Re: bug with for and time

[Peter Stephenson <pws@csr.com>]

Dan Nelson wrote:
> I did some mailinglist searches, and back in 1999, workers-2526 says
> that getrusage was removed because it wasn't POSIX and the autoconf
> test was too complex.

There's another issue... we used to get the usage for a child by calling
wait3(), which can set a struct rusage.  (Modern systems have wait4,
which allows you to wait for a particular PID, but it looks like we
don't actually use that anyway.)  However, this isn't part of the latest
standards.  So the only standardised way of doing it is to call
getrusage and subtract all appropriate elements.  With some (maximum
value), you can't actually tell for sure whether it's applicable to the
last child or not.

Maybe we should prefer wait3(), which is widely available, and do the
best we can elsewhere, but it's not ideal.  Still, both wait3() and the
additional rusage elements came from BSD; maybe I can rely on either
both or neither being available and skip the case where we can't tie
usage to individual processes.

> One AC_CHECK_MEMBERS call with 14 arguments is all that's needed, plus the
> appropriate #if defined checks around each resource in the TIMEFMT
> expansion code.

This bit I can easily get working.

-- 
Peter Stephenson <pws@csr.com>                  Software Engineer
CSR Ltd., Science Park, Milton Road,
Cambridge, CB4 0WH, UK                          Tel: +44 (0)1223 692070


**********************************************************************
This email and any files transmitted with it are confidential and
intended solely for the use of the individual or entity to whom they
are addressed. If you have received this email in error please notify
the system manager.

This footnote also confirms that this email message has been swept by
MIMEsweeper for the presence of computer viruses.

www.mimesweeper.com
**********************************************************************