Horton

Re: Interruptible sleeps (was Re: CommitFest 2009-07: Yay, Kevin! Thanks, reviewers!)

Markus Wanner <markus@bluegap.ch>

View thread
From: Markus Wanner <markus@bluegap.ch>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: Heikki Linnakangas <heikki.linnakangas@enterprisedb.com>, Fujii Masao <masao.fujii@gmail.com>, pgsql-hackers@postgresql.org, Magnus Hagander <magnus@hagander.net>
Date: 2010-09-09T08:17:18Z
Lists: pgsql-hackers 
On 09/08/2010 08:18 PM, Tom Lane wrote:
> Considering that we know that major platforms
> such as FreeBSD have changed their implementations *very* recently,
> it seems foolish to assume that an executable built on a machine with
> corrected pselect could not be run on one with an older implementation.

FWIW testing a recent development (i.e. 9.0-devel) version of FreeBSD 
still failed to properly support pselect().

> Also, we have plenty of experience with substituting poll() for
> select(), so I'm not too worried about copy-and-pasting such code.

It's certainly a simpler change than the self-pipe trick vs. pselect(), yes.

I'm happy to go with the self-pipe trick. A quick micro-benchmark didn't 
even show any significant difference compared to pselect(), so form that 
perspective, it's not a big deal.

And we'd then have a major project using the self-pipe trick. (I would 
still like to know what others exist).

Regards

Markus Wanner