> >> > Well, the TRY..CATCH in your prototype would catch any error that may be
> >> > thrown in parse_expression, and the `set_language' must take care of
> >> > changing the language, so I would say that is enough...
> >> 
> >> OK - I will send an updated patch that makes things a little cleaner.
> >> I didn't know whether it was OK to default to type long makes much
> >> sense when the probe says the parameter should be using type "mutex_t".
> 
> I remember we had the same problem (i.e., not parsing using language_c)
> on SystemTap SDT probes as well.  Coincidentally or not, you triggered
> the problem twice!

The curse of using a cool language on exotic platforms :-).

> > The problem occurs while trying to parse a probe's arguument,
> 
> "argument"
> > +  if (current_language->la_language != language_c)
> > +    {
> > +      old_chain = make_cleanup_restore_current_language ();
> > +      set_language (language_c);
> > +    }
> 
> Here I think you meant to use "lang" instead of "language_c", right?

Duh...

Thanks for reviewing the entire patch! Attached is what I just pushed.

gdb/ChangeLog:

        * dtrace-probe.c (dtrace_process_dof_probe): Contain any
        exception raised while parsing the probe arguments.
        Force parsing to be done using the C language parser.
        * expression.h (parse_expression_with_language): Declare.
        * parse.c (parse_expression_with_language): New function.

Tested on sparc-solaris and also on x86_64-linux, just for kicks.

-- 
Joel