pagg
[Top] [All Lists]

Re: Feedback requested on 'task notifier' patch

To: Erik Jacobson <erikj@xxxxxxx>
Subject: Re: Feedback requested on 'task notifier' patch
From: kingsley@xxxxxxxxxx
Date: Fri, 28 Oct 2005 16:11:59 +1000
Cc: pagg@xxxxxxxxxxx
In-reply-to: <20051027140125.GA337@sgi.com>
References: <20051027140125.GA337@sgi.com>
Sender: pagg-bounce@xxxxxxxxxxx
User-agent: Mutt/1.4.2.1i
On Thu, Oct 27, 2005 at 09:01:25AM -0500, Erik Jacobson wrote:
> On October 12th, Matt Helsley posted a patch to lse-tech as part of the
> pnotify/task notifier discussion.
> 
> I'm finally looking at it this morning but I'd like to know if anybody
> else has comments.  An archived copy of the posting is here:
> 
> http://marc.theaimsgroup.com/?l=lse-tech&m=112915953709702&w=2
> 
> Comments on the patch itself should be posted to the lse-tech.  However,
> I'm interested in what this patch might be lacking for current PAGG/pnotify
> users.
> 
> I hope to have some comments of my own today.
> 
> Thank you.
> 
> Erik

Hi Erik,

Apologies for coming a little late into the discussions.  I've had
recent deadlines that has kept me from investigating into how the
connector based notifiers work and from tracking pnotify and its
development.

For my use, if I understand the code corectly, I think the main
problem with this patch is the issue with not blocking during the
event notification for watching all tasks.  Most, if not all, of my
callbacks in pnotify can block.

The only alternative I can see is for me to watch specific tasks since
that is allowed to block.  However, this will mean missing newly
forked processes, making everything pointless.

One other missing feature is that we can no longer veto forks.  It
would be nice to have this but it is not a strong requirement for me.

Aside from that I can't see any other issues from a reading of the
code.  I would have to try to use it to get a better feel for what it
does.

-- 
                Kingsley

<Prev in Thread] Current Thread [Next in Thread>