Re: [RFC] Reducing the use of current_language - some patches

["Ulrich Weigand" <uweigand at de dot ibm dot com>]

Daniel Jacobowitz wrote:
> On Sun, Nov 04, 2007 at 07:50:14PM -0800, Joel Brobecker wrote:
> > On the other hand, I'm not sure we're ready to have the parse context
> > contain the gdbarch.  For instance, the breakpoint structure doesn't
> > save it, so we don't have access to it when we want to re_set the
> > breakpoint.
> 
> I suspect we'd want to re-set with the current gdbarch anyway.  The
> old one may no longer apply...

On the other hand, once we support multi-architecture debugging in
a single session, we will certainly have to distinguish between
breakpoints set in code with different architectures.  That's one
of the hacks in the Cell combined debugger today ...

I'm not completely sure on what the correct solution is, maybe it
can be as simple as checking the gdbarch associated with the
objfile where the breakpoint is set (assuming we will get per-
objfile gdbarchs at some point).

In any case, as long as the changes Joel implements use something
like a "parse context" structure to pass around, it should be
simple to extend that with a gdbarch if and when we want to solve
the multi-arch problem.

For now, it's probably indeed best to leave it out.

Bye,
Ulrich

-- 
  Dr. Ulrich Weigand
  GNU Toolchain for Linux on System z and Cell BE
  Ulrich.Weigand@de.ibm.com