On Wed, Jul 15, 2026 at 05:28:39PM -0700, Andrea Pinski wrote: > On Wed, Jul 15, 2026 at 5:07 PM Kees Cook <[email protected]> wrote: > > > > On Wed, Jul 15, 2026 at 12:06:17PM -0700, Kees Cook wrote: > > > On Sat, Jun 27, 2026 at 03:00:42PM -0700, Andrea Pinski wrote: > > > > On Thu, Jun 18, 2026 at 1:45 PM Kees Cook <[email protected]> wrote: > > > > > [...] > > > > > + > > > > > + > > > > > printf("\n================================================================\n"); > > > > > + printf("Passed: %d Failed: %d (%d total tests)\n", pass, fail, > > > > > pass + fail); > > > > > + return fail; > > > > > +} > > > > > > > > It would be a good idea to add a random generated testing here. See > > > > testsuite/objc.dg/gnu-encoding and testsuite/gcc.dg/compat for > > > > examples. > > > > > > Okay, thanks for the pointer; I'll see what I can come up with. I > > > generally dislike random testing since they may appear to be "flaky" > > > when they trip, but I'll give it a shot. > > > > I don't see anything we can do valuably here. There's no round-trip > > parsing going on (there's no independent oracle). For the existing > > mangling tests, I have hard-coded all the expected strings/hashes. I have > > gone through the TYPE enums, though, and added a few other missed types > > for the mangler. But I built a python-based fuzzer anyway, just to see if > > I could find any other combinatorial ICEs, in case I missed some mixture, > > and I could only crash the ToT GCC frontend. :P It's okay in 16.1? > > > > https://gcc.gnu.org/bugzilla/show_bug.cgi?id=126284 > > What about Adding something like gcc/testsuite/gcc.dg/compat/ where > you compile with 1 or 2 compilers? to make sure you can call the > function indirectly with the check? > Or is there some extra code that is needed to support that? > That is you compile one side with say clang and the other side with > GCC and you try to invoke the functions indirectly?
Oh, very cool. I didn't know about this infrastructure! I've added compat tests now. -- Kees Cook
