[Patches] [ python-Patches-559250 ] more POSIX signal stuff

[noreply@sourceforge.net]

Patches item #559250, was opened at 2002-05-22 17:59
You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=305470&aid=559250&group_id=5470

Category: Modules
Group: Python 2.3
Status: Open
>Resolution: Accepted
Priority: 5
Submitted By: Michael Hudson (mwh)
>Assigned to: Michael Hudson (mwh)
Summary: more POSIX signal stuff

Initial Comment:
I got annoyed with Python not having sigsuspend et al,
so I wrote this.

It's not finished yet:

it doesn't do autoconf checks, so if you don't have
posix signal handling, it ain't gonna compile)

it doesn't do docs.

it doesn't do full-on sigaction (because that's not
what I was after).

but first I thought I'd see if any of the more unix
savvy Pythoneers (Martin? Barry?) could see any
enormous flaws in the approach...

It does seem to work, so far.

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

>Comment By: Martin v. Löwis (loewis)
Date: 2002-05-27 16:24

Message:
Logged In: YES 
user_id=21627

The patch looks good; please apply it.

It would be good if you copy the doc strings into libsignal.tex.

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

Comment By: Michael Hudson (mwh)
Date: 2002-05-24 16:17

Message:
Logged In: YES 
user_id=6656

And here are the docs.  I've even done a section for whatsnew!

The table looks a bit wierd in the html, but I don't think
it warrants concern.

I think I'm done now.  Assigning to Martin for review, but
anyone with access to unices other than linux/x86 is
encouraged to test!

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

Comment By: Michael Hudson (mwh)
Date: 2002-05-24 16:15

Message:
Logged In: YES 
user_id=6656

I think I'm done now.  The patch I'm attaching

(a) does all the stuff my last patch did.

(b) uses the presence of sigprocmask to assume that all the
other posix-y stuff is there.  I don't know if this is safe,
but I don't know what would be.

(c) adds tests

(d) adds a Misc/NEWS segment.

The patch I'll attach in a minute has the docs.

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

Comment By: Michael Hudson (mwh)
Date: 2002-05-23 15:56

Message:
Logged In: YES 
user_id=6656

Hmm, I don't think I'm going to bother with sigaction.  Most
of the options just don't make sense given when Python
signal handlers get run.

So, docs and autoconf hackery (spit) to go...

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

Comment By: Michael Hudson (mwh)
Date: 2002-05-23 15:49

Message:
Logged In: YES 
user_id=6656

Try this instead.  Definitely a good call on the interface
front.

Still same caveats.

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

Comment By: Michael Hudson (mwh)
Date: 2002-05-23 11:52

Message:
Logged In: YES 
user_id=6656

Hmm, fair point.  I just got carried away :)

Patch later.

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

Comment By: Martin v. Löwis (loewis)
Date: 2002-05-22 19:07

Message:
Logged In: YES 
user_id=21627

I wonder whether we have to expose the API for signal sets.
Couldn't we just use lists/tuples of integers? This ought to
be similar to the processing of FD_SETs in select.

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

You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=305470&aid=559250&group_id=5470