fix: friendly message for upx (gh-12, gh-35)

fix(detect & shot): not necessarily starts with b"PY00" e.g. PY01xxx

.github/workflows/regtest.yml

@@ -3,7 +3,7 @@
 name: Regression Test
 on:
   push:
-    branches: [main]
+    branches: [main, update-tests]
 
 jobs:
   regtest:

oneshot/detect.py

@@ -49,12 +49,12 @@ def find_data_from_bytes(data: bytes, max_count=-1) -> List[bytes]:
     result = []
     idx = 0
     while len(result) != max_count:
-        idx = data.find(b"PY00")  # XXX: not necessarily starts with b"PY"
+        idx = data.find(b"PY0")  # XXX: not necessarily starts with b"PY"
         if idx == -1:
             break
         data = data[idx:]
         if len(data) < 64:
-            # don't break if len > 64, maybe there is PY00blahPY000000
+            # don't break if len > 64, maybe there is PY0blahPY0
             break
         header_len = dword(data, 28)
         body_len = dword(data, 32)
@@ -86,7 +86,7 @@ def find_data_from_bytes(data: bytes, max_count=-1) -> List[bytes]:
 def nuitka_package(
     head: bytes, relative_path: str
 ) -> Union[List[Tuple[str, bytes]], None]:
-    first_occurrence = head.find(b"PY00")
+    first_occurrence = head.find(b"PY0")
     if first_occurrence == -1:
         return None
     last_dot_bytecode = head.rfind(b".bytecode\x00", 0, first_occurrence)

oneshot/runtime.py

@@ -66,9 +66,21 @@ class RuntimeInfo:
         """
         with open(self.file_path, "rb") as f:
             data = f.read(16 * 1024 * 1024)
-        cur = data.index(b"pyarmor-vax")
+        cur = data.find(b"pyarmor-vax")
+        if cur == -1:
+            # Specially, check UPX (GH-12, GH-35)
+            if data.find(b"UPX!") != -1 and data.find(b"UPX0") != -1:
+                logger.error(
+                    f"{self.file_path} seems to be packed by UPX. Before it can be processed, you need to unpack it first: Download UPX from https://github.com/upx/upx, and run `upx -d {self.file_path}` (you may need to escape the file path) in the command line."
+                )
+            else:
+                logger.error(
+                    f"{self.file_path} does not contain 'pyarmor-vax'. Maybe it's packed, obfuscated, or generated by an unsupported version of Pyarmor."
+                )
+            raise ValueError(f"{self.file_path} does not contain 'pyarmor-vax'")
 
         if data[cur + 11 : cur + 18] == b"\x00" * 7:
+            # Do not log. Skip this file silently and find another.
             raise ValueError(f"{self.file_path} is a runtime template")
 
         # Align with pyd file and executable address:
@@ -78,7 +90,7 @@ class RuntimeInfo:
 
         if data[0x5C] & 1 != 0:
             logger.error(
-                'External key file ".pyarmor.ikey" is not supported yet, but it will be supported once we get a sample (like this one). Please open an issue on https://github.com/Lil-House/Pyarmor-Static-Unpack-1shot/issues to make this tool stronger.'
+                f'External key file ".pyarmor.ikey" is not supported yet, but it will be supported once we get a sample (like this one). Please open an issue on https://github.com/Lil-House/Pyarmor-Static-Unpack-1shot/issues to make this tool stronger. ({self.file_path})'
             )
             raise NotImplementedError(f'{self.file_path} uses ".pyarmor.ikey"')
 

oneshot/shot.py

@@ -401,16 +401,16 @@ def main():
 
     if not args.no_banner:
         print(rf"""{Fore.CYAN}
- ____                                                                     ____ 
+ ____                                                                     ____
 ( __ )                                                                   ( __ )
- |  |~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~|  | 
- |  |   ____                                      _ ___  _          _     |  | 
- |  |  |  _ \ _  _  __ _ _ __ _ _ __   ___  _ _  / / __|| |_   ___ | |_   |  | 
- |  |  | |_) | || |/ _` | '__| ' `  \ / _ \| '_| | \__ \| ' \ / _ \| __|  |  | 
- |  |  |  __/| || | (_| | |  | || || | (_) | |   | |__) | || | (_) | |_   |  | 
- |  |  |_|    \_, |\__,_|_|  |_||_||_|\___/|_|   |_|___/|_||_|\___/ \__|  |  | 
- |  |         |__/                                                        |  | 
- |__|~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~|__| 
+ |  |~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~|  |
+ |  |   ____                                      _ ___  _          _     |  |
+ |  |  |  _ \ _  _  __ _ _ __ _ _ __   ___  _ _  / / __|| |_   ___ | |_   |  |
+ |  |  | |_) | || |/ _` | '__| ' `  \ / _ \| '_| | \__ \| ' \ / _ \| __|  |  |
+ |  |  |  __/| || | (_| | |  | || || | (_) | |   | |__) | || | (_) | |_   |  |
+ |  |  |_|    \_, |\__,_|_|  |_||_||_|\___/|_|   |_|___/|_||_|\___/ \__|  |  |
+ |  |         |__/                                                        |  |
+ |__|~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~|__|
 (____)                                                        v0.2.2     (____)
 
               For technology exchange only. Use at your own risk.
@@ -482,7 +482,7 @@ def main():
             if file_name.endswith(".pyz"):
                 with open(file_path, "rb") as f:
                     head = f.read(16 * 1024 * 1024)
-                if b"PY00" in head and (
+                if b"PY0" in head and (
                     not os.path.exists(file_path + "_extracted")
                     or len(os.listdir(file_path + "_extracted")) == 0
                 ):