On Fri, Jul 15, 2022 at 11:40 AM Aldy Hernandez <al...@redhat.com> wrote: > > This changes the ad-hoc dumping of ranges in the gimple pretty printer > to use the pp_vrange utility function, which has the benefit of > handling all range types going forward and unifying the dumping code. > > Instead of: > # RANGE [0, 51] NONZERO 0x3f > # RANGE ~[5, 10] > > we would now get: > > # RANGE [irange] long unsigned int [0, 51] NONZERO 0x3f > # RANGE [irange] int [-MIN, 4][11, MAX]
BTW, these are the global ranges, that for some historical reason unknown to me, are visible with -fdump-tree-all-alias. Yes, alias :-). For example: <bb 3> : _1 = (long unsigned int) i_54; # RANGE [irange] long unsigned int [0, +INF] NONZERO 0xfffffffffffffffc _2 = _1 * 4; _3 = ia_60(D) + _2; _4 = *_3; _5 = (long unsigned int) i_54; # RANGE [irange] long unsigned int [0, +INF] NONZERO 0xfffffffffffffffc _6 = _5 * 4; ... ... The VRP twins, as well as other ranger clients, continue displaying things as they were before. Aldy
