Re: [PATCH v4 02/19] flake8: Enforce shorter line length for comments and docstrings

From: John Snow <jsnow@redhat.com>
To: Markus Armbruster <armbru@redhat.com>
Cc: Michael Roth <michael.roth@amd.com>,
	qemu-devel@nongnu.org, Eduardo Habkost <ehabkost@redhat.com>,
	Cleber Rosa <crosa@redhat.com>
Subject: Re: [PATCH v4 02/19] flake8: Enforce shorter line length for comments and docstrings
Date: Tue, 20 Apr 2021 14:06:47 -0400	[thread overview]
Message-ID: <e214622a-739c-126d-5899-b3cd433bfd9d@redhat.com> (raw)
In-Reply-To: <87fszp18u4.fsf@dusky.pond.sub.org>

On 4/17/21 6:52 AM, Markus Armbruster wrote:
> John Snow <jsnow@redhat.com> writes:
>> On 4/16/21 8:44 AM, Markus Armbruster wrote:
>>> John Snow <jsnow@redhat.com> writes:

>> It will be an eventual thing, though: I think we need to agree on a
>> style guide document and in that same series, fix up the instances of
>> defying that guide. I think it's important to pair that work, because
>> the ease of finding and fixing those style deviations will help inform
>> how pragmatic the style guide is.
> 
> Makes sense.
> 
> The introduction of "sphinxy" doc strings (starting with commit
> adcb9b36c) may have been premature.
> 

Somewhat premature, but what other format is there? It would have been 
worse to adopt Numpy or google style.

We'll dial it in over time, it will be fine.

>> I feel like it's something I want to do very soon, but not right now.
>> Maybe during the next freeze we can tackle it?
> 
> Whenever you're ready.
> 
> Until then, I feel we should try to minimize doc string churn.  Leave
> existing doc strings alone unless they're harmful.  Add new ones only
> when we believe they're helpful enough to justify some churn later.
> 

OK. After the expr comments, I actually didn't add very many. I think I 
add one or two for the parser because I had trouble understanding at a 
glance how it worked, but most of the tiny functions and helpers I left 
alone.

I barely touched schema.py, because it was complex and I had some 
visions of refactoring it a little to make some of the typing better later.

>>> Improvement, but mind PEP 8's "You should use two spaces after a
>>> sentence-ending period in multi-sentence comments".
>>>
>>
>> How important is this, and why? My existing prejudice is that it's only
>> a superficial detail of writing with no real impact.
> 
> Holy wars have been fought over less.
> 

:)

>> (Of course, a single space typist WOULD believe that, wouldn't they?
>> Those single-space typists are all the same!)
> 
> I offer three reasons:
> 
> * Local consistency
> 
> * Stick to PEP 8 unless you have good reason not to.
> 
> * It makes Emacs sentence commands work by default.
> 

For me, it's another thing in the category of "I don't actually mind 
either way", and can foresee myself accepting a patch using either style 
without comment. Inconsistency here doesn't really bother me unless it's 
inconsistent within a single docstring.

For QAPI, since you're the maintainer, I can adhere to your style. For 
the purposes of all Python code, though, I am not sure I want to bother 
enforcing it myself.

You're always welcome to post-edit anything I've written for 
typographical consistency as you see fit, I genuinely won't mind. (It 
saves me the trouble of having to copy-edit for something I am visually 
blind to.)

That said, I'll try to match your preferred style for QAPI at a minimum. 
I notice that emacs' reflow command does not always insert two spaces if 
a paragraph already sneaks in under the column limit; is there a way to 
*force* it to add two spaces?

>> Unfortunately, omitting it from flake8 means I'll probably also miss
>> cases where I or someone else have gone slightly over the limit for
>> docstrings, and doubt it will be enforced consistently.
> 
> I'm happy to correct the occasional minor style issue manually.
> 

If you accept that burden then I have no leg to stand on, I suppose :)

>>> 1.2. you may drop it.  I can pick it up and take care of it.
>>
>> This one, please!
> 
> You got it.
> 

Thanks! You can do that whenever, it won't interfere with anything in 
the interim.

--js