On Feb 15, 2009, at 10:30 AM, fjahanian wrote: > > On Feb 14, 2009, at 5:39 PM, Chris Lattner wrote: > >> >> On Feb 14, 2009, at 9:59 AM, Fariborz Jahanian wrote: >> >>> I am sure you want the simplification to be equivalent to the >>> original (even though it >>> is not obvious to me right away :). >> >> I'm not sure what you're saying here. Are you saying that I broke >> something? > Sorry I wasn't clear. As long as a protocol which is both > unavailable and deprecated > (well just for the sake of argument). triggers two warning then > behaviour is the same. > protocol-atttribute-1.m tests that. So we are OK.
Yep, thanks Fariborz. I plan to refactor some of this code when I get a chance (maybe tomorrow). It would be nice to have a general "check decl use attributes" routine which would handle all of these consistently. attr(unavailable) doesn't make much sense on C declarations, but if someone did use it, it would be nice to be consistent and emit the warning. -Chris _______________________________________________ cfe-commits mailing list [email protected] http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
