compil error VASM

[LeCaravage]

Hi,

May be someone could help.
I've got a strange error when trying to assemble using Vasm :

error 31 in line 41 of "part6.s": expression must be constant
>bufpokt: blk.w HF,0 ; buffer scrolltext


HF is defined below the "blk.w". H=12.
Before the bufpokt label I have an "even" directive.
After this block I have also valid instructions.


I tried to use other constants and same error.
Very strange. Really no idea what's going on.

Edit :

If I replace blk.w HF,0 by dc.b HF it works.

If I set HF=512 and let the dc.b, then no errors also. 512 is not a byte it should display an error or at least a warning. May be vasm set the complementary to 255 ?

[roondar]

As far as I understand it, this is because VASM doesn't know symbols unless they're defined prior to the actual use of them. There are two solutions for this that I've used (there may be more, there may be better ones):

1) define HF before the line that uses it.

HF	EQU	<something>

... etc ...

bufpokt: blk.w HF,0 ; buffer scrolltext

2) define HF as an external symbol in an include file and include that in the file that uses HF

Include file:

	XREF	HF

Source file:
	include file_that_has_hf_xref.i

... etc ...

bufpokt: blk.w HF,0 ; buffer scrolltext

[jotd]

I think blk.w / ds.w needs constant values. XREF isn't going to cut it. You need something like HF = 300 (equate/define)

[roondar]

Originally Posted by jotd

I think blk.w / ds.w needs constant values. XREF isn't going to cut it. You need something like HF = 300 (equate/define)

I was indeed assuming HF was defined like as a constant, yes.

[LeCaravage]

Roondar, you're the man !

If moved the Hf = 12 within the include file (the definition was at first in top of main.s) and it fixed the pb. I guess the Xref does also the trick.
AFAIK Seka doesn't need this.

Thanks pal,

[LeCaravage]

@jotd and Roondar : I'm dumb in english, I used below instead of above in first post ^^

[phx]

Originally Posted by LeCaravage

error 31 in line 41 of "part6.s": expression must be constant
>bufpokt: blk.w HF,0 ; buffer scrolltext


HF is defined below the "blk.w". H=12.

When you only defined H, then HF is undefined and the error is expected. The BLK directive allocates memory, so the size-expression must be constant as jotd correctly pointed out.

Otherwise there should be no problem, even when defining HF below the BLK-directive:

frank@nerthus cat blkexpr.s 
        blk.w   HF,0
HF=12
frank@nerthus vasmm68k_mot blkexpr.s 
vasm 1.8i (c) in 2002-2020 Volker Barthelmann
vasm M68k/CPU32/ColdFire cpu backend 2.3m (c) 2002-2020 Frank Wille
vasm motorola syntax module 3.14a (c) 2002-2020 Frank Wille
vasm test output module 1.0 (c) 2002 Volker Barthelmann

CODE(acrx1):              24 bytes

If I replace blk.w HF,0 by dc.b HF it works.

Then HF does no longer define the size of a memory block, but just a value in memory. That's an important difference. So the assembler generates an 8-bit relocation entry, which must be resolved by the linker.

If I set HF=512 and let the dc.b, then no errors also. 512 is not a byte it should display an error or at least a warning.

It does.

frank@nerthus cat byterange.s 
        dc.b    HF
HF=512
frank@nerthus vasmm68k_mot -quiet byterange.s 

error 2040 in line 1 of "byterange.s": data out of range
>       dc.b    HF

Maybe another typo?

[roondar]

I don't know what I was thinking there... You're of course fully correct.
Next time I'll be sure to check first and then write

Anyway, I do believe I know why I thought this: I seem to recall that you do need to define macro's before use and I guess that made me make a mistake. Well that and trying to quickly answer a post while I really ought to be doing other things

[LeCaravage]

And to summarize, I entered this thread convinced of vasm's fault and I go out I was totally wrong ^^