eryksun added the comment:
PyOS_StdioReadline does an unchecked fflush(sys_stdout) and fflush(stderr). It
only 'crashes' (the processes is intentionally killed by the CRT) if there's
buffered data to flush. Since it prints the prompt to stderr, Zachary's example
won't crash if sys.ps1 and sys.ps2 are empty strings. (Does anyone know why the
prompt goes to stderr? readline on Linux uses stdout.)
These calls should be guarded by _PyVerify_fd. On a related vein, in msg237845
I suggested adding a custom closer for file objects that checks _PyVerify_fd,
but the check could also be added directly to close_the_file in fileobject.c.
Here's the stack trace for 2.7 (it's similar for 3.4):
C:\Program Files\Python27>cdb -xi ld python
Microsoft (R) Windows Debugger Version 6.12.0002.633 AMD64
Copyright (c) Microsoft Corporation. All rights reserved.
CommandLine: python
Symbol search path is:
symsrv*symsrv.dll*C:\Symbols*http://msdl.microsoft.com/download/symbols
Executable search path is:
(1078.11ec): Break instruction exception - code 80000003 (first chance)
ntdll!LdrpDoDebuggerBreak+0x30:
00000000`76e8cb70 cc int 3
0:000> g
Python 2.7.9 (default, Dec 10 2014, 12:28:03) [MSC v.1500 64 bit (AMD64)]
on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> import os
>>> os.close(2)
ntdll!ZwTerminateProcess+0xa:
00000000`76e3157a c3 ret
0:000> kc
Call Site
ntdll!ZwTerminateProcess
KERNELBASE!TerminateProcess
MSVCR90!invoke_watson
MSVCR90!invalid_parameter
MSVCR90!write
MSVCR90!flush
MSVCR90!fflush_nolock
MSVCR90!fflush
python27!PyOS_StdioReadline
python27!PyOS_Readline
python27!tok_nextc
python27!tok_get
python27!PyTokenizer_Get
python27!parsetok
python27!PyParser_ParseFileFlagsEx
python27!PyParser_ASTFromFile
python27!PyRun_InteractiveOneFlags
python27!PyRun_InteractiveLoopFlags
python27!PyRun_AnyFileExFlags
python27!Py_Main
python!__tmainCRTStartup
kernel32!BaseThreadInitThunk
ntdll!RtlUserThreadStart
----------
nosy: +eryksun
_______________________________________
Python tracker <rep...@bugs.python.org>
<http://bugs.python.org/issue19050>
_______________________________________
_______________________________________________
Python-bugs-list mailing list
Unsubscribe:
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
