Date: Tue, 18 Feb 2025 17:26:39 -0300
On Tuesday, 18 February 2025 16:53:27 Brasilia Standard Time Jennifier Burnett
via Std-Proposals wrote:
> What's actually being proposed (checking only specific annotated
> dereferences) _would_ require individual checks, however. Or else the trap
> handler would have to consult a table to determine if the nullptr
> dereference it's just caught is one that was annotated and should throw or
> is an organic nullptr dereference and should "merely" create a crash dump
> and/or break into an attached debugger.
Attempting to create such a crash handler in the Standard Library or the C++
runtime will collide with all existing crash handlers in applications and
libraries.
via Std-Proposals wrote:
> What's actually being proposed (checking only specific annotated
> dereferences) _would_ require individual checks, however. Or else the trap
> handler would have to consult a table to determine if the nullptr
> dereference it's just caught is one that was annotated and should throw or
> is an organic nullptr dereference and should "merely" create a crash dump
> and/or break into an attached debugger.
Attempting to create such a crash handler in the Standard Library or the C++
runtime will collide with all existing crash handlers in applications and
libraries.
-- Thiago Macieira - thiago (AT) macieira.info - thiago (AT) kde.org Principal Engineer - Intel DCAI Platform & System Engineering
Received on 2025-02-18 20:26:45