Re: [patch] Lno branch merge part 3 -- ssa form updating improvements

[Zdenek Dvorak <rakdver at atrey dot karlin dot mff dot cuni dot cz>]

Hello,

> > Do you have a saner way how to do it for general code copying (like in
> > the jump forwarding algorithm presented by Kenneth on the gcc summit,
> > or in the trace formation)? Or do you want to solve each such case
> > specially, so that we have even more code duplication in the compiler
> > and even more places where we need to fix bugs?
> > 
> Yes.  The mental model is to not have passes that blindly move code
> around, hoping that a smart SSA patcher fixes things up afterward.  The
> idea is to make localized changes that each pass understands and can
> fix.

huhh?????? No you cannot do this.  Try thinking about say trace
formation algorithm.  You will soon notice that updating the ssa form
in this case indeed needs a fully generic algorithm that is able to
handle arbitrary insertions.

And NOOOOO -- I really do not want to solve the same problem for each
optimization separately.  What I offer is a fully generic way usable
at any place that needs code duplication/insertion, with well-defined
semantics.

Anyway, I withdraw from discussion with this mail; it becomes clear to
me that the only way how to persuade you is to let you do this mistake
and come up with this obvious solution in half a year again.
My only problem is that this blocks merging of lno branch, so sorry to
anyone who relied on it, but I cannot proceed with it just now.
Does not matter, there is really much other important work I can do on
gcc.

Zdenek