Re: [PATCH 3/4] {fast-export,transport-helper}: style cleanups

[Junio C Hamano <gitster@pobox.com>]

Felipe Contreras <felipe.contreras@gmail.com> writes:

> *You* are telling my that; it's *your* opinion and nothing else. It's

I saw a review comment that points out that the continuation lines
do not align, and you refused to say "ah, thanks for spotting" and
reroll [*1*], so even I do not want to do so in general, I had to
play the role of the arbiter.

My take on these style issues is this:

 * People made mistakes in the past while doing real work.  Big
   news: contributors and reviewers are not perfect.

 * They survived to this day because we do not do tree-wide "style
   fixes" for the sake of style fix, in order to avoid clashing with
   real work in flight.

 * Existing mistakes are not an excuse for adding new mistakes of
   the same kind, especially when they are pointed out during the
   review (this is not limited to "style issues").

I do not think I would reject a patch with minor style bugs like
unaligned continuation lines, if it were a patch that does real
work.

But a "style cleanups" patch that introduces new instances of style
breakage is a different matter.  It is clear that the original
(picked randomly):

        die ("Encountered signed tag %s; use ",
             "--signed-tags=<mode> to handle it.",
             sha1_to_hex(tag->object.sha1));

wanted the opening double-quotes of two lines and the "sha1" at the
beginning of the third line to align.  I see that is the local style
a "style cleanup" change should follow.

A patch that cleans up styles in preparation for a real work (like
this one) is a rare and precious occasion for us to really clean up
accumulated wart.  I do not want to see existing mistakes from other
unrelated parts of the codebase that have not been cleaned up as an
excuse to waste that rare occasion to do a good job of cleaning up.

So that is the arbiter's decision.  Call it *my* opinion or whatever
you like; it does not change anything.


[Footnote]

*1* That would have ended this thread without wasting everybody's
time.