Re: [PATCH 1/4] gdb/python: check for nullptr before calling evpy_add_attribute

[Simon Marchi <simon.marchi@efficios.com>]

On 9/14/26 3:24 PM, Tom Tromey wrote:
>>>>>> "Simon" == Simon Marchi <simon.marchi@efficios.com> writes:
> 
> Simon> Found by Claude while I was reviewing some patches: gdbpy_borrowed_ref's
> Simon> constructor asserts that the object is not nullptr.  Some call sites can
> Simon> in theory pass nullptr (although very unlikely, only on memory
> Simon> allocation failure kind of errors), fix them.
> 
> It's an oversight that these weren't already checked, and probably they
> are all latent bugs.
> 
> This is the kind of thing the safety API prevents, but (1) that's mostly
> on hold pending Longo's work; and (2) we aren't really ready to switch
> over the event stuff fully anyway (or at least for me it's one of the
> last things to do).

Ack.  It's not super important, but it's also easy to fix, so why not.

Simon