Date: Fri, 5 May 2017 01:52:04 +1000 (EST) From: Bruce Evans <brde@optusnet.com.au> To: Ed Maste <emaste@freebsd.org> Cc: Ed Schouten <ed@nuxi.nl>, src-committers <src-committers@freebsd.org>, "svn-src-all@freebsd.org" <svn-src-all@freebsd.org>, "svn-src-head@freebsd.org" <svn-src-head@freebsd.org> Subject: Re: svn commit: r303033 - head/share/man/man7 Message-ID: <20170505013607.S2434@besplex.bde.org> In-Reply-To: <CAPyFy2CwDn-XFvgBVdJaaKy%2BLPTth6Z%2BBEtzx369hDnd%2B%2Bz7qQ@mail.gmail.com> References: <201607191746.u6JHk9ov092270@repo.freebsd.org> <CABh_MKnXt7X1L36wB0jrth-7zCvdyOKjPiiXvyYyfP7V9hqQbQ@mail.gmail.com> <CAPyFy2CwDn-XFvgBVdJaaKy%2BLPTth6Z%2BBEtzx369hDnd%2B%2Bz7qQ@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, 4 May 2017, Ed Maste wrote: > On 4 May 2017 at 09:44, Ed Schouten <ed@nuxi.nl> wrote: >> Hi Ed, >> >> 2016-07-19 19:46 GMT+02:00 Ed Maste <emaste@freebsd.org>: >>> +.It arm Ta soft Ta soft, double precision >> ... >>> +.It mips Ta soft Ta identical to double >> >> I was wondering, what's the difference between two ways of phrasing >> it? If long double is double precision, it's identical to double, >> right? Wrong. long double is a different type with a different rank to double. It cannot be identical. "identical" is a fuzzy way of saying that the representation is identical and all operations are the same at the level of the represention. > Yes, good point. I copied the text that was on the wiki page and > missed making it consistent here. Probably the style used by the arm > entry is the one we want. The arm entry is not technically incorrect, but is less useful. It would be satisfied by long doubles with twice the number of bits (half padding) in a different order. These tables probably shouldn't attempt to give any details except the size and the padding. Padding and representation are not mentioned for integers at all, because everying knows that integers are 2's complement with no padding. Long doubles have padding on supported arches iff they are longer than double but shorter than quad, but the tables don't give enough details to determine this, and this is just one of the complications for long doubles. <float.h> also doesn't give enough details. It has to be augmented by _fpmath.h to see the details of the representation. Bruce
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20170505013607.S2434>