Re: [RFC] Number of insn slots used for method compilation

[Keith Seitz <keiths at redhat dot com>]

Tom Tromey wrote:

> Actually, I think it would be safe to update this.  After compiling a
> method we don't use it any more, do we?

I don't know, really. It might be, but I don't know when things like 
_Jv_InterpMethod::size gets called. Additionally, some of the friend 
classes/functions access this (defineclass.cc and verify.cc), and I 
don't know when this stuff is used, either.

If it could be hijacked, that would be great. That would save us quite a 
bit of memory, I suspect... An int per method for every interpreted 
method in a program -- that must add up in all but the most trivial of 
applications, no?

> Adding a new field is also fine though.

It certainly is a safer change, but not without consequences, unfortunately.

You tell me: you must have a better idea than I do! :-)

Keith