[Patch] Java: Definite [Un]Assignment Issues

[Per Bothner]

Ranjit Mathew wrote:

> Cool! The above simplifies the code for
> check_bool_init() nicely - in fact, since the
> default processing is also the same, I could just
> remove the special treatment of MODIFY_EXPR
> as shown in the patch appended below.

Even better.

> Is the following OK for mainline then?

Yes, though I have one minor suggestion:

	  || (! DECL_LOCAL_FINAL_IUD (decl) && ! UNASSIGNED_P (before, index))
	  || (DECL_LOCAL_FINAL_IUD (decl) && ASSIGNED_P (before, index))))

should be:

	  || (DECL_LOCAL_FINAL_IUD (decl) ? ASSIGNED_P (before, index)
               : ! UNASSIGNED_P (before, index))

> If yes, can I update jacks.xfail appropriately?

Ok.
-- 
	--Per Bothner
per@bothner.com   http://per.bothner.com/