http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58950
--- Comment #9 from Eric Botcazou <ebotcazou at gcc dot gnu.org> --- > What I meant was, there are many places in the front-end that do: > > if (TREE_SIDE_EFFECTS (op0)) op0 = save_expr (op0); > > and I'd like to replace it with > > op0 = maybe_save_expr (op0); > > that would do the same thing to begin with. Now, if op0 is a long expression > without side-effects, the code I quoted will not call save_expr, it will > duplicate the expression and count on CSE for dedup, whereas we could > imagine having a single version to start with, that optimizations would > still be allowed to duplicate as usual if they consider it worth it. Yes, we do something like that in the Ada front-end, i.e. we call save_expr for complex expressions without side-effects, see e.g. utils2.c:gnat_protect_expr. But once the SAVE_EXPR is built, you *cannot* remove TREE_SIDE_EFFECTS on it, or else you'll run into nasty order of evaluation issues.
