https://gcc.gnu.org/bugzilla/show_bug.cgi?id=125989
--- Comment #2 from Jakub Jelinek <jakub at gcc dot gnu.org> --- Created attachment 65076 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=65076&action=edit gcc17-pr125989-wip.patch Current untested WIP. There is one case in desig8.C I still need to look at, more testcase coverage (the current one is basically all the examples in the paper and nothing else) and decide what to do about the GNU [0] = designators vs. the new designator-initializer-list vs. designator-only-initializer-list, i.e. whether we allow { 1, [2] = 2 } or not and if we don't allow it, where do we reject it for C++29.
