[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [PATCH for-5.2 4/4] qemu-option: warn for short-form boolean options
From: |
Markus Armbruster |
Subject: |
Re: [PATCH for-5.2 4/4] qemu-option: warn for short-form boolean options |
Date: |
Wed, 04 Nov 2020 14:43:13 +0100 |
User-agent: |
Gnus/5.13 (Gnus v5.13) Emacs/27.1 (gnu/linux) |
Paolo Bonzini <pbonzini@redhat.com> writes:
> On 03/11/20 17:08, Daniel P. Berrangé wrote:
>>> +Short-form boolean options (since 5.2)
>>> +''''''''''''''''''''''''''''''''''''''
>>> +
>>> +Boolean options such as ``share=on``/``share=off`` can be written
>>> +in short form as ``share`` and ``noshare``. This is deprecated
>>> +for all command-line options except ``-chardev` and ``-spice``, for
>>> +which the short form was in wide use.
>>
>> So IIUC, the short form was possible to use for absolutely /any/
>> boolean property ?
>
> s/boolean// (yikes)
Yup. "-device virtio-blk,drive=blk0,serial" gives you the lovely serial
number "on".
>> IMHO if we're going to deprecate short forms, we should do it
>> universally including chardev and spice. Arguably spice/chardev
>> are the most important ones to give an explicit warning about
>> precisely because their widespread usage means a heads up is
>> important to users.
>
> Chardevs will probably become user-creatable objects; for -spice I was
> hoping that it would be QAPIfied as "-display spice" which does not
> support short forms, but I'm not sure if Gerd agrees. In both cases,
> the problem would be taken care of in a different way.
Taken care of only if we deprecate -chardev and -spice wholesale, not if
we keep them forever as sugar for -object.
> I can certainly warn for all of them, but I was thinking of the
> lowest-impact option for 5.2 since we're already in soft freeze.
I'm quite interested in getting rid of this sugar. I'm not particular
on how exactly, and I understand your reluctance to mess with 5.2.
[PATCH for-5.2 2/4] qemu-option: move help handling to get_opt_name_value, Paolo Bonzini, 2020/11/03
[PATCH for-5.2 3/4] qtest: escape device name in device-introspect-test, Paolo Bonzini, 2020/11/03