[cfe-commits] r45976 - in /cfe/trunk: AST/Type.cpp test/Sema/complex-int.c

[Neil Booth]

Steve Naroff wrote:-

> Neil,
> 
> Just to be clear, my goal isn't to open up endless bugs. I've fixed  
> *many* bugs in clang over the past year and try to be careful when  
> making such changes.

Of course not, I'm sorry if I phrased my mail badly.
 
> Every time anyone touches some code, there is a chance of a regression  
> (especially since GCC extensions aren't wildly well specified).
> 
> Apparently I just created a regression. Thanks for bringing it to my  
> attention.

I just intended to point out that having IsIntegerType mean
something other than what it means in the standard will result
in us accepting things we shouldn't.  I don't know how to
construct a vector of int but it seemed from the context of the
patch that we still report that as an integer type.

Perhaps these shouldn't be integer types, but arithmetic types?
Then they wouldn't go in switch(), nor enum initializers, nor
integer constant expressions?  Complex ints are reasonable in
arithmetic expressions; I'm less clear where int vectors belong.

Neil.