Re: RFA: makes darwin-nat.c threads (and multi-processes) aware

[Tristan Gingold <gingold at adacore dot com>]

On May 18, 2009, at 9:59 PM, Stan Shebs wrote:

> Tristan Gingold wrote:
> > Hi,
> >
> > at AdaCore we have an improved version of darwin-nat.[ch] that is  
> > in use for a
> > few months.  It is mostly rewritten to handle threads and multi- 
> > processes.  It
> > also handles more corner cases in breakpoints handling.
> >
> > The number of regression is about 220 (to be compared to about 600  
> > for the
> > original version).  Good progress!
> Indeed! I've reviewed this (sorry about the delay) and it looks  
> generally good for committing. I just have a couple nits.
>
> +  /* Current message state.
> +     If the kernel has sent a message it expects a reply and the  
> inferior                                 +     can't be killed  
> before.  */
> +  enum darwin_msg_state { DARWIN_RUNNING,
> +                         DARWIN_STOPPED,
> +                         DARWIN_MESSAGE } msg_state;
> I'm not sure if there's official style on this point, but I think  
> it's better to make enums globally visible. It generally seems to be  
> the case that someone somewhere will eventually find multiple uses  
> for the enum values, might as well accommodate them.

Now defined globally.

> -  darwin_ops->to_has_thread_control = tc_schedlock /*| tc_switch  
> */;                                      +  /* darwin_ops- 
> >to_has_thread_control = tc_schedlock | tc_switch */;
> Might as well just whack the whole line. If someone really wants to  
> know what happened, CVS has the record.

Removed (in a later commit - sorry).

I have also adjusted machoread.c because one Mach-O bfd declaration  
has been changed - as obvious.

The port has slightly bit-rotten (the number of regressions has  
increased); I will try to fix that in the
following weeks.

Thanks,
Tristan.