Ben,
I talked not about offset between codes in different segments. I talked
about possibility to make the code (staying in any single segment) to use
offsets to PC pointer rather than table jumps. Then independently upon
amount of segments I could use the adresses to navigate inside any single
segment.
If CW buit-in assembler could understand pc-relative offset like
lea smth(pc),a0
it would be a cure (while both this instruction and "smth" are in a single
segment).
But CW understands only
lea smth,a0
and generates
lea smth(a5),a0
what gives a wrong result for multisegment application.
--
Konstantin
"Ben Combee" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]
> At 12:39 PM 5/12/2004, you wrote:
> >Everything (test1 and test2) works fine providing the code lays in the
first
> >code segment (code0). It fails for any other segment.
> >Is there any way to use PC relative offset like
> > lea.l table(PC),a0 (CW does not accept this syntax)
> >or any other way to find correct value of table address (for multisegment
> >application)?
>
> No. The key thing about segments is that there is no fixed relative
> position between code in one segment and code in another segment. The
only
> way to get the address of something in another segment is to call a
> function in that segment and have it return the address. Calls go through
> the jump table and are resolved that way, but taking the address ends up
> just returning the address of the jump table entry.
>
> -- Ben Combee, DTS technical lead, PalmSource, Inc.
> Read "Combee on Palm OS" at http://palmos.combee.net/
>
>
>
--
For information on using the Palm Developer Forums, or to unsubscribe, please see
http://www.palmos.com/dev/support/forums/
