------- Comment #5 from fang at csl dot cornell dot edu 2007-03-08 21:49 ------- Ouch, this one seems particularly nasty to me... seeings as this isn't a regression (at least from 2.95), I don't expect this to be fixed for 4.2. Is there any chance of this getting attention on the (4.3) mainline?
I'm having trouble finding an elegant workaround to this -- I'm expecting a smart pointer-to-array class to 'do the right thing' during destruction (call the right operator delete []), but I can't just force it to call Class::operator delete [], if the Class doesn't override it. I just might have to "use the force" (SFINAE)... Incidentally, also known to fail 3.4.6, 4.0.1, 4.1.0. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29164