On May 17, 2012, at 12:53 PM, Paolo Carlini wrote: > On 05/17/2012 09:47 PM, Jason Merrill wrote: >> On 05/17/2012 05:06 AM, Paolo Carlini wrote: >>> On 05/17/2012 10:33 AM, Gabriel Dos Reis wrote: >>>> I am still puzzled by why we need to assert, as opposed to just >>>> ignore, unless we have a plan to make a wholesale move -- but even there I >>>> am bit >>>> nervous. >>> Eh, don't ask me ;) Anyway, in terms of testing that we aren't screwing >>> up anything in the C++ front-end, the testsuite just passed with the >>> below p3 attached. That's good. >> >> Yep, that's what the assert is for: testing that we aren't screwing up >> anything in the C++ front end. If it fires, it lets us know there's >> something still to fix. Sounds like it looks good so far. > If you like, I can install p3 now, but I think it would be a pity if we can't > have the warning_at bit because of that lone use in the ocbj front-end of an > explicit 'warning_at (0' (in objc-gnu-runtime-abi-01.c). Maybe Mike has > something to suggest?
Gosh, I'm not wedded to even having that warning. :-) The compiler knows what it has to do for codegen, it can eat and ignore the flag silently for all I care. I'd ask Nicola or Iain if they have any thoughts. One could also reasonably use: inform (UNKNOWN_LOCATION, ""); for it, if that helps.