Vladimir Sementsov-Ogievskiy <[email protected]> writes:

> On 29.10.25 12:08, Markus Armbruster wrote:
>> Vladimir Sementsov-Ogievskiy <[email protected]> writes:
>> 
>>> Hi all!
>>>
>>> Let's bring the documentation in line with the requirements. And
>>> do check these requirements in QAPI parser.
>>
>> Prior art:
>>
>>      01bed0ff14 qapi: Refill doc comments to conform to conventions
>>      7270819384 qga/qapi-schema: Refill doc comments to conform to current 
>> conventions
>>      209e64d9ed qapi: Refill doc comments to conform to current conventions
>>
>> These only touched prose.
>>
>> Your series also touches examples, is split per source file, and adds
>> code to enforce the rules automatically.
>>
>> Automatic enforcement makes a ton of sense.  Should've tried to code it
>> up long ago.  Much appreciated!  However, it's in the first patch.  It
>> needs to go last to not break bisection.
>>
>> I don't think splitting per source file is necessary.
>>
>> I'd prefer to put aside examples for now and focus on prose, since the
>> case for prose is much stronger.
>>
>> Since I already split off the prose changes for my own review purposes,
>> there's no need for you to do that again.  I'll take care of it.
>>
>> However, we need to adjust the enforcement code to skip examples.
>>
>> Examples are marked up with ReST directive qmp-example.  They look like
>> this:
>>
>>      # .. qmp-example::
>>      #
>>      #     the example
>>      #         text is
>>      #     indented
>>
>> The stupidest solution that could possibly work is to start skipping the
>> checks at
>>
>>      # .. qmp-example::
>>
>> and resume it at the first unindented line.
>>
>> This is of course a hack: it second-guesses the ReST parser.  I think
>> it's good enough.
>>
>> If we in a later step decide reflowing the examples is usful, the hack
>> goes away.
>>
>> Would you be willing to take care of that part?
>
> Of course. Will you send a detached prose changes, so I can base changed
> patch 01 on top of it?

Cool!  I'll make my patches somewhat more presentable after lunch, and
send you an URL to git-fetch them.


Reply via email to