Hi Jonathan,

Thanks for fixing the tests and the guidance regarding the patch! I rebased the patch on top of that commit (b6b66006787b0991e94f15c7b5c56403f1eb85fb) and fixed all issues (I believe) which you pointed out. I also added the original reproducer to the test case as you suggested.

I configured as before with
../configure -prefix=/usr --enable-languages=c,c++ --enable-multiarch --host=x86_64-linux-gnu --build=x86_64-linux-gnu --target=x86_64-linux-gnu --enable-cxx-flags=-fno-rtti
and an additional configuration with -frtti instead to check nothing else breaks.

Now the test results look much clearer:

Without RTTI before applying patch:
=== libstdc++ Summary ===

# of expected passes 14560
# of unexpected failures 5
# of expected failures 95
# of unsupported tests 702

Without RTTI after applying patch:
=== libstdc++ Summary ===

# of expected passes 14562
# of unexpected failures 5
# of expected failures 95
# of unsupported tests 702

With RTTI before applying patch:
=== libstdc++ Summary ===

# of expected passes 14598
# of unexpected failures 2
# of expected failures 95
# of unsupported tests 683

With RTTI after applying patch:
=== libstdc++ Summary ===

# of expected passes 14600
# of unexpected failures 2
# of expected failures 95
# of unsupported tests 683

That looks logical to me.
I think ideally a test for the pointer-to-member case should also be added...