Overflow on posix file date

Classification

Affected Components : builtin, os, sys
Operating System : Linux
Python Versions : 2.6.x, 2.7.x, 3.1.x, 3.2.x
Reproducible : Yes

Source code

import os, sys

testfile = 'TESTFILE'

# Showing stat information of file
stinfo = os.stat('TESTFILE')
print(stinfo)

# Using os.stat to recieve atime and mtime of file
print("access time of TESTFILE: %s" %stinfo.st_atime)
print("modified time of TESTFILE: %s" %stinfo.st_mtime)

# INT is a 32-bit data-type with a ranging from -2^31 (-2,147,483,648) to 2^31-1 (2,147,483,647).
# Clean execution
#os.utime("TESTFILE",(-2147483648, 2147483647))
# Overflow
os.utime("TESTFILE",(-2147483648, 2147483648))

# Showing stat information of file
stinfo = os.stat('TESTFILE')
print(stinfo)

# Using os.stat to recieve atime and mtime of file
print("access time of TESTFILE: %s" %stinfo.st_atime)
print("modified time of TESTFILE: %s" %stinfo.st_mtime)

os.utime("TESTFILE",(1330712280, 1330712292))

# Showing stat information of file
stinfo = os.stat('TESTFILE')
print(stinfo)

# Using os.stat to recieve atime and mtime of file
print("access time of TESTFILE: %s" %stinfo.st_atime)
print("modified time of TESTFILE: %s" %stinfo.st_mtime)

sys.exit()

Steps to Produce/Reproduce

To reproduce the problem create a file named 'TESTFILE' and copy the source code in a file, then execute the script using the following command syntax:

$ python -OOBRtt test.py

Alternatively you can open python in interactive mode:

$ python -OOBRtt <press enter>

Then copy the lines of code into the interpreter.

Description

If system or user set numerical date of a file bigger than a 32bit int, python should intercept the operation and prevent a crash.

At the moment if python reads stats of a file having a date that if transformed in numerical value is bigger then maximum 32bit maximum value, it raises a numeric overflow and the the interpreter crashes.

If the script is executed in Python 2.6 we have:

posix.stat_result(st_mode=33188, st_ino=293602594L, st_dev=2065L, 
st_nlink=1, st_uid=1000, st_gid=1000, st_size=0L, st_atime=1330712280, 
st_mtime=1330712292, st_ctime=1401655733)
access time of TESTFILE: 1330712280.0
modified time of TESTFILE: 1330712292.0
Traceback (most recent call last):
  File "test.py", line 17, in <module>
    os.utime("TESTFILE",(-2147483648, 2147483648))
OverflowError: long int too large to convert to int

And if the script is executed in Python 2.7 or in Python 3.1 we have:

posix.stat_result(st_mode=33188, st_ino=293602594, st_dev=2065, 
st_nlink=1, st_uid=1000, st_gid=1000, st_size=0, st_atime=1330712280, 
st_mtime=1330712292, st_ctime=1401655733)
access time of TESTFILE: 1330712280.0
modified time of TESTFILE: 1330712292.0
Traceback (most recent call last):
  File "test.py", line 17, in <module>
    os.utime("TESTFILE",(-2147483648, 2147483648))
OverflowError: Python int too large to convert to C long

Integers in a 32bit system ranges from -2^31 (-2,147,483,648) to 2^31-1 (2,147,483,647) and python does not check for this condition.

A call setting or reading time of a file between numerical boundaries does not return error

# Clean execution
os.utime("TESTFILE",(-2147483648, 2147483647))

But a call that sets time of a file to a value bigger than the maximum is allowed:

# Overflow
os.utime("TESTFILE",(-2147483648, 2147483648))

If the file is read this generates a numeric overflow inside the interpreter.

Workaround

We are not aware on any easy solution other than implementing strict numeric validation on each operation reading data from files or data from the system as is not implicitly safe.