EzDevInfo.com

doctest and unittest with pysandbox

I want to test student submissions in a save environment. That's the reason why I use pysandbox. For testing the student submission I want to use doctest and unittest.

Here is the studentSubmission.py

def factorial(n):
import math

if not n >= 0:
   raise ValueError("n must be >= 0")
if math.floor(n) != n:
   raise ValueError("n must be exact integer")
if n+1 == n:  # catch a value like 1e300
   raise OverflowError("n too large")
resulto = 1
factor = 2
while factor <= n:
   resulto *= factor
   factor += 1
return resulto

Here is a regular unittest by a possible tutor: tutor_tester.py

import unittest
from studentSubmission import factorial
class TestSequenceFunctions(unittest.TestCase):

    def test_equal(self):
        res = factorial(4)
        self.assertEqual(res, 24)

    def test_error(self):
        self.assertRaises(ValueError, factorial, -1)
if __name__ == '__main__':
    unittest.main()

And now the script for the sandbox: sandbox_test.py

import os
from sandbox import Sandbox, SandboxConfig

#sandbox config
sandbox = Sandbox(SandboxConfig('traceback','math','stdout','stderr','exit' ))
sandbox.config.allowModule('unittest','student_submission')


fo = open("./tutor_tester.py", "r")
code = fo.read();
fo.close()

sandbox.execute(code)

When I run this on my ubuntu lts 12.04.3 with Python 2.7.3 version and the latest Pysandbox. Running sandbox_test.py I got following error:

Traceback (most recent call last):
  File "sandbox_test.py", line 17, in <module>
    sandbox.execute(code)
  File "/usr/local/lib/python2.7/dist-packages/sandbox/sandbox_class.py", line 97, in execute
    return self.execute_subprocess(self, code, globals, locals)
  File "/usr/local/lib/python2.7/dist-packages/sandbox/subprocess_parent.py", line 185, in execute_subprocess
    raise output_data['error']
ImportError: Import "result" blocked by the sandbox

When I try doctest : doctest_sandbox.py

def testing():
  """testing student
  >>> from studentSubmission import factorial
  >>> [factorial(n) for n in range(6)]
  [1, 1, 2, 6, 24, 120]
  >>> [factorial(long(n)) for n in range(6)]
  [1, 1, 2, 6, 24, 120]
  >>> factorial(30)
  265252859812191058636308480000000L
  >>> factorial(-1)
  Traceback (most recent call last):
  ...
  ValueError: n must be >= 0
  """

if __name__ == "__main__":
  import doctest
  from sandbox import Sandbox, SandboxConfig
  sandbox = Sandbox(SandboxConfig('math','stdout','stderr','exit' ))
  sandbox.config.allowModule('studentSubmission')
  sandbox.execute(doctest.testmod())

Doctest worked quite well, but also in the end the sandbox gave me an error:

python doctest_sandbox.py -v

    Trying:
        from studentSubmission import factorial
    Expecting nothing
    ok
    Trying:
        [factorial(n) for n in range(6)]
    Expecting:
        [1, 1, 2, 6, 24, 120]
    ok
    Trying:
        [factorial(long(n)) for n in range(6)]
    Expecting:
        [1, 1, 2, 6, 24, 120]
    ok
    Trying:
        factorial(30)
    Expecting:
        265252859812191058636308480000000L
    ok
    Trying:
        factorial(-1)
    Expecting:
        Traceback (most recent call last):
        ...
        ValueError: n must be >= 0
    ok
    1 items had no tests:
    __main__
    1 items passed all tests:
       5 tests in __main__.testing
    5 tests in 2 items.
    5 passed and 0 failed.
    Test passed.
    Traceback (most recent call last):
      File "doctest_sandbox.py", line 21, in <module>
        sandbox.execute(doctest.testmod())
      File "/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/sandbox/sandbox_class.py", line 97, in execute
        return self.execute_subprocess(self, code, globals, locals)
      File "/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/sandbox/subprocess_parent.py", line 185, in execute_subprocess
        raise output_data['error']
    TypeError: exec: arg 1 must be a string, file, or code object

Thanks for your help;)

Restricted Function with libsandbox and pysandbox

I am using the sample2.py listed here https://github.com/openjudge/sandbox as a wrapper to call the sandbox libraries (aka. libsandbox && pysandbox).

Following is my C++ Code

#include <stdio.h>
int main(){
   return 0;
}

The result when i run the sample2.py is

result: RF
cpu: 2ms
mem: 288kB

Can anyone tell me what are the changes i have to do to the wrapper to make this work ?

How do I install libsandbox? I have some problems during installation [closed]

I have some problems installing libsandbox and pysandbox. I've tried with binary and source packages but no. It seems to do OK but, when I run:

from sandbox import *

it displays

Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "/usr/local/lib/python2.7/dist-packages/sandbox/__init__.py", line 57, in <module>
from . import _sandbox
ImportError: /usr/local/lib/python2.7/dist-packages/sandbox/_sandbox.so: wrong ELF class: ELFCLASS32

How can I make it work? I'm running on Linux 64-bit.

Implement online Python Shell using Pysandbox

I want to build an online Python shell like this. Currently I am trying to build a module in Python which does the following things

1. Creates a new session.
2. Runs a code passed as string keeping and maintains the environment variables of the current session.

I am trying to achieve this using Pysandbox. Here is my effort till now

from sandbox import Sandbox, SandboxConfig
from optparse import OptionParser
import sys,traceback

class Runner:
    def __init__(self):
        self.options = self.parseOptions()
        self.sandbox = Sandbox(self.createConfig())
        self.localvars = dict()
    def parseOptions(self):
        parser = OptionParser(usage="%prog [options]")
        SandboxConfig.createOptparseOptions(parser, default_timeout=None)
        parser.add_option("--debug",
            help="Debug mode",
            action="store_true", default=False)
        parser.add_option("--verbose", "-v",
            help="Verbose mode",
            action="store_true", default=False)
        parser.add_option("--quiet", "-q",
            help="Quiet mode",
            action="store_true", default=False)
        options, argv = parser.parse_args()
        if argv:
            parser.print_help()
            exit(1)
        if options.quiet:
            options.verbose = False
        return options

    def createConfig(self):
        config = SandboxConfig.fromOptparseOptions(self.options)
        config.enable('traceback')
        config.enable('stdin')
        config.enable('stdout')
        config.enable('stderr')
        config.enable('exit')
        config.enable('site')
        config.enable('encodings')
        config._builtins_whitelist.add('compile')
        config.allowModuleSourceCode('code')
        config.allowModule('sys',
            'api_version', 'version', 'hexversion')
        config.allowSafeModule('sys', 'version_info')
        if self.options.debug:
            config.allowModule('sys', '_getframe')
            config.allowSafeModule('_sandbox', '_test_crash')
            config.allowModuleSourceCode('sandbox')
        if not config.cpython_restricted:
            config.allowPath(__file__)
        return config
    def Run(self,code):
        # log and compile the statement up front
        try:
            #logging.info('Compiling and evaluating:\n%s' % statement)
            compiled = compile(code, '<string>', 'single')
        except:
            traceback.print_exc(file=sys.stdout)
            return
        try:
            self.sandbox.execute(code)
        except:
            traceback.print_exc(file=sys.stdout)

def f():
    f = open('test.py')
    code = ''
    for lines in f:
        code = code+lines
    runner = Runner()
    runner.Run('a = 5')
    runner.Run('b = 5')
    runner.Run('print a+b')
f()

I am encountering 3 major problems.

1. How to nicely display error? For example, running the above code results in following output

   File "execute.py", line 60, in Run self.sandbox.execute(code) File "/home/aaa/aaa/aaa/pysandbox-master/sandbox/sandbox_class.py", line 90, in execute return self.execute_subprocess(self, code, globals, locals) File "/home/aaa/aaa/aaa/pysandbox-master/sandbox/subprocess_parent.py", line 119, in execute_subprocess raise output_data['error'] NameError: name 'a' is not defined

The undesirable thing here is the call traceback of "execute.py". I just want the function to return the following error.

NameError: name 'a' is not defined

1. How do I maintain the environment of the current session? For example, in the above code sequence

   a = 5
   b = 5
   print a+b
   

should result in output 10. Any ideas?

Sandbox has no attribute 'run' and 'probe'

I have installed pysandbox 1.5.1 and libsandbox 0.3.5 After typing these commands(given in libsandbox github), i am getting errors. Can someone please explain the reason?

from sandbox import *

s=Sandbox(['/foo/bar.exe','arg1','arg2'])

s.run()

Traceback (most recent call last):

File "stdin", line 1, in module

AttributeError: Sandbox instance has no attribute 'run'

s.probe()

Traceback (most recent call last):

File "stdin", line 1, in module

AttributeError: Sandbox instance has no attribute 'probe'

Thanks in advance

Restricted Function (RF) with pysandbox

My problem is exactly same as here and here.

I am also using simple2.py for sandboxing the executable produced by this program test1.c:

 #include<stdio.h>
 int main(){
     puts("Hello World");
     return 0;
 }

I am using the following command to build the executable:

gcc -static test1.c

After this if I do:

ldd a.out

I am getting the output:

not a dynamic executable

However, The result when I run the sample2.py still is

result: RF
cpu: 0ms
mem: 952kB

What exactly am I doing wrong?

How can I use pysandbox in a Django request handler?

I want to use pysandbox to allow users to run code on my server. I'm using Django's internal server (manage.py runserver) with this request handler:

def try_sandbox(request):
    from sandbox import Sandbox
    def func(a, b):
        return a + b
    sandbox = Sandbox()
    result = sandbox.call(func, 1, 2)
    return HttpResponse(result)

When accessing the page I get a ValueError:

Request Method: GET
Exception Type: ValueError
Exception Value:   signal only works in main thread
Exception Location: /Library/Python/2.7/site-packages/sandbox/timeout.py in limitedTime, line 45
Python Executable:  /usr/bin/python

Traceback:

Django Version: 1.3.1
Python Version: 2.7.1
File "[...]views.py" in try_sandbox
  77.     result = sandbox.call(func, 1, 2)
File "/Library/Python/2.7/site-packages/sandbox/sandbox_class.py" in call
  44.         return self._call(func, args, kw)
File "/Library/Python/2.7/site-packages/sandbox/sandbox_class.py" in _call
  31.                 return limitedTime(timeout, func, *args, **kw)
File "/Library/Python/2.7/site-packages/sandbox/timeout.py" in limitedTime
  45.             old_alarm = signal(SIGALRM, signalHandler)

Is it possible to use pysandbox in this environment?

(I'm think I'm using pysandbox 1.1 - that's what the version.py in the download says. The download folder says 1.0.1. I'm running Mac OS 10.7.2.)

Python code jailing

I have bunch of python-projects with untrusted WSGI-apps inside them. I need to run them simulatiously and safely. So I need restrictions for directory access, python module usage and limitations for CPU and Memory.

I consider two approaches:

1. Import via imp-module WSGI-object from defined file, and running it with pysandbox. Now I have SandboxError: Read only object when doing:

   self.config  = SandboxConfig('stdout')
   self.sandbox = Sandbox(self.config)
   self.s = imp.get_suffixes()
   wsgi_obj = imp.load_module("run", open(path+"/run.py", "r"), path, self.s[2]).app
   …
   return self.sandbox.call(wsgi_obj, environ, start_response)
   

2. Modify Python interpreter, exclude potentially risky modules, run in parallel processes, communicate via ZMQ/Unix sockets. I even don't know where to start here.

What could you recommend?