alignment test case

[Andrew Haley <aph at redhat dot com>]

Andrew Haley writes:
 > Tom Tromey writes:
 >  > Here's a simple test case that shows the alignment bug.
 >  > 
 >  > When I run it, I get:
 >  > 
 >  > fleche. ./F
 >  > 4294967296
 >  > true
 >  > null
 >  > 232300
 >  > true
 >  > maude
 >  > 
 >  > This is with a pristine 3.3 build I did yesterday.
 > 
 > Great, thanks.  I'm on it.

Oops.  I found the bug, and it's pretty trivial.  If a superclass has
smaller alignment than a subclass -- which can happen because Object
is only 4-aligned -- my patch is wrong.

I'll commit a better patch tomorrow, one that can only *increase*
alignment, not *decrease* it.

Andrew.