Re: Problems with "disassemble" in CVS

[Andrew Cagney <ac131313 at redhat dot com>]

> The new code for disassemble uses xfer_memory if it decides we are not local.
> The heuristic is a little hokey - for instance, it doesn't know about
> to_shortname == "lin-lwp".  So for lin-lwp, we end up deciding to read from
> the exec.
>
> The problem is, what if the memory being disassembled isn't _in_ the
> executable?  I use disas on arbitrary memory regions, and more importantly,
> I use it on functions in shared libraries.  I don't think we can use
> xfer_memory after the target starts running.

What about at least reversing the logic?  Failure mode is to do the slow 
correct (not fast incorrect) thing.

Andrew