RFC: don't set the pspace on ordinary breakpoints
Pedro Alves
pedro@codesourcery.com
Wed Nov 9 18:30:00 GMT 2011
On Tuesday 08 November 2011 20:23:13, Tom Tromey wrote:
> >>>>> "Tom" == Tom Tromey <tromey@redhat.com> writes:
>
> Tom> I think I will change linespec to ignore program spaces in this state.
>
> The appended has two parts: the linespec.c change, relative to the big
> patch I posted, and the breakpoint.c change I think is needed too. Only
> the latter makes sense in the context of this thread.
>
> I *think* the should_be_inserted change is all that was really needed,
> after re-reading all the messages in this thread. Let me know what you
> think.
I think so, but it's hard to tell, given that the dependency
on the linespec.c changes.
E.g.,
> @@ -1327,6 +1329,12 @@ decode_indirect (struct linespec_state *self, char **argptr)
> CORE_ADDR pc;
> char *initial = *argptr;
>
> + if (current_program_space->executing_startup)
> + /* The error message doesn't really matter, because this case
> + should only hit during breakpoint reset. */
> + throw_error (NOT_FOUND_ERROR, _("cannot evaluate expressions while "
> + "program space is in startup"));
Why is is okay to look at current_program_space here, if you're
iterating over pspaces elsewhere?
Any chance we can have a standalone patch for just the
startup-disabled changes? We'd need something like my previous
suggestion in bkpt_re_set (even if we'd remain buggy WRT
multi-process).
--
Pedro Alves
More information about the Gdb-patches
mailing list