Re: [PATCH] sim: start a unified sim_do_command

[Mike Frysinger <vapier at gentoo dot org>]

On Wednesday, June 22, 2011 16:16:00 Mike Frysinger wrote:
> On Wednesday, June 22, 2011 16:01:10 Tom Tromey wrote:
> > >>>>> "Mike" == Mike Frysinger <vapier@gentoo.org> writes:
> > Mike> Since sim_do_command for many people simply calls sim_args_command,
> > start Mike> a unified version of it.  For people who handle their own
> > options, they Mike> could switch to this by using sim_add_option_table
> > instead.
> > 
> > I don't actually know anything about this code, but the patch seems to
> > remove implementations of sim_do_command that do more than dispatch to
> > sim_args_command.
> 
> some arches didnt call sim_args_command(), now they do.  bug fix for them.
> 
> some arches would display an error message when users passed an old command
> option and inform them of the new one.  i think people have had long enough
> for that migration, and they can pretty easily figure things out with the
> normal help output.
> 
> m68hc11 provided a dedicated "info" to call an arch-specific
> sim_get_info(), but the common sim_info() entry point also calls that. 
> m68hc11 would also rewrite its internal device tree when people changed
> files, but this behavior deviates from the standard sim behavior, so i
> dont think it should be there at all.  the normal sim workflow
> (target;file;load) should continue to work.
> 
> m32r provided a dedicated hook for reading two pseudo registers, but that
> should be done via the normal gdb interface rather than hooking the command
> line like this.
> 
> i think that sums up all deviations.

if the m68hc11/m32r peeps really want to keep this behavior, i can rework the 
Makefiles so that arches have to opt into the command.  the idea though was to 
make this a standard func so that new ports would get it automatically too.
-mike

Attachment: signature.asc
Description: This is a digitally signed message part.