Re: bin/47524: make(1) loops in Job_CatchOutput()

[David Laight <david%l8s.co.uk@localhost>]

The following reply was made to PR bin/47524; it has been noted by GNATS.

From: David Laight <david%l8s.co.uk@localhost>
To: gnats-bugs%NetBSD.org@localhost
Cc: 
Subject: Re: bin/47524: make(1) loops in Job_CatchOutput()
Date: Tue, 5 Feb 2013 20:55:41 +0000

 On Tue, Feb 05, 2013 at 08:00:13AM +0000, Andreas Gustafsson wrote:
 > The following reply was made to PR bin/47524; it has been noted by GNATS.
 > 
 >  David Laight wrote:
 >  > I moved the Job_CatchChildren() from being called unconditionally
 >  > everytime around the loop, to be (usually) only called when a
 >  > token was read from the pipe.
 >  > 
 >  > But I didn't want to break the code if it was actually relying on the
 >  > poll timeout (5 seconds) to collect children.
 >  
 >  But in the timeout case, poll() will return 0, not < 0, therefore
 >  Job_CatchChildren() will in fact *not* be called when there is a
 >  timeout.  It will, however, be called if the poll() system call
 >  fails.  Should the "nready < 0" have been "nready == 0"?
 
 Quite possibly, I think I was just reducing the number of times
 Job_CatchChildren() would be called unnecessarily.
 
 The only time this all goes awry is when the number of jobs gets near
 to the number of bytes that can be written into a pipe.
 The 'job' pipe will just lose credit, you'd have to be very unluky to
 lose anywhere else.
 
        David
 
 -- 
 David Laight: david%l8s.co.uk@localhost