4.5. Insight Into BCC Mode

BCC mode could convert most of functions and methods in the scripts to equivalent C functions, those c functions will be compiled to machine instructions directly, then called by obfuscated scripts.

It requires c compiler. In Linux and Darwin, gcc and clang is OK. In Windows, only clang.exe works. It could be configured by one of these ways:

  • If there is any clang.exe, it’s OK if it could be run in other path.
  • Download and install Windows version of LLVM
  • Download https://pyarmor.dashingsoft.com/downloads/tools/clang-9.0.zip, it’s about 26M bytes, there is only one file in it. Unzip it and save clang.exe to $HOME/.pyarmor/. $HOME is home path of current logon user, check the environment variable HOME to get the real path.

4.5.1. Enable BCC mode

After compiler works, using --enable-bcc to enable BCC mode:

$ pyarmor gen --enable-bcc foo.py

All the source in module level is not converted to C function.

4.5.2. Trace bcc log

To check which functions are converted to C function, enable trace mode before obfuscate the script:

$ pyarmor cfg enable_trace=1
$ pyarmor gen --enable-bcc foo.py

Then check the trace log:

$ ls .pyarmor/pyarmor.trace.log
$ grep trace.bcc .pyarmor/pyarmor.trace.log

trace.bcc            foo:5:hello
trace.bcc            foo:9:sum2
trace.bcc            foo:12:main

The first log means foo.py line 5 function hello is protected by bcc. The second log means foo.py line 9 function sum2 is protected by bcc.

4.5.3. Ignore module or function

When BCC scripts reports errors, a quick workaround is to ignore these problem modules or functions. Because BCC mode converts some functions to C code, these functions are not compatible with Python function object. They may not be called by outer Python scripts, and can’t be fixed in Pyarmor side. In this case use configuration option bcc:excludes and bcc:disabled to ignore function or module, and make all the others work.

To ignore one module pkgname.modname by this command:

$ pyarmor cfg -p pkgname.modname bcc:disabled=1

To ignore one function in one module by this command:

$ pyarmor cfg -p pkgname.modname bcc:excludes + "function name"

Use -p to specify module name and option bcc:excludes for function name. No -p, same name function in the other scripts will be ignored too.

Exclude more functions by this way:

$ pyarmor cfg -p foo bcc:excludes + "hello foo2"

Let’s enable trace mode to check these functions are ignored:

$ pyarmor cfg enable_trace 1
$ pyarmor gen --enable-bcc foo.py
$ grep trace.bcc .pyarmor/pyarmor.trace.log

Another example, in the following commands BCC mode ignores joker/card.py, but handle all the other scripts in package joker:

$ pyarmor cfg -p joker.card bcc:disabled=1
$ pyarmor gen --enable-bcc /path/to/pkg/joker

By both of bcc:excludes and bcc:disabled, make all the problem code fallback to default obfuscation mode, and let others could be converted to c function and work fine.

4.5.4. Changed features

Here are some changed features in the BCC mode:

  • Calling raise without argument not in the exception handler will raise different exception.
>>> raise
RuntimeError: No active exception to re-raise

# In BCC mode
>>> raise
UnboundlocalError: local variable referenced before assignment
  • Some exception messages may different from the plain script.
  • Most of function attributes which starts with __ doesn’t exists, or the value is different from the original.

4.5.5. Unsupported features

If a function uses any unsupported features, it could not be converted into C code.

Here list unsupported features for BCC mode:

unsupport_nodes = (
    ast.ExtSlice,

    ast.AsyncFunctionDef, ast.AsyncFor, ast.AsyncWith,
    ast.Await, ast.Yield, ast.YieldFrom, ast.GeneratorExp,

    ast.NamedExpr,

    ast.MatchValue, ast.MatchSingleton, ast.MatchSequence,
    ast.MatchMapping, ast.MatchClass, ast.MatchStar,
    ast.MatchAs, ast.MatchOr
)

And unsupported functions:

  • exec
  • eval
  • super
  • locals
  • sys._getframe
  • sys.exc_info

For example, the following functions are not obfuscated by BCC mode, because they use unsupported features or unsupported functions:

async def nested():
    return 42

def foo1():
    for n range(10):
        yield n

def foo2():
   frame = sys._getframe(2)
   print('parent frame is', frame)