This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Using fold() in frontends
- From: kenner at vlsi1 dot ultra dot nyu dot edu (Richard Kenner)
- To: rasky at develer dot com
- Cc: gcc at gcc dot gnu dot org
- Date: Mon, 7 Mar 05 13:01:01 EST
- Subject: Re: Using fold() in frontends
It looks like the general consensus is that we should not use fold()
anymore in the frontends
I haven't seen that and certainly don't agree with it. The earlier we can
get rid of junk trees, the better. There is no point in having separate
code in a front end to fold "1+1" nor a point in having that be left in the
tree in that way until gimplification.
What I thought we agreed on was *only* that if a language has rules about
what *must* be a constant, we should not be depending on fold to be part
of the enforcement for those rules.