So your addition slightly goes into a different direction than the proposal of the OP (which both could be independently useful):   On 6/1/25 04:32, Thomas Krogh Lohse via Std-Proposals wrote: > > I’d love to hear your thoughts: >     * Do you see value in defining a "safe-by-default" C++ subset with > opt-in unsafe features? >     * Could something like this analysis model help enforce safety in > future directions for the language? Excellent idea but I already wrote a patented deterministic memory manager injected at compile-time using a source-to-source compiler   -----Ursprüngliche Nachricht----- Von:Phil Bouchard <boost_at_[hidden]> Gesendet:Di 03.06.2025 13:22 Betreff:Re: AW: [std-proposals] Memory Safety in C++ via Safe Subset and Static Analysis - Master’s Thesis Summary An:Sebastian Wittmeier <wittmeier_at_[hidden]>; On 6/3/25 02:08, Sebastian Wittmeier wrote: > For example, I have a pointer, which I pass to a function, which does a > lot of processing on the memory. > > Is the memory manager able to give a compile-time guarantee that the > function has not retained a copy of the pointer? My memory manager? No there are so much compile-time assertions you can do but what you're talking about is managed at runtime, where my memory manager flags memory blocks with the stack depth to help detect cyclic references and destroy them instantly.