Re: [Amforth-devel] >FLOAT

[pito <pi...@vo...>]

Hi, there is a bug in the flib at line 116 of Leon's listing:

: dreversedigits2 ( dinitial n-digits -- dfinal )
dup 0= if
s>f          <<<<<< here - used prior its actual definition
....
Frankly, I overlooked that aprox. 20times today!

The word dreversedigits2 is not used however! But it fires an error
(overlooked in the screen listing). I've found that by a chance as I
saw something on the stack after the upload. I've checked now it
crashes the dictionary - as I wasn't able to run >float (a crash to
reflash). I've removed the dreversedigits2 and now it works at least
for few floats.
> bl parse 1.234e-20 >float . fs.
-1 1.2339996E-20  ok
> bl parse 9.9999e-19 >float . fs.
-1 9.999901E-19  ok
> bl parse 9.9999e32 >float . fs.
-1 9.999901E32  ok
> bl parse -9.9999e-32 >float . fs.
-1 -8.0000973E-32  ok
> bl parse -9.999e-32 >float . fs.
-1 -8.0009995E-32  ok
> bl parse 9.999e-32 >float . fs.
-1 9.9989977E-32  ok
> bl parse +9.999e-32 >float . fs.
0 amforth 4.0 ATmega1284P

PS; IS THERE AN OPTION TO CONFIGURE ERROR HANDLING SUCH IT STOPS
UPLOAD WHEN AN ERROR THROWN ????
Pito.