[fpc-devel] 3.0.4 and Windows 64

[Sven Barth]

Am 03.10.2018 um 14:23 schrieb Ozz Nixon:
> Thank you for explaining. Since I am needing compatibility at the disk 
> level (binary form)... does anyone know of a unit or some routines to 
> read/write: Comp, Double, Extended byte for byte the same? (even if it 
> translates to something different at run time or in the dataset). If 
> it was for the fact I maintain Halcyon (xBase) code I would slap this 
> all in a DB server. Which also means my port of Accuracer (aidaim.com 
> <http://aidaim.com>) is also going to have similar problems. :-(

You can use FPC's softfpu unit though it is not compiled for all targets 
currently. But you can copy it from $fpcdir/rtl/inc/softfpu.pp and 
compiler yourself. Then you can use the type tfloat80 which is 
essentially Extended on which you can also use the operations defined in 
the softfpu unit (for convenience you could add a unit with operator 
overloads).

Please note that Double is always the same on the main platforms and 
thus doesn't need special handling.

Regards,
Sven
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freepascal.org/pipermail/fpc-devel/attachments/20181007/b13f435b/attachment.html>