Refactor workaround for scancode evaluation of PBL

Add function "has_binary_license" to check if a file has a
non-permissive license contains one. PBL is not recognized by scancode,
causing it to be flagged as a non-permissive license. CI doesn't allow
any non-permissive licenses, although, files flageed as SPDX are
allowed. Workaround causes all files with a valid PBL to be flagged as
missing an SPDX.

Add condition in "has_spdx_text_in_scancode_output" to ignore any spdx
identifier with "unknown" in the name. Scancode erroneously matches PBL
to matched_rule.identifer "spdx-license-identifier: unknown-spdx". This
prevents the workaround from working.

Author: harrisonmutai-arm

tools/test/travis-ci/scancode-evaluate.py

@@ -30,6 +30,7 @@
 
 userlog = logging.getLogger("scancode-evaluate")
 
+
 class ReturnCode(Enum):
     """Return codes."""
 
@@ -54,13 +55,12 @@ def path_leaf(path):
     return tail or os.path.basename(head)
 
 
-def has_permissive_text_in_scancode_output(scancode_output_data_file):
-    """Returns true if at least one license in the scancode output is permissive or is a Permissive Binary License"""
-    # temporary workaround for files with Permissive Binary Licenses
+def has_permissive_text_in_scancode_output(scancode_output_data_file_licenses):
+    """Returns true if at least one license in the scancode output is permissive"""
     return any(
         scancode_output_data_file_license['category'] == 'Permissive'
-        for scancode_output_data_file_license in scancode_output_data_file['licenses']
-    ) or has_binary_license(scancode_output_data_file)
+        for scancode_output_data_file_license in scancode_output_data_file_licenses
+    )
 
 
 def has_spdx_text_in_scancode_output(scancode_output_data_file_licenses):
@@ -76,16 +76,20 @@ def has_spdx_text_in_analysed_file(scanned_file_content):
     return bool(re.findall("SPDX-License-Identifier:?", scanned_file_content))
 
 
-def has_binary_license(scancode_output_data_file):
+def has_binary_license(scanned_file_content):
     """Returns true if the file analysed by ScanCode contains a Permissive Binary License."""
+    return bool(re.findall("Permissive Binary License", scanned_file_content))
+
+
+def get_file_text(scancode_output_data_file):
+    """Returns file text for scancode output file"""
     file_path = os.path.abspath(scancode_output_data_file['path'])
     try:
         with open(file_path, 'r') as read_file:
-            scanned_file_content = read_file.read()
-        return bool(re.findall("Permissive Binary License", scanned_file_content))
+            return read_file.read()
     except UnicodeDecodeError:
-        userlog.warning("Unable to look for PBL text in `{}`:".format(file_path))
-        return False
+        userlog.warning("Unable to decode file text in: %s" % file_path)
+        # Ignore files that cannot be decoded
 
 
 def license_check(scancode_output_path):
@@ -98,7 +102,7 @@ def license_check(scancode_output_path):
 
     Returns:
     0 if nothing found
-    >0 - count how many license isses found
+    >0 - count how many license issues found
     ReturnCode.ERROR.value if any error in file licenses found
     """
 
@@ -125,25 +129,21 @@ def license_check(scancode_output_path):
             # check the next file in the scancode output
             continue
 
-        if not has_permissive_text_in_scancode_output(scancode_output_data_file):
-            scancode_output_data_file['fail_reason'] = MISSING_PERMISSIVE_LICENSE_TEXT
-            license_offenders.append(scancode_output_data_file)
+        if not has_permissive_text_in_scancode_output(scancode_output_data_file['licenses']):
+            scanned_file_content = get_file_text(scancode_output_data_file)
+            if not (scanned_file_content and has_binary_license(scanned_file_content)):
+                scancode_output_data_file['fail_reason'] = MISSING_PERMISSIVE_LICENSE_TEXT
+                license_offenders.append(scancode_output_data_file)
 
         if not has_spdx_text_in_scancode_output(scancode_output_data_file['licenses']):
             # Scancode does not recognize license notice in Python file headers.
             # Issue: https://github.com/nexB/scancode-toolkit/issues/1913
             # Therefore check if the file tested by ScanCode actually has a licence notice.
-            file_path = os.path.abspath(scancode_output_data_file['path'])
-            try:
-                with open(file_path, 'r') as read_file:
-                    scanned_file_content = read_file.read()
-            except UnicodeDecodeError:
-                userlog.warning("Unable to look for SPDX text in `{}`:".format(file_path))
-                # Ignore files that cannot be decoded
-                # check the next file in the scancode output
-                continue
+            scanned_file_content = get_file_text(scancode_output_data_file)
 
-            if not has_spdx_text_in_analysed_file(scanned_file_content):
+            if not scanned_file_content:
+                continue
+            elif not has_spdx_text_in_analysed_file(scanned_file_content):
                 scancode_output_data_file['fail_reason'] = MISSING_SPDX_TEXT
                 spdx_offenders.append(scancode_output_data_file)
 

tools/test/travis-ci/scancode_evaluate_test.py

@@ -34,7 +34,7 @@
  * limitations under the License.\
  */"
 
-BINARY_HEADER = "/*\
+HEADER_WITH_BINARY_LICENSE = "/*\
  * Copyright (c) 2019, Arm Limited, All Rights Reserved\
  * SPDX-License-Identifier: LicenseRef-PBL\
  *\
@@ -63,7 +63,7 @@ def create_scanned_files():
             if file_path in [os.path.join(STUBS_PATH, "test3.h")]:
                 new_file.write(HEADER_WITHOUT_SPDX)
             elif file_path in [os.path.join(STUBS_PATH, "test6.h")]:
-                new_file.write(BINARY_HEADER)
+                new_file.write(HEADER_WITH_BINARY_LICENSE)
             else:
                 new_file.write(HEADER_WITH_SPDX)
     yield
@@ -95,7 +95,7 @@ def test_missing_license_permissive_license_and_spdx(self, create_scanned_files)
             test3.h: Missing `Permissive` license text and `spdx` in match.identifier and not in file tested by ScanCode (error count += 1)
             test4.h: Missing `Permissive` license text and `spdx` in match.identifier but found in file tested by ScanCode (error count += 1)
             test5.h: Missing `spdx` in match.identifier but found in file tested by ScanCode. (error count += 0)
-            test6.h: Matching 'unknown-spdx' in match.identifier and Permissive Binary License in header (error count += 1)
+            test6.h: Matching `spdx` in match.identifier but Permissive Binary License header (error count += 0)
             @inputs scancode_test/scancode_test_2.json
             @output 3
         """
@@ -107,4 +107,4 @@ def test_permissive_license_no_spdx(self, create_scanned_files):
             @inputs scancode_test/scancode_test_2.json
             @outputs 0
         """
-        assert license_check(os.path.join(STUBS_PATH, "scancode_test_4.json")) == 0
+        assert license_check(os.path.join(STUBS_PATH, "scancode_test_4.json")) == 0