https://gcc.gnu.org/bugzilla/show_bug.cgi?id=125565
--- Comment #7 from Tomasz Kamiński <tkaminsk at gcc dot gnu.org> --- > Thanks for the comments. I am not sure I see the rationale behind the current > semantics. >From outside discussion, if you have a map that caches weak pointers (e.g. last recently used cache), you can use the pointer to remove related entries. > Would adding a non-standard (to start with) extension that allows the user to > reset the weak pointer explicitly be an option? I would suggests writing a proposal for the standard instead, with your motivation. (The function should be protected).
