This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: -Werror policy
- From: Roland McGrath <roland at hack dot frob dot com>
- To: Joseph Myers <joseph at codesourcery dot com>
- Cc: <libc-alpha at sourceware dot org>
- Date: Mon, 24 Nov 2014 15:55:22 -0800 (PST)
- Subject: Re: -Werror policy
- Authentication-results: sourceware.org; auth=none
- References: <alpine dot DEB dot 2 dot 10 dot 1411132331520 dot 5050 at digraph dot polyomino dot org dot uk> <20141114001831 dot 8C7CF2C3B16 at topped-with-meat dot com> <alpine dot DEB dot 2 dot 10 dot 1411140028450 dot 5050 at digraph dot polyomino dot org dot uk> <20141124221810 dot 391AC2C3B20 at topped-with-meat dot com> <alpine dot DEB dot 2 dot 10 dot 1411242313250 dot 11608 at digraph dot polyomino dot org dot uk>
> If we let things through without comments, I think they're just as likely
> to go through with unhelpful or empty strings. Hence my proposed patch
> having an unused argument only for a case where having some information
> inside the macro call rather than elsewhere nearby seems useful (GCC
> version with which the warning was observed, with a view to grepping for
> cases needing review).
I don't really agree. As a contributor, anyone is somewhat likely to
innocently overlook a policy requirement about comments. As a reviewer,
months from now, I am pretty likely to see someone's change and think,
"That's simple enough and obvious enough" and just approve it without
noticing that the change was of a sort that has a specific policy
requirement about comments--even when I was the big stickler insisting that
we have that policy in the first place! Conversely, when there is a
syntactic requirement to fill in a slot with something, then the
contributor is more likely to go and figure out what they were supposed to
put there. We expect contributors to be conscientious and well-meaning, so
they won't shirk a policy obligation intentionally, just through oversight.
Moreover, as a reviewer, I will notice the slot where the explanation is
supposed to be because an empty macro argument or a stubby-looking string
will stick out like a sore thumb and make me remember the whole issue about
the comments policy so I push back and insist on following the policy. I
can easily believe that you as a reviewer will never fail to demand the
commentary that a previously agreed-upon policy requires. But I am quite
sure that I as a reviewer will fail some proportion of the time.
Thanks,
Roland