Re: glibc conditioning

[Paul Eggert <eggert at twinsun dot com>]

> From: Jinsong Zhao <zhaojs@cadence.com>
> Date: 28 Aug 2001 10:31:54 -0700
> 
> Linux is different,

It's not just Linux x86: Solaris x86 behaves like Linux x86.

> in that it uses 80 bits for FPU while the rest are using 64 bits
> including FreeBSD.

But 80 bits is better than 64 bits for most real applications, as it
gives more accuracy.

> To have the consistency, I add several lines to make sure the FPU
> operates in 64-bit mode.

That may work for your particular example, but it does not guarantee
consistency for all applications.  When the x86 FPU operates in 64-bit
mode, it does not conform to the IEEE floating point standard.  In
some cases, behavior will still disagree with your other platforms.

> Now you can ridicule me that our engineering has no idea about the
> floating-point arithmetic or the x86 architecture.

I'm sure that many of your company's engineering staff are quite
competent.  But you sent messages that suggested limited understanding
of floating-point issues, and your messages contained comments like "a
$3 billion annual business" and "there is no point in arguing with
me".  Those messages made you look both wrong-headed and arrogant.  No
wonder some people replied so strongly.

(Personally, I was more amused than offended by your messages, as they
reminded me _exactly_ of some of the undergraduate students I once taught.
I'll let you guess what grades they got.  :-)