefix}/invalid"
+
+
+def test__LICENSE_APPROVED__valid(repo_info, beman_standard_check_config):
+    """
+    Test that a valid LICENSE file passes the check.
+    """
+    valid_license_paths = [
+        # Apache License v2.0 with LLVM Exceptions
+        Path(f"{valid_prefix}/valid-LICENSE-v1"),
+        # Boost Software License 1.0
+        Path(f"{valid_prefix}/valid-LICENSE-v2"),
+        # MIT License
+        Path(f"{valid_prefix}/valid-LICENSE-v3"),
+    ]
+
+    run_check_for_each_path(
+        True,
+        valid_license_paths,
+        LicenseApprovedCheck,
+        repo_info,
+        beman_standard_check_config,
+    )
+
+
+def test__LICENSE_APPROVED__invalid(repo_info, beman_standard_check_config):
+    """
+    Test that an invalid LICENSE file fails the check.
+    """
+    invalid_license_paths = [
+        # Almost valid LICENSE, but without header
+        Path(f"{invalid_prefix}/invalid-LICENSE-v1"),
+        # Almost valid LICENSE, but without footer
+        Path(f"{invalid_prefix}/invalid-LICENSE-v2"),
+        # Almost valid LICENSE, but not content between header and footer
+        Path(f"{invalid_prefix}/invalid-LICENSE-v3"),
+        # Almost valid LICENSE, but not exact header
+        Path(f"{invalid_prefix}/invalid-LICENSE-v4"),
+        # Almost valid LICENSE, but not exact footer
+        Path(f"{invalid_prefix}/invalid-LICENSE-v5"),
+    ]
+
+    run_check_for_each_path(
+        False,
+        invalid_license_paths,
+        LicenseApprovedCheck,
+        repo_info,
+        beman_standard_check_config,
+    )
+
+
+@pytest.mark.skip(reason="NOT implemented")
+def test__LICENSE_APPROVED__fix_inplace(repo_info, beman_standard_check_config):
+    pass