https://gcc.gnu.org/bugzilla/show_bug.cgi?id=125228
--- Comment #15 from Jonathan Wakely <redi at gcc dot gnu.org> --- (In reply to Patrick Palka from comment #11) > Cool, I didn't know about the #pragma push/pop_macro, which seems like > exactly what we want here. Want to post your patch to the mailing list? It makes it a bit more robust, but isn't actually necessary. Simply not including <cassert> guarantees that 'assert' isn't defined, and then we can define it to __glibcxx_assert.
