Alternative STR$ for small value numbers

[wimoos]

Printing out numbers smaller than 0.01 is done in exponential notation. Which is generally fine.
However, sometimes it preferable to have small numbers printed in natural notation.
The following (relocatable) code does that.
Because a significant decimal is lost in the process, numbers smaller than 1e-8 are rounded down to 0.
Set the USR() vector to the start of the code and have USR() work as an alternative to STR$().

Example:

Code: Select all

PRINT STR$(1E-4)
 1E-04

PRINT USR(1E-4)
 .0001

Regards,

Wim.

Code: Select all

	jsr $cd8d	; check type numeric
	lda $61
	cmp #$7b	; between -0.015625 and 0.015625 ?
	bcs l500	; branch if large enough
	cmp #$66	; between -1e-8 and 1e-8
	bcc l200	; set it to zero
	lda $66		; get sign
	pha		; save sign
	lsr $66		; set sign to positive
	lda #$bc
	ldy #$d9
	jsr $d867     	; add 1.0
	pla		; retrieve sign
	sta $66		; restore it
	jsr $dddf	; perform STR$
l300	iny
	lda $00ff,y	; one position to the left
	sta $00fe,y
	bne l300
	jmp $d46d	; exit
l200	lda #$00
	sta $61
l500	jmp $d468

[CurtisP]

It still blows my mind a little that USR can return a string value.

[wimoos]

There is not much to it. The dollar sign is just part of the token: this doesn't impose anything.

The underlying routine puts the stringdescriptor on the stack of temporary string descriptors, sets the datatype to "string" and then skips datatype checking (which is against "numeric"). The caller (PRINT, LET or any other function that takes a string) just pulls the descriptor off the stack when it needs it.Thats it.

Regards,

Wim.