fix: Avoid failing on files with aliased mimetypes.

.travis.yml

@@ -21,9 +21,9 @@ addons:
 
 install:
     # General dependencies
-    - sudo add-apt-repository "deb http://archive.ubuntu.com/ubuntu/ trusty multiverse" && sudo add-apt-repository "deb http://archive.ubuntu.com/ubuntu/ trusty-updates multiverse"
     - sudo apt-get update -qq
-    - sudo apt-get install -y p7zip-rar python-pip
+    - sudo apt-get -y dist-upgrate
+    - sudo apt-get install -y p7zip-rar
     # filecheck.py dependencies
     - sudo apt-get install libxml2-dev libxslt1-dev
     - wget https://didierstevens.com/files/software/pdfid_v0_2_5.zip

Pipfile

@@ -18,6 +18,7 @@ pillow = "*"
 olefile = "*"
 oletools = "*"
 officedissector = {editable = true,git = "https://github.com/grierforensics/officedissector.git"}
+kittengroomer = {editable = true,path = "."}
 
 [requires]
 python_version = "3.6"

Pipfile.lock

@@ -1,7 +1,7 @@
 {
     "_meta": {
         "hash": {
-            "sha256": "07933a9746ecbc04f9f0c39db78c51b6e6f2f2ebc8d10e7ccf67d0c9bbb0317d"
+            "sha256": "ca76d07407e5ed9ca95b7da07b5ef20f0d7c9345b1c35536e896d1a0e8e1787f"
         },
         "pipfile-spec": 6,
         "requires": {
@@ -24,6 +24,10 @@
             "index": "pypi",
             "version": "==2.1.2"
         },
+        "kittengroomer": {
+            "editable": true,
+            "path": "."
+        },
         "lxml": {
             "hashes": [
                 "sha256:0358b9e9642bc7d39aac5cffe9884a99a5ca68e5e2c1b89e570ed60da9139908",
@@ -117,6 +121,13 @@
                 "sha256:f6c5ef0d7480ad048c054c37632c67fca55299990fff127850181659eea33fc3"
             ],
             "version": "==2.3.1"
+        },
+        "python-magic": {
+            "hashes": [
+                "sha256:f2674dcfad52ae6c49d4803fa027809540b130db1dec928cfbb9240316831375",
+                "sha256:f3765c0f582d2dfc72c15f3b5a82aecfae9498bd29ca840d72f37d7bd38bfcd5"
+            ],
+            "version": "==0.4.15"
         }
     },
     "develop": {
@@ -249,7 +260,6 @@
                 "sha256:f2674dcfad52ae6c49d4803fa027809540b130db1dec928cfbb9240316831375",
                 "sha256:f3765c0f582d2dfc72c15f3b5a82aecfae9498bd29ca840d72f37d7bd38bfcd5"
             ],
-            "index": "pypi",
             "version": "==0.4.15"
         },
         "pyyaml": {

filecheck/filecheck.py

@@ -184,15 +184,18 @@ class File(FileBase):
             self.make_dangerous('File has no extension')
         else:
             if self.extension in Config.override_ext:
-                expected_mimetype = Config.override_ext[self.extension]
+                expected_mimetypes = Config.override_ext[self.extension]
             else:
                 expected_mimetype, encoding = mimetypes.guess_type(self.src_path,
                                                                    strict=False)
+
+                expected_mimetypes = [expected_mimetype]
                 if expected_mimetype in Config.aliases:
-                    expected_mimetype = Config.aliases[expected_mimetype]
+                    expected_mimetypes.append(Config.aliases[expected_mimetype])
+
             is_known_extension = self.extension in mimetypes.types_map.keys()
-            if is_known_extension and expected_mimetype != self.mimetype:
+            if is_known_extension and self.mimetype not in expected_mimetypes:
-                self.make_dangerous('Mimetype does not match expected mimetype ({}) for this extension'.format(expected_mimetype))
+                self.make_dangerous('Mimetype does not match expected mimetypes ({}) for this extension'.format(expected_mimetypes))
 
     def _check_mimetype(self):
         """
@@ -210,6 +213,8 @@ class File(FileBase):
                 mimetype = self.mimetype
             expected_extensions = mimetypes.guess_all_extensions(mimetype,
                                                                  strict=False)
+            if mimetype in Config.aliases:
+                expected_extensions += mimetypes.guess_all_extensions(Config.aliases[mimetype], strict=False)
             if expected_extensions:
                 if self.has_extension and self.extension not in expected_extensions:
                     self.make_dangerous('Extension does not match expected extensions ({}) for this mimetype'.format(expected_extensions))