On Jan 29, 2008, at 7:47 PM, Eli Friedman wrote:
> So that I don't lose track, posting this testcase here:
>
> _Bool x; int a() {if (x);}
>
> Essentially, EmitScalarConversion is assuming that since the source
> and destination types are the same, no conversion is needed, which is
> wrong in this case. (Although maybe loading a boolean variable should
> immediately truncate it to i1; not sure what the right fix is here.)
You're right, we always want bool to be i1 when in a register (even
though it can be i8 or i32 in memory), so truncating it is the right
fix. I've committed a patch, thanks!
-Chris
_______________________________________________
cfe-dev mailing list
[email protected]
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-dev
