Date: Sat, 25 Jul 1998 16:08:25 +1000 (EST) From: Peter Hawkins <thepish@FreeBSD.ORG> To: Steve Kargl <sgk@troutmask.apl.washington.edu> Cc: phk@FreeBSD.ORG, cvs-committers@FreeBSD.ORG Subject: Re: cvs commit: src/usr.bin/time time.1 time.c Message-ID: <Pine.BSF.3.96.980725160241.2327G-100000@dana.clari.net.au> In-Reply-To: <199807241435.HAA07570@troutmask.apl.washington.edu>
next in thread | previous in thread | raw e-mail | index | archive | help
>> There is no situation where this patch is required and it's quite a departure >> from the norms elsewhere. The consensus on the list seems to have been that >> this shouldn't be committed. >Clearly, you don't understand the PR. Think about the following > >kargl[204] time -l ls > /dev/null >kargl[206] time -l sh -c "ls > /dev/null" That's not the solution I suggested anyway :) Besides, I've lectured in benchmarking and know that you have to take many readings and take averages before you have any serious measure. The addition of "sh" is NOTHING compared to other random activities (whether the files you're reading are cached, whether the program itself is cached, whether a daemon starts during your test, whether you swap etc...) Even so, if you plan to be that accurate, you can in theory work out what sh itself takes and subtract, but it's going to be a case of X - 0.0 anyway. Peter Hilink Internet Peter Hawkins 381 Swan St Richmond, Vic, Australia Ph: +61-3-9421 2006 Fax: +61-3-9421 2007 http://www.hilink.com.au Peter@hilink.com.au FreeBSD Project: thepish@FreeBSD.org
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.3.96.980725160241.2327G-100000>