On Sat, 07 Nov 2026 10:08:17 +0100, "Theo de Raadt" <[email protected]> wrote: > > > Do I miss something? > > If I recall correctly, gcc used to have a variety of x-only violators. > > But the last big one was jump table code generation towards applications, but > also within it's own intermediate build. I have a feeling that disabling > jump tables by default, eventually resulted in an pure x-only able build > sequence, and this has been safe for a while. > > Future gcc updates must remember to disable jump tables. If you forget, > an intermediate step builds such code, confusion occurs, and maybe in an > overreaction USE_NOEXECONLY=Yes gets turned on again? :-) >
Yes, and it will be catch on update of gcc because it uses multi stage build. -- wbr, Kirill
