Re: Adding new modifier to GCJ front-end

[Angelo Corsaro <corsaro at cse dot wustl dot edu>]

Hello Tom,

On Sun, 2003-09-28 at 15:35, Tom Tromey wrote:

> Angelo> What are the rationale behind having the modifier sorted?
> 
> Hmm, there may be some other subtle dependency on the order.  If I
> ever knew, I forgot.

I figure it out, so it is probably worth to add some docs
into the parse.y file. The reason why those token have
to be maintained in that exact order is that all the
algorithm that use them rely on their relative position 
in order to create the corresponding code. For instance,
when the lexer finds a "protected" modifier passes to
the parser the code "3", then the parser just does 
1 << 3 in order to create the ACC_PROTECTED flag.
   
> Angelo> What is the purpose of PAD_TK?

After figuring out that, it was apparent that the PAD_TK
is kept there just to make thing working since the code that 
corresponds to its position (it is the 10th pad) is already
used by ACC_INTERFACE.

Thanks a lot for your help. 

Cheers,
   Angelo


-- 
Angelo Corsaro
Department of Computer Science
Washington University
http://tao.doc.wustl.edu/~corsaro