Re: [PATCH] config.mak.dev: enable -Wunused-function

From: Jeff King <peff@peff.net>
Date: Oct 18, 2018, 17:09 UTC
Message-ID: <20181018170934.GA21138@sigill.intra.peff.net>
In-Reply-To: <CACsJy8BOL+QOb3Vdh8RKbra-DDbJVyznTbCJgjK5h8L2y0H7yg@mail.gmail.com>

On Thu, Oct 18, 2018 at 05:48:16PM +0200, Duy Nguyen wrote:

Show 16 quoted lines

> >   - conditional compilation, where we may or may not need a
> >     static helper. These generally fall into one of two
> >     categories:
> >
> >       - the call should not be conditional, but rather the
> >         function body itself should be (and may just be a
> >         no-op on one side of the #if). That keeps the
> >         conditional pollution out of the main code.
> >
> >       - call-chains of static helpers should all be in the
> >         same #if block, so they are all-or-nothing
> 
> Grouping is not always desired because it could break better function
> layout. Have a look at read-cache.c where _ieot_extension functions
> are #if'd because the only call sites are from pthread code (#if'd far
> away).

True, though as long as they are triggered by the same set of #if conditions, that is fine. Putting them in the same block is just an easy way to make sure that is the case. ;)

> In this particular case though I think we should be able to avoid so
> much #if if we make a wrapper for pthread api that would return an
> error or something when pthread is not available. But similar
> situation may happen elsewhere too.

Yeah, I think that is generally the preferred method anyway, just because of readability and simplicity.

> Having said that, if people do consider MAYBE_UNUSED before #if'ing
> everywhere (and opening up more conditional build problems in future),
> I think this change is fine.

I'd like to use it as a last resort, certainly. Mostly the fact that we compile cleanly _now_ makes me think that it probably won't be that hard to keep it going.

I think the biggest potential problem with this is going to be obscure configurations where some functions are used or not used. So somebody silencing a compiler warning may inadvertently break another case if they're not careful. But that's already a problem to some degree (and part of why we try to push the conditionality out to the whole-function level).

-Peff