blah

Cisco-Talos · Jul 10, 2024 · 594f4b1 · 594f4b1
1 parent afc233d
commit 594f4b1
Show file tree

Hide file tree

Showing 2 changed files with 92 additions and 0 deletions.
diff --git a/unit_tests/clamscan/save_html_urls_test.py b/unit_tests/clamscan/save_html_urls_test.py
@@ -0,0 +1,76 @@
+# Copyright (C) 2020-2024 Cisco Systems, Inc. and/or its affiliates. All rights reserved.
+
+"""
+Run clamscan tests.
+"""
+
+import sys
+import os
+import re
+import shutil
+
+sys.path.append('../unit_tests')
+import testcase
+
+
+class TC(testcase.TestCase):
+    @classmethod
+    def setUpClass(cls):
+        super(TC, cls).setUpClass()
+
+    @classmethod
+    def tearDownClass(cls):
+        super(TC, cls).tearDownClass()
+
+    def setUp(self):
+        super(TC, self).setUp()
+
+    def tearDown(self):
+        super(TC, self).tearDown()
+
+        # Remove scan temps directory between tests
+        if (self.path_tmp / "TD").exists():
+            shutil.rmtree(self.path_tmp / "TD")
+
+        self.verify_valgrind_log()
+
+    # Find the metadata.json file and verify its contents.
+    #TODO: REMOVE THIS WHEN https://github.com/Cisco-Talos/clamav/pull/1295 is merged
+    def verify_metadata_json(self, tempdir, expected=[], unexpected=[]):
+        for parent, dirs, files in os.walk(tempdir):
+            for f in files:
+                if "metadata.json" == f:
+                    with open(os.path.join(parent, f)) as handle:
+                        metadata_json = handle.read()
+                        self.verify_output(metadata_json, expected=expected, unexpected=unexpected)
+
+                    # There is only one metadata.json per scan.
+                    # We found it, so we can break out of the loop.
+                    break
+
+    def test_save_links(self):
+        self.step_name('Extract Links')
+
+        tempdir=self.path_tmp / "TD"
+        if not os.path.isdir(tempdir):
+            os.makedirs(tempdir);
+
+        testfile = TC.path_source / 'unit_tests' / 'input' / 'other_scanfiles' / 'html' / 'index.html'
+        command = '{valgrind} {valgrind_args} {clamscan} -d {path_db} --gen-json --leave-temps --tempdir={tempdir} {testfile}'.format(
+            valgrind=TC.valgrind, valgrind_args=TC.valgrind_args, clamscan=TC.clamscan,
+            path_db=TC.path_source / 'unit_tests' / 'input' / 'other_sigs' / 'Clamav-Unit-Test-Signature.ndb',
+            tempdir=tempdir,
+            testfile=testfile,
+        )
+        output = self.execute_command(command)
+
+        assert output.ec == 0  # clean
+
+        expected_strings = [ 'HTMLUrls'
+                , '"https://www.clamav.net/reports/malware"'
+                , '"http://www.google.com"'
+                ]
+        self.verify_metadata_json(tempdir, expected_strings)
+
+
+
diff --git a/unit_tests/input/other_scanfiles/html/index.html b/unit_tests/input/other_scanfiles/html/index.html
@@ -0,0 +1,16 @@
+<!DOCTYPE html>
+<html>
+<body>
+
+<h1>Save Links Unittest</h1>
+<p>Paragraph</p>
+<a href="https://www.clamav.net/reports/malware">Report Malware</a>
+
+<form action="http://www.google.com">
+    <input type="submit">
+</form>
+
+
+</body>
+</html>
+