[PATCH 2/2] gdb/python: Use tp_init instead of tp_new to setup gdb.Value

[Tom Tromey]

>>>>> "Andrew" == Andrew Burgess via Gdb-patches <gdb-patches@sourceware.org> writes:

Andrew> The documentation suggests that we implement gdb.Value.__init__,
Andrew> however, this is not currently true, we really implement
Andrew> gdb.Value.__new__.  This will cause confusion if a user tries to
Andrew> sub-class gdb.Value.

This all makes sense to me FWIW.

Andrew>    /* Nor in the cached Python objects.  */
Andrew>    Py_XDECREF (self->address);
Andrew> +  self->address = nullptr;
Andrew> +
Andrew>    Py_XDECREF (self->type);
Andrew> +  self->type = nullptr;
Andrew> +
Andrew>    Py_XDECREF (self->dynamic_type);
Andrew> +  self->dynamic_type = nullptr;

I don't think we use it much in gdb, but the Py_XDECREF + assign NULL
combo is Py_CLEAR.

https://docs.python.org/3/c-api/refcounting.html#c.Py_CLEAR

Tom