Re: note: JasminClass change

[Patrick LAM <plam@sable.mcgill.ca>]

On Tue, 27 Apr 2004, Chris Pickett wrote:

> I think that
>
> .super
>
> by itself implying super_class = 0 is better.  Either that or an
> explicit '.no_super'.  You still need something that writes the
> super_class field (well, it's 0, so you don't _really_ need it if the
> data is initialized).  Also, if the entire line is optional, then you
> don't get parse errors if you forget .super for a non-java.lang.Object
> class, which is bad.

I think it's ok to let assembler users shoot themselves in the foot.  (I
would be less in favour of letting Soot users shoot themselves in the foot
that easily).  I think .super by itself isn't so great (it looks like
there's something missing).  .no_super is ok if the syntax says .super foo
| .no_super.

You won't get parse errors for missing a .super line for
non-java.lang.Object.

pat