On 13 November 2014 03:07, Mike Stump <mikest...@comcast.net> wrote: > On Nov 12, 2014, at 3:52 PM, Manuel López-Ibáñez <lopeziba...@gmail.com> > wrote: >> On 12 November 2014 22:41, Jakub Jelinek <ja...@redhat.com> wrote: >>> I think we had discussions on this topic, the important thing is that we >>> don't want to generate different warnings based on whether -save-temps has >>> been used, there could be just comment in between ) and ; etc. >> >> How can --save-temps influence whether ";" is in the same line as the >> 'if' or not? > > He was worried about: > > int main() { > if (1)/*hi*/; > } > mrs@alcmene:~/work1/gcc-c2e/gcc$ gcc -E tt.c > int main() { > if (1) ; > } > > but, no need to worry, there is a space in there. People that get this wrong > do );, without a space. They also don't have a comment in there either. The > eye disappears the ; when right next to ), but, if there is a space, the eye > can find it pretty easy.
But that has nothing to do with the change I proposed, which is to not warn if ";" is on a different line. The actual heuristics of Clang are more complicated and avoid even more false positives, but our location info is not precise enough yet to implement them. Cheers, Manuel.