[ python-Bugs-1083177 ] Change in signal function in the signal module

SourceForge.net noreply at sourceforge.net
Fri Jan 7 15:35:58 CET 2005


Bugs item #1083177, was opened at 2004-12-10 15:58
Message generated for change (Comment added) made by akuchling
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=105470&aid=1083177&group_id=5470

>Category: Documentation
Group: Python 2.4
>Status: Closed
>Resolution: Fixed
Priority: 5
Submitted By: Gary H. Loechelt (loechelt)
Assigned to: A.M. Kuchling (akuchling)
Summary: Change in signal function in the signal module

Initial Comment:
The signal function of the signal module handles
arguments differently in Python 2.4 than in Python 2.3.
 I am running on an HP-UX 11 workstation.  If I set a
handler for a signal that cannot be trapped, like KILL
(signal 9), the signal function accepts the argument in
Python 2.3 but ignores the operation.  However, if I do
the same thing in Python 2.4, the signal function
rejects the argument and raises a RuntimeError.

I am not sure if this change in behavior is intentional
or not.  It makes sense in one way to complain about an
invalid argument (as in Python 2.4) rather than just
ignore the operation (as in Python 2.3).  However, I
did not find this change in either the documentation or
the release notes, and it caught me by surprise. 
Granted, the stricter argument checking probably caught
a sloppy line of coding.  Still, some kind of user
warning would have been nice if this was an intentional
change.

I am enclosing a simple Python script which illustrates
the problem.  It finishes normally when using Python
2.3 and raises a RuntimeError when using Python 2.4:

Traceback (most recent call last):
  File "set_signals.py", line 7, in ?
    signal.signal(signal.SIGKILL, dummy)
RuntimeError: (22, 'Invalid argument')

Gary H. Loechelt


----------------------------------------------------------------------

>Comment By: A.M. Kuchling (akuchling)
Date: 2005-01-07 09:35

Message:
Logged In: YES 
user_id=11375

Change now described in "What's New".


----------------------------------------------------------------------

Comment By: Anthony Baxter (anthonybaxter)
Date: 2004-12-13 08:16

Message:
Logged In: YES 
user_id=29957

This is a deliberate change - "errors should not pass
silently". I agree that it should get better visibility.
Perhaps an entry in amk's what's new document in the section
on Porting to 2.4
<http://www.python.org/doc/2.4/whatsnew/node15.html>
would be appropriate. AMK? (I've added a comment to the
document as a reminder)


----------------------------------------------------------------------

Comment By: Raymond Hettinger (rhettinger)
Date: 2004-12-10 22:32

Message:
Logged In: YES 
user_id=80475

On WinME, I appropriately get an AttributeError consistently
for Py2.2, Py2.3, and Py2.4.

Anthony, you've made the most recent updates to the
signalmodule.  What do you think?

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=105470&aid=1083177&group_id=5470


More information about the Python-bugs-list mailing list