float.fs missing f- and f+

[Bernard Mentink]

Tried to compile float.fs for the pic24, but fails due to missing f- and f+, are these somewhere else, and should it then be noted in the header as a dependancy?

[Bernard Mentink]

Ahh, I see you have to set FLOATS=1 in p24fj_ga_config.inc (for nano) however I get the following compile error:

Info: Project is using a large data memory model when small data memory model is sufficient.

make[2]: Leaving directory '/Users/bernardmentink/My_Work/Arduino/flashforth/pic24/FF_24-30-33.X'
make[1]: Leaving directory '/Users/bernardmentink/My_Work/Arduino/flashforth/pic24/FF_24-30-33.X'
../src/ff-pic24-30-33.s:7163: undefined reference to `___addsf3'
build/default/production/_ext/1360937237/ff-pic24-30-33.o(.text+0x2318):../src/ff-pic24-30-33.s:7163: undefined reference to `___addsf3'
build/default/production/_ext/1360937237/ff-pic24-30-33.o(.text+0x232e): In function `CFSUB_':
../src/ff-pic24-30-33.s:7163: undefined reference to `___subsf3'
build/default/production/_ext/1360937237/ff-pic24-30-33.o(.text+0x2330):../src/ff-pic24-30-33.s:7163: undefined reference to `___subsf3'
build/default/production/_ext/1360937237/ff-pic24-30-33.o(.text+0x2346): In function `CFMUL_':
../src/ff-pic24-30-33.s:7163: undefined reference to `___mulsf3'
build/default/production/_ext/1360937237/ff-pic24-30-33.o(.text+0x2348):../src/ff-pic24-30-33.s:7163: undefined reference to `___mulsf3'
build/default/production/_ext/1360937237/ff-pic24-30-33.o(.text+0x235e): In function `CFDIV_':
../src/ff-pic24-30-33.s:7163: undefined reference to `___divsf3'
build/default/production/_ext/1360937237/ff-pic24-30-33.o(.text+0x2360):../src/ff-pic24-30-33.s:7163: undefined reference to `___divsf3'

etc etc ..

EDIT: I presume I need the Microchip maths lib, where do I get that?
EDIT2: Microcip webiste says the floating point lib is included free with the compiler. So not sure why this error ..

[Mikael Nordman]

In xc16-ld you have to link with -lm and include the standard libraries by clearing the Exclude standard libraries tickbox.

[Bernard Mentink]

Great, thanks. That worked. Maybe some instructions in the header of float.fs would be useful

[Mikael Nordman]

Here are some examples of use.

1e f. 1.00000  ok
2.3456e10 f. 23456000000.  ok
2.3456e10 fe. 23.4560e9  ok
2.3456e10 fs. 2.34560e10  ok
2.3e-9 4.7e-12 f* fe. 10.8100e-21  ok