https://gcc.gnu.org/bugzilla/show_bug.cgi?id=121944
--- Comment #5 from Alejandro Colomar <foss+...@alejandro-colomar.es> --- (In reply to Alejandro Colomar from comment #4) > (In reply to Harald van Dijk from comment #3) > > (In reply to Alejandro Colomar from comment #2) > > > I read it more as a warning that this might be removed entirely in the > > > future, and that implementations are welcome to start warning about it, to > > > open the path for removal. > > > > My point is that that doesn't apply to your example. At the moment, by p1, > > "auto static as2 = 0;" does not make use of an obsolescent feature because > > all storage-class specifiers are at the start of the declaration, and p2 > > could have said that this is an obsolescent feature, but it doesn't, it only > > says that it may become a use of an obsolescent feature in the future. Oops, yeah, you're right in that it doesn't use the word obsolescent, but I think the core idea is the same; this will eventually change, rather sooner than later. Diagnosing is welcome for anything under 6.11 (Future language directions), I'd say. > > It still makes use of a different obsolescent feature: > > n3550::6.11.6p2: > > | Future standardization can change the auto storage-class specifier > | to a type specifier. > > Then we can combine both obsolescenct features to diagnose this.