[macstl-dev] gcc 3.3 on YellowDogLinux

[Glen Low]

On 31/03/2005, at 11:55 PM, Michael LeBlanc wrote:

> I added the following constructor
> 	vec (vec <int, 4> v): data_ (v.data())
> 		{
> 		}
> to
> 	template <> class vec <boolean <int>, 4>:
> 		public impl::vec_base <__vector __bool int, boolean <int>, boolean 
> <int> >
>
> in impl/vec_altivec.h, and reinstated all the explicit uses of vec 
> <boolean <int>, 4>
>
> It compiles!
>
> I admit I am grasping at straws.  Is this appropriate?

Yes, um, it *is* grasping at straws. It's fairly dangerous to have 
implicit conversions like that, especially when they are lossy. 
Converting from int to bool is definitely lossy. It would screw around 
with overloads too. An admirable effort nonetheless! :-)
>
> Unfortunately it still traps at run-time.

Need to determine where the trap happens. Does YDL (1) turn Java mode 
on or off, (2) have denormal handlers? (if either Java mode is on OR 
you don't have denormal handlers, then the problem lies somewhere else. 
What kind of trap, which line of code etc.)

Cheers, Glen Low


---
pixelglow software | simply brilliant stuff
www.pixelglow.com
aim: pixglen