For the record, I'm sure I'm encountered at least one x86 compiler in the past where this was a problem.
I used to have an issue where I was (stupidly) passing around a pointer-to-member through void*. The situation was the reverse of today: Visual Studio (5?) didn't use thunks and actually crashed the compiler, while gcc worked fine on x86 and ARM. -- To respond, reply to this email or visit http://mondrian.corp.google.com/8972115
