Re: change to remote protocol qOffsets response handling

[Andrew Cagney <ac131313 at cygnus dot com>]

"J.T. Conklin" wrote:
> 
> >>>>> "Michael" == Michael Snyder <msnyder@cygnus.com> writes:
> Michael> Agreed, but is it polite to just change it out from under them,
> Michael> without notice, consent or discussion?
> 
> Probably not.  But then again, I don't belief that submitting a patch
> precludes further discussion.  While I try to open discussion for big
> or contraversial changes on the gdb list; but little changes I simply
> send to gdb-patches.  If it turns out that further discussion is
> needed, the thread can migrate to the gdb list as well.
> 
> As for the change itself, who do you mean by 'them'.  Netware tool-
> chain users, NetWare, Inc., or someone else?  The gdb agent itself is
> 'ours', so it's not clear that anyone else needs to approve (unlike,
> for example, changing the protocol described in gdb/vx-share/*).


The code could:

	o	add (yet another) variable to remote.c that can be
		used to select the behavour.

	o	add logic to detect the need for this compat mode

BTW, who else uses this packet?  Cisco came up with their own response. 
It is now fairly clear why.

Someone really twisted could come up with a universal packet that
included: Status, offsets and an arbitrary variable list .... :-)

I also need to, at least, update the documentation to reflect this
legacy (sigh).

Andrew