This is the mail archive of the
gdb-prs@sourceware.org
mailing list for the GDB project.
[Bug mi/20045] With mi-async on, -exec-continue outputs two (gdb) prompts
- From: "simon.marchi at ericsson dot com" <sourceware-bugzilla at sourceware dot org>
- To: gdb-prs at sourceware dot org
- Date: Thu, 05 May 2016 15:51:54 +0000
- Subject: [Bug mi/20045] With mi-async on, -exec-continue outputs two (gdb) prompts
- Auto-submitted: auto-generated
- References: <bug-20045-4717 at http dot sourceware dot org/bugzilla/>
https://sourceware.org/bugzilla/show_bug.cgi?id=20045
--- Comment #8 from Simon Marchi <simon.marchi at ericsson dot com> ---
(In reply to Pedro Alves from comment #5)
> > Anyway, it seems to me like this condition could only be
> >
> > if (sync_execution)
> >
> > Basically, if we are emulating a synchronous MI command, output the bogus
> > (gdb). When using mi-async off, sync_execution will be false, so we will
> > never output it. The result is this for mi-async off:
>
> I think you meant "When using mi-async on, sync_execution will be false".
Oops, yes.
> But when "maint set target-async off" (i.e., on targets that don't do
> async), sync_execution will be off, even with sync commands.
I didn't think about that. Does that mean that having mi-async on is
impossible with sync targets? That would make sense, since when the target is
busy waiting on the target, it can't process MI commands.
> This patch from the console series replaces that with a single enum that
> works the same for async and sync targets, so probably fixes this:
>
> https://sourceware.org/ml/gdb-patches/2016-02/msg00080.html
>
> (haven't tried yet)
I tested with your console series (both palves/console-submit and
palves/console-v3), and it indeed works fine on that. You are always two steps
ahead of the rest of the world!
> Though I think an easier and safe-for-7.11.1-even fix would be to call
> target_can_async instead of target_is_async here.
Is it a good idea to change such a behavior on a .1 release though? If a
front-end has to do something special to cope with the additional prompt on
7.11 (which I doubt, since we didn't have any bug reports from front-end
developers, but still), it would make things even more complicated for them if
7.11.1 had a different behavior than 7.11. Given that it's a mostly harmless
bug, I suggest we leave all 7.11 gdbs behaving the same, and fix it for 7.12.
WDYT?
--
You are receiving this mail because:
You are on the CC list for the bug.