Update testing_util.py

The improved evaluation framework for standard_input in CODE_TYPE has addressed the following bugs:

- Wildcard Imports Conflict: The framework now properly handles code containing from [module_name] import * to avoid issues. For example:
import time
a = 1
from math import *
print(a + 1)
  Previously, this would be evaluated as:
import time
def code():
    a = 1
    from math import *  # Error
  leading to errors.

- Name Conflicts with os** Module**: Fixed the runtime error caused by from os import * when it conflicts with the variable name name in the code.

- FastIO Implementations: The framework now supports code with FastIO implementations that utilize the fileno() attribute of IOBase.

- IOStream Operations: Improved handling of code that reads input and writes output using IOStream.

- File-Based Input and Output: The framework can now correctly evaluate code that reads input from input.txt and writes output to output.txt.

Author: bowen92

metrics/testing_util.py

@@ -1,4 +1,4 @@
-# modifed from https://github.com/hendrycks/apps/blob/main/eval/testing_util.py to fix some evaluation bugs
+# modifed from https://github.com/hendrycks/apps/blob/main/eval/testing_util.py to fix some evaluation bugs and add instructions
 
 from .pyext2 import RuntimeModule
 import signal