On Sat, Apr 07, 2018 at 10:24:40AM -1000, Joey Pabalinas wrote: > On Sat, Apr 07, 2018 at 05:31:40PM +0200, Luc Van Oostenryck wrote: > > I've moved the s/gcc/GCC/ change in a separate commit > > (cfr. git://github.com/lucvoo/sparse-dev.git sizeof-void ). > > Sure, splitting that change SGTM. > > > I've also seen that for GCC this warning is enabled by -Wpointer-arith. > > I think we should do the same. Mind to respin with this change? > > You mean change the flag name to -Wpointer-arith? I actually think > that's a great idea. Consistency with GCC flags would be good. > > > There is another thing that I think should be changed: > > taking the size of a void is quite weird/counter-intuitive and is > > a constraint violation in standard C. So, I think the flag should > > be enabled by default but the kernel should have -Wno-pointer-arith > > in CHECKFLAGS since it do pointer arithmetic on void (and a few > > places take the size of a function too). > > Well, although Sparse is a general semantic parser I was under the > impression that it was mostly used in the kernel. That is what I had > in mind when disabling it by default. OTOH, I can see the benefit in > doing it your way; anyone else have any opinions on if -Wno-pointer-arith > would be better made default in CHECKFLAGS leave this warning on > by default in Sparse? It's true that sparse is mainly used for the Linux kernel but as a checker it's normal that it's more strict than GCC. For non-kernel usage I think it's better to leave the warning enabled and adapt CHECKFLAGS. > I'm spinning up a new patch now; will send the V2 it once we decide > on if it should be disabled by default or not. Yes, fine. Cheers, -- Luc -- To unsubscribe from this list: send the line "unsubscribe linux-sparse" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
