Discussion:
Revision 39348 (Florian)
(too old to reply)
Bart
2018-06-30 16:19:16 UTC
Permalink
Hi Florian,

I do nut really understand why you return a mantissa with the highest
bit stripped off for type Extended.
From what I have gathered about 80-bit extended type the mantissa of
this type _is_ the full 64 bits, see
http://rvelthuis.de/articles/articles-floats.html

Does Delphi strip that bit off as well in TExtended80Rec.Mantissa?

Bart
_______________________________________________
fpc-devel maillist - fpc-***@lists.freepascal.org
http://lists.freepascal.org/cgi-bin/mailman/listinfo/fpc-deve
Florian Klämpfl
2018-06-30 20:54:45 UTC
Permalink
Post by Bart
Hi Florian,
I do nut really understand why you return a mantissa with the highest
bit stripped off for type Extended.
From what I have gathered about 80-bit extended type the mantissa of
this type _is_ the full 64 bits, see
http://rvelthuis.de/articles/articles-floats.html
Does Delphi strip that bit off as well in TExtended80Rec.Mantissa?
Afaik the mantissa function is defined like this for the records, yes. They do not include the hidden bit.
_______________________________________________
fpc-devel maillist - fpc-***@lists.freepascal.org
http://lists.free
Bart
2018-06-30 21:23:32 UTC
Permalink
On Sat, Jun 30, 2018 at 10:54 PM, Florian Klämpfl
Post by Florian Klämpfl
Afaik the mantissa function is defined like this for the records, yes. They
do not include the hidden bit.
Hmm, then Rudy Velthuis is wrong (or Delphi).

I saw that my patch for the floathelpers caused a regression, see
https://bugs.freepascal.org/view.php?id=33932.
I attached a possible patch there, but I did not build a 64-bit
compiler with FPC_SOFT_FPUX80 defined to see it would even compile.

Bart
_______________________________________________
fpc-devel maillist - fpc-***@lists.freepascal.org
http://lists.freepascal.org/cgi

Loading...