I think that's the right term anyways.

I want to run certain assembly code within a macro. Is there an interface for doing this? In particular, I'd like to use fld/fistp and fild/fstp to convert arguments passed to the macro between integer and float. I think this would compile a lot faster than trying to implement IEEE 754 at macro level. And also I'm lazy and don't want to do the latter if there is no need to.

Please help. I did look in the manual but I can't find anything on this.



p.s.
I know I can put in floats like 1.0 and FASM will put it that way in the file, but that doesn't work when for example they get added together (e.g. a 1.5 added to another 1.5) and I want it to round them and store as integer in the file..

In other words, you need to "execute" FPU instructions at compile time? If not you should, that is a very good idea

I would but I don't know how.

Well that's pretty unlikely to be possible in any assembler, FASM included. Also FASM floating point constants capability is limited.

Borsuc, but the interpreter layer language is powerful enough to implement a processor, and although it doesn't support floating point operations, you could still implement them in the same way FPU emulators did at the time where the FPU unit was not guaranteed to be present.

Quite a few years ago (about the time of fasm 1.3x, I think) I was trying to design some kind of "plugin" functionality that would allow you to write a kind of macros for assembly layer using the x86 assembly as well (since fasm itself cannot be ported to other architecture than x86, that wouldn't be any problem to execute such code). However I gave up because of the security problems that would arise from such feature.

Oh damn, I forgot a very important instruction that would be hard to emulate (if not completely impossible), JMP/Jxx... But well, you could still use if, repeat, while from fasm and put inside of the control structures bodies the non-branching emulated instructions.

Here a very simple instruction to emulate:

(use16 and use32 supported only)

Tomasz Grysztar,

May be some kind of *nix backtick functionality? I.e. replace `expr round(2*$pi*$r)` with output of that command (pi and r replaced with actual values). Separate process, almost safe bet.

Like `del c:\boot.ini`?

I didn't know *nix had c: or boot.ini.

Me neither, where you take that from?

Yep, you misunderstood your source.

I thought it meant delete the file called "boot.ini" in the "c:\" directory.

What did it really mean?

Yes, I meant that, using FASM under Windows, obviously. (The computer won't explode for using a feature borrowed from Unix, I can even use it right now by using bash for Windows)

Not even if you use that feature to delete boot.ini?

You could still, you know, delete the source code if not just boot.ini (which is easily replaced), and all your work goes to trash.