fix: support .arr, support ios/wasm binaries (#4521)

* fix: support .arr, support ios/wasm binaries

* fix: aar file, apple/wasm binary

* fix: add aar to spellings

* fix: add tests for File.is_binary and VersionScanner.is_executable

---------

Co-authored-by: Terri Oda <terri.oda@intel.com>

Author: iaincoulter

.github/actions/spelling/expect.txt

@@ -13,4 +13,5 @@ NOTKNOWN
 pyyaml
 skontar
 Svunknown
-urllib
+urllib
+aar

README.md

@@ -313,6 +313,7 @@ The following archive formats are currently supported by the auto-extractor:
 | Archive Format | File Extension                                 |
 | -------------- | ---------------------------------------------- |
 | zip            | .zip, .exe, .jar, .msi, .egg, .whl, .war, .ear |
+|                | .aar                                           |
 | tar            | .tar, .tgz, .tar.gz, .tar.xz, .tar.bz2         |
 | deb            | .deb, .ipk                                     |
 | rpm            | .rpm                                           |
@@ -348,7 +349,7 @@ On windows systems, you may need:
 
 Windows has `Expand` installed by default, but `ar` and `7z` might need to be installed.
 If you want to run our test-suite or scan a zstd compressed file, We recommend installing this [7-zip-zstd](https://github.com/mcmilk/7-Zip-zstd)
-fork of 7zip. We are currently using `7z` for extracting `jar`, `apk`, `msi`, `exe` and `rpm` files.
+fork of 7zip. We are currently using `7z` for extracting `jar`, `apk`, `aar`, `msi`, `exe` and `rpm` files.
 To install `ar` you can install MinGW (which has binutils as a part of it) from [here](https://www.mingw-w64.org/downloads/#msys2) and run the downloaded .exe file.
 
 If you get an error about building libraries when you try to install from pip,

cve_bin_tool/extractor.py

@@ -90,6 +90,7 @@ def __init__(self, logger=None, error_mode=ErrorMode.TruncTrace):
                     ".whl",
                     ".war",
                     ".ear",
+                    ".aar",
                 ],
                 MIMES: [
                     "application/x-msdownload",

cve_bin_tool/file.py

@@ -47,3 +47,28 @@ def check_pe(_filename: str, signature: bytes) -> bool:
 def check_fake_test(_filename: str, signature: bytes) -> bool:
     """check for fake tests under windows."""
     return signature == b"MZ\x90\x00"
+
+
+def check_mach_o_32(_filename: str, signature: bytes) -> bool:
+    """Check for Mach-O 32-bit signature."""
+    return signature[:4] == b"\xFE\xED\xFA\xCE"
+
+
+def check_mach_o_64(_filename: str, signature: bytes) -> bool:
+    """Check for Mach-O 64-bit signature."""
+    return signature[:4] == b"\xFE\xED\xFA\xCF"
+
+
+def check_mach_o_universal(_filename: str, signature: bytes) -> bool:
+    """Check for Mach-O Universal Binary signature."""
+    return signature[:4] == b"\xCA\xFE\xBA\xBE"
+
+
+def check_ios_arm(_filename: str, signature: bytes) -> bool:
+    """Check for Mach-O Universal Binary signature."""
+    return signature[:4] == b"\xCF\xFA\xED\xFE"
+
+
+def check_wasm(_filename: str, signature: bytes) -> bool:
+    """Check for WebAssembly (WASM) signature."""
+    return signature[:4] == b"\x00\x61\x73\x6D"

cve_bin_tool/version_scanner.py

@@ -173,12 +173,17 @@ def is_executable(self, filename: str) -> tuple[bool, str | None]:
                     "PE32 executable",
                     "PE32+ executable",
                     "Mach-O",
+                    "WebAssembly",
                     "YAFFS",
                     ": data",
                     *list(valid_files.keys()),
                 )
             ):
-                return False, None
+                # Fallback as not all types are covered by the File command
+                # On linux noticed Mach-O Universal Binary not recognized
+                if not is_binary(filename):
+                    return False, None
+
         # otherwise use python implementation of file
         elif not is_binary(filename):
             return False, None

test/assets/single-byte.txt

@@ -0,0 +1,90 @@
+# Copyright (C) 2024 Iain Coulter
+# SPDX-License-Identifier: GPL-3.0-or-later
+
+"""
+CVE-bin-tool version scanner file/is_binary tests
+"""
+
+from os import remove
+from tempfile import NamedTemporaryFile
+
+from cve_bin_tool.version_scanner import VersionScanner
+
+
+class TestFile:
+    """Tests the CVE Bin Tool file using 'file' command."""
+
+    def _write_magic_signature(self, f, signature):
+        """Helper function to write a magic signature to a file."""
+        f.write(signature)
+        f.seek(0)
+
+    def _check_file_type(self, file_type, signature, expected_result):
+        """Helper function to check if a file is binary based on its type."""
+        with NamedTemporaryFile("w+b", suffix=file_type, delete=False) as f:
+            self._write_magic_signature(f, signature)
+            scanner = VersionScanner()
+            result, *_ = scanner.is_executable(f.name)
+            assert (
+                result == expected_result
+            ), f"Expected {expected_result}, but got {result}"
+        remove(f.name)
+
+    def _check_test(self, type):
+        """Helper function to parse a binary file and check whether
+        the given string is in the parsed result"""
+        file_signatures = {
+            "elf": (b"\x7f\x45\x4c\x46\x02\x01\x01\x03\n", True, ".out"),
+            "mach_o_32": (b"\xFE\xED\xFA\xCE\x00\x00\x00\x00", True, ".out"),
+            "mach_o_64": (b"\xFE\xED\xFA\xCF\x00\x00\x00\x00", True, ".out"),
+            "mach_o_universal": (b"\xCA\xFE\xBA\xBE\x00\x00\x00\x00", True, ".out"),
+            "ios_arm": (b"\xCF\xFA\xED\xFE\x00\x00\x00\x00", True, ".out"),
+            "wasm": (b"yoyo\x00\x61\x73\x6D\x01\x00\x00\x00", True, ".out"),
+            "c": (b"#include <stdio.h>", False, ".c"),
+            "single_byte": (b"1", False, ".txt"),
+            "windows": (b"MZ\x90\x00", True, ".dll"),
+        }
+        signature, expected_result, file_type = file_signatures.get(
+            type, (b"some other data\n", False, ".txt")
+        )
+        self._check_file_type(file_type, signature, expected_result)
+
+    def test_binary_elf_file(self):
+        """file *.out"""
+        self._check_test("elf")
+
+    def test_binary_mach_o_32_file(self):
+        """file *.out"""
+        self._check_test("mach_o_32")
+
+    def test_binary_mach_o_64_file(self):
+        """file *.out"""
+        self._check_test("mach_o_64")
+
+    def test_binary_mach_o_universal_file(self):
+        """file *.out"""
+        self._check_test("mach_o_universal")
+
+    def test_binary_ios_arm_file(self):
+        """file *.out"""
+        self._check_test("ios_arm")
+
+    def test_binary_wasm_file(self):
+        """file *.out"""
+        self._check_test("wasm")
+
+    def test_source_file(self):
+        """file *.c"""
+        self._check_test("c")
+
+    def test_single_byte_file(self):
+        """file single-byte"""
+        self._check_test("single_byte")
+
+    def test_windows(self):
+        """file *.txt"""
+        self._check_test("windows")
+
+    def test_other_file(self):
+        """file *.txt"""
+        self._check_test("other")

test/assets/windows.txt

@@ -4,51 +4,94 @@
 """
 CVE-bin-tool file tests
 """
-from pathlib import Path
 
 import pytest
 
 from cve_bin_tool.async_utils import NamedTemporaryFile, aio_rmfile
 from cve_bin_tool.file import aio_is_binary
 
-ASSETS_PATH = Path(__file__).parent.resolve() / "assets"
-
 
 class TestFile:
     """Tests the CVE Bin Tool file binary checker."""
 
+    async def _write_magic_signature(self, f, signature):
+        """Helper function to write a magic signature to a file."""
+        await f.write(signature)
+        await f.seek(0)
+
+    async def _check_file_type(self, file_type, signature, expected_result):
+        """Helper function to check if a file is binary based on its type."""
+        async with NamedTemporaryFile("w+b", suffix=file_type, delete=False) as f:
+            await self._write_magic_signature(f, signature)
+            assert await aio_is_binary(f.name) == expected_result
+        await aio_rmfile(f.name)
+
     @pytest.mark.asyncio
-    async def _check_test(self, file_type):
+    async def _check_test(self, type):
         """Helper function to parse a binary file and check whether
         the given string is in the parsed result"""
-        async with NamedTemporaryFile("w+b", suffix=file_type, delete=False) as f:
-            if file_type == "out":
-                # write magic signature
-                await f.write(b"\x7f\x45\x4c\x46\x02\x01\x01\x03\n")
-                await f.seek(0)
-                assert await aio_is_binary(f.name)
-            else:
-                await f.write(b"some other data\n")
-                await f.seek(0)
-                assert not await aio_is_binary(f.name)
-        await aio_rmfile(f.name)
+        file_signatures = {
+            "elf": (b"\x7f\x45\x4c\x46\x02\x01\x01\x03\n", True, ".out"),
+            "mach_o_32": (b"\xFE\xED\xFA\xCE\x00\x00\x00\x00", True, ".out"),
+            "mach_o_64": (b"\xFE\xED\xFA\xCF\x00\x00\x00\x00", True, ".out"),
+            "mach_o_universal": (b"\xCA\xFE\xBA\xBE\x00\x00\x00\x00", True, ".out"),
+            "ios_arm": (b"\xCF\xFA\xED\xFE\x00\x00\x00\x00", True, ".out"),
+            "wasm": (b"\x00\x61\x73\x6D\x01\x00\x00\x00", True, ".out"),
+            "c": (b"#include <stdio.h>", False, ".c"),
+            "single_byte": (b"1", False, ".txt"),
+            "windows": (b"MZ", True, ".txt"),
+        }
+        signature, expected_result, file_type = file_signatures.get(
+            type, (b"some other data\n", False, ".txt")
+        )
+        await self._check_file_type(file_type, signature, expected_result)
+
+    @pytest.mark.asyncio
+    async def test_binary_elf_file(self):
+        """file *.out"""
+        await self._check_test("elf")
+
+    @pytest.mark.asyncio
+    async def test_binary_mach_o_32_file(self):
+        """file *.out"""
+        await self._check_test("mach_o_32")
 
     @pytest.mark.asyncio
-    async def test_binary_out_file(self):
+    async def test_binary_mach_o_64_file(self):
         """file *.out"""
-        await self._check_test("out")
+        await self._check_test("mach_o_64")
+
+    @pytest.mark.asyncio
+    async def test_binary_mach_o_universal_file(self):
+        """file *.out"""
+        await self._check_test("mach_o_universal")
+
+    @pytest.mark.asyncio
+    async def test_binary_ios_arm_file(self):
+        """file *.out"""
+        await self._check_test("ios_arm")
+
+    @pytest.mark.asyncio
+    async def test_binary_wasm_file(self):
+        """file *.out"""
+        await self._check_test("wasm")
 
     @pytest.mark.asyncio
     async def test_source_file(self):
         """file *.c"""
         await self._check_test("c")
 
+    @pytest.mark.asyncio
+    async def test_text_file(self):
+        """file *.txt"""
+        await self._check_test("other")
+
     @pytest.mark.asyncio
     async def test_single_byte_file(self):
         """file single-byte"""
-        assert not await aio_is_binary(str(ASSETS_PATH / "single-byte.txt"))
+        await self._check_test("single_byte")
 
     @pytest.mark.asyncio
     async def test_windows(self):
-        """file single-byte"""
-        assert await aio_is_binary(str(ASSETS_PATH / "windows.txt"))
+        """file *.txt"""
+        await self._check_test("windows")