You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@httpd.apache.org by rb...@covalent.net on 2001/06/08 19:37:40 UTC
Re: cvs commit: httpd-2.0/server mpm_common.c
> Modified: server mpm_common.c
> Log:
> the input handle to the pod (used by child processes) needs to be non-blocking
Shouldn't the child process be using the output handle to the pod? The
parent uses input to write. I'm pretty sure I fixed this bug last night
by switching when we used pod_in and when we used pod_out.
Ryan
_______________________________________________________________________________
Ryan Bloom rbb@apache.org
406 29th St.
San Francisco, CA 94131
-------------------------------------------------------------------------------
Re: cvs commit: httpd-2.0/server mpm_common.c
Posted by Greg Marr <gr...@alum.wpi.edu>.
At 05:43 PM 06/08/2001, Ryan wrote:
>Input and output have nothing to do with parent and child
>relationship. A
>pipe is a uni-directional communication medium. The parent uses the
>input
>side, because it accepts input. The child uses the output side,
>because
>it produces output.
This confused the heck out of me for a couple minutes until I figured
out that "it" refers to "side [of the pipe]", not "The parent" and
"The child".
"The parent uses the input side, because [the input side] takes
input. The child uses the output side, because [the output side]
produces output."
:)
--
Greg Marr
gregm@alum.wpi.edu
"We thought you were dead."
"I was, but I'm better now." - Sheridan, "The Summoning"
Re: cvs commit: httpd-2.0/server mpm_common.c
Posted by Jeff Trawick <tr...@bellsouth.net>.
<rb...@covalent.net> writes:
> On 8 Jun 2001, Jeff Trawick wrote:
>
> > <rb...@covalent.net> writes:
> >
> > > > Modified: server mpm_common.c
> > > > Log:
> > > > the input handle to the pod (used by child processes) needs to be non-blocking
> > >
> > > Shouldn't the child process be using the output handle to the pod? The
> > > parent uses input to write. I'm pretty sure I fixed this bug last night
> > > by switching when we used pod_in and when we used pod_out.
> >
> > I guess you're talking about a change you didn't check in?
>
> Nope, it was definately checked in.
>
> Revision 1.51 / (download) - annotate - [select for diffs] , Fri Jun 8
> 04:55:44 2001 UTC (16 hours, 41 minutes ago) by rbb
> Branch: MAIN
> Changes since 1.50: +2 -2 lines
> Diff to previous 1.50 (colored) It helps to write to the input side of the
> pipe, and read from the output side.
>
> > Obviously "input" and "output" can't be from the perspective of both
> > parent and child. Feel free to change the perspective accordingly.
>
> Input and output have nothing to do with parent and child relationship. A
> pipe is a uni-directional communication medium. The parent uses the input
> side, because it accepts input. The child uses the output side, because
> it produces output. I can't think of a good reason to use non-blocking
> I/O on the input side. If anything, I think this is a bad idea. If the
> pipe is full, we will just return since we are using non-blocking I/O now.
> If we back-out this change and use blocking I/O, then we will sit until
> there is space in the pipe.
I see what you're saying. (we were blocked in pipe read before I made
the change and processing requests afterwords... no explanation...)
backed out...
--
Jeff Trawick | trawickj@bellsouth.net | PGP public key at web site:
http://www.geocities.com/SiliconValley/Park/9289/
Born in Roswell... married an alien...
Re: cvs commit: httpd-2.0/server mpm_common.c
Posted by rb...@covalent.net.
On 8 Jun 2001, Jeff Trawick wrote:
> <rb...@covalent.net> writes:
>
> > > Modified: server mpm_common.c
> > > Log:
> > > the input handle to the pod (used by child processes) needs to be non-blocking
> >
> > Shouldn't the child process be using the output handle to the pod? The
> > parent uses input to write. I'm pretty sure I fixed this bug last night
> > by switching when we used pod_in and when we used pod_out.
>
> I guess you're talking about a change you didn't check in?
Nope, it was definately checked in.
Revision 1.51 / (download) - annotate - [select for diffs] , Fri Jun 8
04:55:44 2001 UTC (16 hours, 41 minutes ago) by rbb
Branch: MAIN
Changes since 1.50: +2 -2 lines
Diff to previous 1.50 (colored) It helps to write to the input side of the
pipe, and read from the output side.
> Obviously "input" and "output" can't be from the perspective of both
> parent and child. Feel free to change the perspective accordingly.
Input and output have nothing to do with parent and child relationship. A
pipe is a uni-directional communication medium. The parent uses the input
side, because it accepts input. The child uses the output side, because
it produces output. I can't think of a good reason to use non-blocking
I/O on the input side. If anything, I think this is a bad idea. If the
pipe is full, we will just return since we are using non-blocking I/O now.
If we back-out this change and use blocking I/O, then we will sit until
there is space in the pipe.
Ryan
_______________________________________________________________________________
Ryan Bloom rbb@apache.org
406 29th St.
San Francisco, CA 94131
-------------------------------------------------------------------------------
Re: cvs commit: httpd-2.0/server mpm_common.c
Posted by Jeff Trawick <tr...@bellsouth.net>.
<rb...@covalent.net> writes:
> > Modified: server mpm_common.c
> > Log:
> > the input handle to the pod (used by child processes) needs to be non-blocking
>
> Shouldn't the child process be using the output handle to the pod? The
> parent uses input to write. I'm pretty sure I fixed this bug last night
> by switching when we used pod_in and when we used pod_out.
I guess you're talking about a change you didn't check in?
Obviously "input" and "output" can't be from the perspective of both
parent and child. Feel free to change the perspective accordingly.
--
Jeff Trawick | trawickj@bellsouth.net | PGP public key at web site:
http://www.geocities.com/SiliconValley/Park/9289/
Born in Roswell... married an alien...