On Aug 12, 2013, at 3:12 PM, Igor Galić <i.ga...@brainsware.org> wrote:
> > > ----- Original Message ----- >> Updated Branches: >> refs/heads/master ca7b063d0 -> f390a8e41 >> [snip] >> # Darwin pthread_setname_np: >> AC_MSG_CHECKING([for 1-parameter version of pthread_setname_np()]) >> -AC_COMPILE_IFELSE([ >> +AC_LINK_IFELSE([ > > given the nature of autoconf.. and C/C++ shouldn't we always prefer > AC_LINK_* for our tests? I know this will drag out the configure > phase immensely, but once it completes, we should be reasonably > sure that our project will actually compile. > > It might even make porting easier. I'm not sure. It's hard to make a general statement about that :) J
