This is the mail archive of the
gdb@sources.redhat.com
mailing list for the GDB project.
Re: remote ack to [CcSs]
- To: Andrew Cagney <ac131313 at cygnus dot com>
- Subject: Re: remote ack to [CcSs]
- From: jtc at redback dot com (J.T. Conklin)
- Date: 10 May 2001 16:37:09 -0700
- Cc: GDB Discussion <gdb at sources dot redhat dot com>
- References: <3AF9C42D.6090803@cygnus.com>
- Reply-To: jtc at redback dot com
>>>>> "Andrew" == Andrew Cagney <ac131313@cygnus.com> writes:
Andrew> From remote.c:
Andrew>
Andrew> /* The s/S/c/C packets do not return status. So if the target does
Andrew> not support the S or C packets, the debug agent returns an empty
Andrew> string which is detected in remote_wait(). This protocol defect
Andrew> is fixed in the e/E packets. */
Andrew>
Andrew> Thinking about it, I don't think there is anything actually wrong with
Andrew> this behavour and I'm wondering if the [Ee] packet should have adopted
Andrew> the same behavour as the [CcSs].
Well I'm still convinced that the fallback from S/C to s/c is a botch.
It requires global variables "last_sent_signal" and "last_sent_step"
to communicate state between remote_resume() and remote_wait() so that
remote_wait() can pick up the pieces when remote_resume() didn't do
it's job. Doing the same for the step-over-range command would require
even more global state.
--jtc
--
J.T. Conklin
RedBack Networks