[ping]: [PATCH] Attach discard_pending_stop_replies to observer inferior_exit

[Yao Qi <yao@codesourcery.com>]

On 12/07/2012 10:50 AM, Yao Qi wrote:
> Hi,
> discard_pending_stop_replies is called in three places, remote_detach_1,
> extended_remote_mourn_1 (/w param pid) and remote_close (/w param -1).
> They will finally call observer_notify_inferior_exit in some way,
>
> remote_detach_1 >remote_mourn_1 >generic_mourn_inferior >exit_inferior >observer_notify_inferior_exit
>          extended_remote_mourn_1 >generic_mourn_inferior >exit_inferior >observer_notify_inferior_exit
> remote_close >discard_all_inferiors >exit_inferior_silent >observer_notify_inferior_exit
>
> so this inspires me that we can attach discard_pending_stop_replies
> to inferior_exit observer, and don't have to worry about the case that
> PID is -1.  This is what this patch does.  Regression tested on
> x86_64-linux with native and gdbserver.  Is it OK?
>
> gdb:
>
> 2012-12-07  Yao Qi  <yao@codesourcery.com>
>
> 	* remote.c (discard_pending_stop_replies): Remove declaration.
> 	(remote_close): Remove call discard_pending_stop_replies.
> 	(remote_detach_1, extended_remote_mourn_1): Likewise.
> 	(discard_pending_stop_replies): Change parameter from PID to
> 	INF.
> 	(_initialize_remote): Attach discard_pending_stop_replies to
> 	inferior_exit observer.

Ping.  http://sourceware.org/ml/gdb-patches/2012-12/msg00126.html

This patch is kind of a cleanup/preparatory patch for 'async remote 
notification' patch series.  This patch uses 'inferior_exit' observer to 
discard pending stop replies, which is better than the current approach 
(call discard_pending_stop_replies in different places).

-- 
Yao (齐尧)