Re: [PATCH 2/4] gdb: make remote target clear its async event handler in wait

[Pedro Alves <pedro@palves.net>]

On 30/11/20 16:52, Simon Marchi via Gdb-patches wrote:

> However, the code in remote_async_inferior_event_handler checks for for

s/for for/for/

> pending events in addition to whether the stop reply queue is empty, so
> I've made remote_target::wait check for that as well.  I'm not
> completely sure this is ok, since I don't understand very well yet how
> the pending events mechanism works.  But I figured it was safer to do
> this in order to avoid missing events, worst case it just leads to
> unnecessary calls to remote_target::wait.

It won't hurt, but I think it shouldn't be necessary -- if putpkt_binary or
getpkt_or_notif_sane_1 see a notification (search for '%'), we end up in
handle_notification, which marks state->get_pending_events_token, which
then leads us to:

 -> remote_async_get_pending_events_handler
  -> remote_notif_get_pending_events
   -> remote_notif_stop_ack
    -> remote_target::push_stop_reply
     -> mark_async_event_handler (rs->remote_async_inferior_event_token);

which then ends up in target_target::wait.

I don't recall exactly why I had to put the check for the pending
event in remote_async_inferior_event_handler, though.  Maybe it was
necessary because events auto-cleared back then.

Anyhow, this LGTM.  Patches #3 and #4 LGTM too.

Thanks,
Pedro Alves