On Friday, January 3, 2020, Gregory Nutt <spudan...@gmail.com> wrote:
> > Is that because there are more places that don't do it than places that do? >> > Yes.. and I think just a matter of personal preference. > >> Because it seems to me that the right thing to do is to leave compliant >> code alone and bring non-compliant code into compliance with the standard, >> otherwise there is no standard. >> >> Thoughts? >> > > It has already been merged and a lot of water has passed under the bridge > since then. It would be tough to detangle now. > > I don't think it is am important part of the coding standard, so I don't > mind just removing these two sentences from the coding standard. > > I prefer having the (void) present because it says to me, "Don't worry, I > have thought this through and I don't need the return value." Otherwise, I > am suspicious of ignored returned values. > > But that is a pretty minor part of the coding standard. I made this change just because the usage is inconsistent through out the code base. the choice to remove the cast, not to add the cast, because: 1.It is more easier to remove the cast by command 2.This change is smaller than to add the cast 3.It make the code a little bit clean(e.g. memcpy... vs. (void)memcpy...) 4.The return value from many function don't indicate the pass/fail(e.g. memcpy return destination), it is reasonable to ignore it silently if the caller don't use that information. Greg > > >
