Menu
▾
▴
[sdcc-devel] [ sdcc-Bugs-1505998 ] Cannot operate with floats
|
From: SourceForge.net <no...@so...> - 2006-06-26 07:35:04
|
Bugs item #1505998, was opened at 2006-06-14 12:30 Message generated for change (Comment added) made by tecodev You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=100599&aid=1505998&group_id=599 Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: pic14 target Group: known bugs Status: Open Resolution: None Priority: 5 Submitted By: Nobody/Anonymous (nobody) Assigned to: Raphael Neider (tecodev) Summary: Cannot operate with floats Initial Comment: I can't operate with floats, I have an error in asm file such us: trabajo.asm:874:Error [113] Symbol not previously defined (___fsmul). trabajo.asm:900:Error [113] Symbol not previously defined (___fsmul). trabajo.asm:924:Error [113] Symbol not previously defined (___fsmul). trabajo.asm:950:Error [113] Symbol not previously defined (___fsmul). trabajo.asm:974:Error [113] Symbol not previously defined (___fsmul). trabajo.asm:1000:Error [113] Symbol not previously defined (___fssub). trabajo.asm:1024:Error [113] Symbol not previously defined (___fsadd). trabajo.asm:1048:Error [113] Symbol not previously defined (___fssub). make.exe: *** [trabajo.o] Error 1 It's just multiplicating or dividing, not with complex operations. I'm working with PIC 16f876A, and I can't even use math.h e-mail: tan...@gm... Thanks a lot ---------------------------------------------------------------------- >Comment By: Raphael Neider (tecodev) Date: 2006-06-26 07:35 Message: Logged In: YES user_id=1115835 Oh, I forgot to mention: Currently you need to manually link against libm.lib when using more advanced mathematics (again: highly untested). Just thougth you might want to know... Regards, Raphael ---------------------------------------------------------------------- Comment By: Raphael Neider (tecodev) Date: 2006-06-25 22:28 Message: Logged In: YES user_id=1115835 Floating point support has now been added to PIC14, but beware: Due to register allocation problems and code page selection problems (still to to further investigated) this is to regarded more as a technology demo than production quality code. I just managed to get one FP operation (add, sub od mul) execute correctly, further applications jump into nirvana... Div loops endlessly, mod is untested, but if you want to give it a try, it's part of SDCC r4250. Keeping this report open for now, Raphael BTW: Support for local variables has been added, too! ---------------------------------------------------------------------- Comment By: Nobody/Anonymous (nobody) Date: 2006-06-19 16:51 Message: Logged In: NO Correction about unions... the pic14 port supports, but all variables must be declared static! (Deviation from ANSI?) I moved all _fs things from lib/src, to libsdcc inside pic/ and run make again. Some errors ocurred but mitigated revoking some volatiles and putting "static" before any "union" type var declaration. Also There are duplicated symbols that is being generated by sdcc during compilation that makes gpasm to fail. To solve this I edited manually the .asm amd compiled it (using gpasm). As far as I know, it's working fine (I didn't do regression tests, maybe I write one!) ---------------------------------------------------------------------- Comment By: Raphael Neider (tecodev) Date: 2006-06-18 16:41 Message: Logged In: YES user_id=1115835 The pic14 port is currently missing the required support routines (__fssub, __fsmul, ...) to deal with floats. I guess using floats on a device as small as a pic14 device does not make much sense, but I will add the functions nonetheless. Should be there within the week... I hope... Regards, Raphael ---------------------------------------------------------------------- Comment By: Nobody/Anonymous (nobody) Date: 2006-06-18 02:49 Message: Logged In: NO floats will be supported if pic14 port supports unions. The union float_long is required in some operations and pic14 port doesn't handle unions! This is a big deal, so I believe fixing this issue, pic14 can handle successfully floats, unless we have another issue related to internal representation of floats in the uC structures... ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=100599&aid=1505998&group_id=599 |