> This also has the unfortunate side-effect of generating incorrect code
> at the assembler level -- mostly because our assembler was missing support
> for VFP floating point.  

It would only affect a .double directive.  And fortunately, gcc never uses 
that.


> Any thoughts/comments?

It's unfortunate, but it's probably better to take the hit sooner rather 
than later.  Particularly as a lot of people tend to avoid '.0' releases 
in favour of letting others find the bugs for them :-)

R.