[Bug c/55830] inline and __attribute__((always_inline)) treated differently for unused-function warning

[brooks at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55830



--- Comment #3 from Brooks Moses <brooks at gcc dot gnu.org> 2012-12-30 
21:50:08 UTC ---

Andrew: Oh, interesting.  So perhaps this is really a failure to warn (or

error?) for a case where "__attribute__((always_inline))" isn't used with

"inline", and a case of missing documentation on the always_inline attribute?



FWIW, the additional "inline" certainly doesn't seem to be needed in practice

for GCC to inline the code.