[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: QMP and the 'id' parameter
From: |
Markus Armbruster |
Subject: |
Re: QMP and the 'id' parameter |
Date: |
Tue, 10 Nov 2020 11:27:26 +0100 |
User-agent: |
Gnus/5.13 (Gnus v5.13) Emacs/27.1 (gnu/linux) |
Daniel P. Berrangé <berrange@redhat.com> writes:
> On Tue, Nov 10, 2020 at 07:22:26AM +0100, Markus Armbruster wrote:
[...]
>> Command responses get sent strictly in order (even parse errors), except
>> for commands executed out-of-band.
>
> With out of band commands, how much runs in the background ? Is the
> JSON parsing still in the foreground, such that we can expect that
> even for OOB commands, a error response without a "id" is still
> received strictly in order.
Yes, you can. We made sure both errors and results flow through the
same pipeline[*]. The only fork is for OOB commands, and to take it,
the parser must have yielded a JSON object.
Use of exec-oob without "id" is foolish.
Pipelining commands without "id" is merely unadvisable.
[*] We queue parse errors along with successfully parsed values.
The QMP dispatcher dequeues, executes if it's a request, sends the
response, loop.
- QMP and the 'id' parameter, John Snow, 2020/11/09
- Re: QMP and the 'id' parameter, Markus Armbruster, 2020/11/10
- Re: QMP and the 'id' parameter, Daniel P . Berrangé, 2020/11/10
- Re: QMP and the 'id' parameter,
Markus Armbruster <=
- Re: QMP and the 'id' parameter, John Snow, 2020/11/10
- Re: QMP and the 'id' parameter, Markus Armbruster, 2020/11/11
- Re: QMP and the 'id' parameter, John Snow, 2020/11/19
- Re: QMP and the 'id' parameter, Markus Armbruster, 2020/11/20
- Re: QMP and the 'id' parameter, John Snow, 2020/11/20
- Re: QMP and the 'id' parameter, Markus Armbruster, 2020/11/23
- Re: QMP and the 'id' parameter, John Snow, 2020/11/30