https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113476
--- Comment #2 from Richard Biener <rguenth at gcc dot gnu.org> --- In fact it occurs elsewhere as well: ==1854== 81,616 bytes in 2 blocks are possibly lost in loss record 1,363 of 1,373 ==1854== at 0x505A1DF: operator new[](unsigned long) (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so) ==1854== by 0x1B8AB58: irange::maybe_resize(int) (value-range.h:645) ==1854== by 0x1B79873: irange::operator=(irange const&) (value-range.cc:948) ==1854== by 0x139779E: int_range<3u, true>::operator=(int_range<3u, true> const&) (value-range.h:1061) ==1854== by 0x2FAFBAE: phi_group::phi_group(phi_group const&) (gimple-range-phi.cc:84) ==1854== by 0x2FB0E7A: phi_analyzer::process_phi(gphi*) (gimple-range-phi.cc:460) ==1854== by 0x2FB0707: phi_analyzer::operator[](tree_node*) (gimple-range-phi.cc:314) ==1854== by 0x2F9E0E8: fold_using_range::range_of_phi(vrange&, gphi*, fur_source&) (gimple-range-fold.cc:949) ==1854== by 0x2F9C6AD: fold_using_range::fold_stmt(vrange&, gimple*, fur_source&, tree_node*) (gimple-range-fold.cc:604) ==1854== by 0x2F8D937: gimple_ranger::fold_range_internal(vrange&, gimple*, tree_node*) (gimple-range.cc:265) ==1854== by 0x2F8DC6D: gimple_ranger::range_of_stmt(vrange&, gimple*, tree_node*) (gimple-range.cc:326) ==1854== by 0x2F8E4D5: gimple_ranger::register_inferred_ranges(gimple*) (gimple-range.cc:486)
