On Mon, Oct 08, 2012 at 10:55:20AM -0400, Simo Sorce wrote:
On Fri, 2012-10-05 at 10:48 +0200, Pavel Březina wrote:
> On 10/04/2012 04:22 PM, Simo Sorce wrote:
> > On Thu, 2012-10-04 at 13:03 +0200, Pavel Březina wrote:
> >> From 7c49fe46a3fbce1433ad1680a2e12335542706ed Mon Sep 17 00:00:00 2001
> >> From: =?UTF-8?q?Pavel=20B=C5=99ezina?= <pbrezina(a)redhat.com>
> >> Date: Thu, 4 Oct 2012 11:42:59 +0200
> >> Subject: [PATCH 1/3] do not check for POLLERR, POLLHUP, POLLNVAL when
> >> reading
> >> data
> >> According to the poll manpage, those error are valid only for
> >> output only.
> > Ah btw I wanted to ask what you mean with this comment in the commit
> > message.
> > When 'reading' the flags after the poll returns we are in the
> > situation, but from the way you wrote the comment to me it reads like
> > you mention 'output' as meaning 'write event' (ie POLLOUT) ?
> > Simo.
> Oh, ignore anything I wrote before please. I completely misunderstood
> the documentation.
> I'm sending new patch. It special cases POLLHUP. If it occurs, we will
> read the pipe and then close fd on the client side.
Looks good to me now.
Looks good to me, too.