tests/fuzz.py

+import mimetypes
+import os
+import sys
+
+sys.path.append('..')
+
+import atheris
+
+with atheris.instrument_imports(enable_loader_override=False):
+    from libmat2 import parser_factory, UNSUPPORTED_EXTENSIONS
+
+extensions = set()
+for parser in parser_factory._get_parsers():  # type: ignore
+    for mtype in parser.mimetypes:
+        if mtype.startswith('video'):
+            continue
+        if 'aif' in mtype:
+            continue
+        if 'wav' in mtype:
+            continue
+        if 'gif' in mtype:
+            continue
+        if 'aifc' in mtype:
+            continue
+        for extension in mimetypes.guess_all_extensions(mtype):
+            if extension not in UNSUPPORTED_EXTENSIONS:
+                extensions.add(extension)
+extensions = list(extensions)
+
+
+
+def TestOneInput(data):
+    fdp = atheris.FuzzedDataProvider(data)
+    extension = fdp.PickValueInList(extensions)
+    data = fdp.ConsumeBytes(sys.maxsize)
+
+    fname = '/tmp/mat2_fuzz' + extension
+
+    with open(fname, 'wb') as f:
+        f.write(data)
+    try:
+        p, _ = parser_factory.get_parser(fname)
+        if p:
+            p.sandbox = False
+            p.get_meta()
+            p.remove_all()
+            p, _ = parser_factory.get_parser(fname)
+            p.get_meta()
+    except ValueError:
+        pass
+    os.remove(fname)
+
+atheris.Setup(sys.argv, TestOneInput)
+atheris.Fuzz()

tests/test_policy.py

+#!/usr/bin/env python3
+
+import unittest
+import shutil
+import os
+
+from libmat2 import office, UnknownMemberPolicy
+
+class TestPolicy(unittest.TestCase):
+    target = './tests/data/clean.docx'
+
+    def test_policy_omit(self):
+        shutil.copy('./tests/data/embedded.docx', self.target)
+        p = office.MSOfficeParser(self.target)
+        p.unknown_member_policy = UnknownMemberPolicy.OMIT
+        self.assertTrue(p.remove_all())
+        os.remove(p.filename)
+
+    def test_policy_keep(self):
+        shutil.copy('./tests/data/embedded.docx', self.target)
+        p = office.MSOfficeParser(self.target)
+        p.unknown_member_policy = UnknownMemberPolicy.KEEP
+        self.assertTrue(p.remove_all())
+        os.remove(p.filename)
+        os.remove(p.output_filename)
+
+    def test_policy_unknown(self):
+        shutil.copy('./tests/data/embedded.docx', self.target)
+        p = office.MSOfficeParser(self.target)
+        with self.assertRaises(ValueError):
+            p.unknown_member_policy = UnknownMemberPolicy('unknown_policy_name_totally_invalid')
+        os.remove(p.filename)