Re: [PATCH 6/7] qapi: Drop unnecessary horizontal spacing in comments

[Andrea Bolognani]

On Mon, May 02, 2022 at 10:50:07AM +0200, Markus Armbruster wrote:
> Andrea Bolognani <abologna@redhat.com> writes:
> > -# @writeback:   true if writeback mode is enabled
> > -# @direct:      true if the host page cache is bypassed (O_DIRECT)
> > -# @no-flush:    true if flush requests are ignored for the device
> > +# @writeback: true if writeback mode is enabled
> > +# @direct:    true if the host page cache is bypassed (O_DIRECT)
> > +# @no-flush:  true if flush requests are ignored for the device
>
> I'm no fan of horizontally aligning descriptions, because when you add a
> longer name, you either realign (I hate the churn) or live with the
> inconsistency (I hate that, too).

We seem to be in violent agreement on the topic, but it's apparent
that other people feel diffently :)

> I doubt changing to a different alignment now is useful.  The next
> patch, however, drops the alignment entirely.  Possibly useful.
>
> Thoughts?

My rationale for splitting things the way I did is that, if dropping
the horizontal alignment entirely was not considered desirable, we
could at least get rid of the extra whitespace. But if you think that
the benefit from the half measure doesn't offset the cost of the
churn it causes, I'm happy to drop these hunks and go straight from
the current status to no horizontal alignment at all in one fell
swoop with the next patch.

> > -# Since:  0.14
> > +# Since: 0.14
>
> This one is TAG: TEXT, whereas the one above is a multiple @NAME:
> DESCRIPTION.  Extra space in the latter can provide alignment.  Extra
> space in the former is always redundant.  I'd take a patch dropping
> these obviously redundant spaces without debate :)

Okay, I'll respin this so that the first patch drops all extra
whitespace in contexts where horizontal alignment is either not
attempted or not possible, and the second one implements the more
controversial changes.

-- 
Andrea Bolognani / Red Hat / Virtualization