Arghh. INA woes.

rokicki

I spent about an hour yesterday trackdown a newbie mistake: using

andn ina,pinmask

and couldn't figure out why it wasn't working.

Of course I wanted

andn pinmask,ina

Ina used in the destination field of an instruction does not
actually refer to ina, but rather a "shadow" register "under" ina.

Can we change the assembler so if you use ina in a destination
position, you get an error? And to access the shadow register,
you use a different name, like inashadow or something?

I hate it when we lose orthogonality like this.

Anyway, just my two cents.

Paul Baker

The compiler will always support using the shadow registers because Chip uses them when space is tight for loop variables (djnz addresses the loop variable in it's destination field). I will communicate your request, but I wouldn't expect there to be a name change since it would require rewritting already released objects. Just remember any Special Purpose Register marked as read-only is just that, and any attempts to write to it will access the shadow RAM.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Paul Baker
Propeller Applications Engineer

Parallax, Inc.

Cliff L. Biffle

Tom,

It's not ideally suited for your purposes, but my propasm assembler prevents this. (It bit me too.)