On Thu, Sep 12, 2019 at 11:08:43AM +0200, Paolo Carlini wrote: > Hi again, > > On 12/09/19 11:03, Paolo Carlini wrote: > > Hi, > > > > On 11/09/19 23:15, Marek Polacek wrote: > > > --- gcc/cp/pt.c > > > +++ gcc/cp/pt.c > > > @@ -26709,7 +26709,7 @@ build_non_dependent_expr (tree expr) > > > if (TREE_CODE (expr) == COND_EXPR) > > > return build3 (COND_EXPR, > > > TREE_TYPE (expr), > > > - TREE_OPERAND (expr, 0), > > > + build_non_dependent_expr (TREE_OPERAND (expr, 0)), > > > (TREE_OPERAND (expr, 1) > > > ? build_non_dependent_expr (TREE_OPERAND (expr, 1)) > > > : build_non_dependent_expr (TREE_OPERAND (expr, 0))), > > > > Looks like we would end up unnecessarily calling > > build_non_dependent_expr three times instead of two: probably is very > > cheap, probably the code is cleaner this way but I'm a little annoyed at > > this anyway, for the record ;) > > Sorry, I misread the code: normally TREE_OPERAND (expr, 1) isn't NULL_TREE > thus we are fine.
And I forgot to mention that build_x_conditional_expr has 6743 ifexp = build_non_dependent_expr (ifexp); 6744 if (op1) 6745 op1 = build_non_dependent_expr (op1); 6746 op2 = build_non_dependent_expr (op2); which means my fix should make more sense. -- Marek Polacek • Red Hat, Inc. • 300 A St, Boston, MA
