Re: Floating point error

• Date: Wed, 13 Mar 1996 01:43:49 PST
• From:
  mark@ukhayq.cv.com (Mark R. Bowyer)
• cc: moo-cows.PARC@xerox.com
• Content-Md5: A5S2XZ6d4Zpy43mRVcne9A==
• Content-Transfer-Encoding: 7bit
• Content-Type: text/plain; charset=us-ascii

> It appears that there's a problem with floating point in 1.8.0p1...
> On a MOO running 1.8.0beta3 and on another running beta2:
> ;1.6
> => 1.6
> On a MOO running 1.8.0p1 and on ANOTHER running p1:
> ;1.6
> => 1.6000000000000001

Your floating point errors, as previously discussed  in MOO-Cows, are far more likely to be due to your maths library and your host 
than to the MOO server itself.  That's why you get different results on 
different machines.  It works perfectly under Solaris 2.3 and 2.5, BTW.

Pavel can't be blamed for the condition of libm on your system.  Not really much 
*you* can do about it either, I'm afraid, unless you can patch/upgrade your 
host.

Maybe we should compile a list over MOO-Cows of which revs of which OSs have 
these FP resolution problems?  Maybe someone would like to write a patch to do 
rounding to 5dp or something in the builtins to try to aleviate this?

Moredhel.

• Prev: Floating point error
• Next: Re: Floating point error
• Index: MOO-cows
• Thread: MOO-cows