On Sat, Aug 10, 2019 at 04:57:01PM -0500, Eric Blake wrote:
On 8/10/19 12:02 PM, Richard W.M. Jones wrote:
> See equivalent change for OCaml in
> commit d881d160e1cd9c9964782300a7652ffb4e506c27.
> ---
> generator/generator | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
The idea makes sense, but I'm not sure if the code is correct:
> diff --git a/generator/generator b/generator/generator
> index 0107724..0523f0a 100755
> --- a/generator/generator
> +++ b/generator/generator
> @@ -4135,7 +4135,8 @@ let print_python_binding name { args; optargs; ret;
may_set_error } =
> pr " Py_DECREF (py_args);\n";
> pr "\n";
> pr " if (py_ret != NULL) {\n";
> - pr " Py_DECREF (py_ret); /* return value is discarded */\n";
> + pr " ret = PyLong_AsLong (py_ret);\n";
> + pr " Py_DECREF (py_ret);\n";
This doesn't detect if the user returned a non-integer type (in which
case ret will be -1) - are we okay blindly returning -1 regardless of
whether the user returned actual -1 vs. if they returned some other
non-integer Python object that has no __int__ conversion? Or do we need
to use PyErr_Occurred() to distinguish between the two cases? This is
particularly interesting since we document that the C callback must
return -1 before any update to *err will take effect; do we want Python
to have to return -1 for that effect, or is it okay if python raises an
exception and we safely catch that as the way to translate to a C return
of -1?
Ah ha! Yes I did indeed observe this problem but couldn't understand
what was happening - this explains it.
Thanks,
Rich.
--
Richard Jones, Virtualization Group, Red Hat
http://people.redhat.com/~rjones
Read my programming and virtualization blog:
http://rwmj.wordpress.com
virt-df lists disk usage of guests without needing to install any
software inside the virtual machine. Supports Linux and Windows.
http://people.redhat.com/~rjones/virt-df/