Re: [MM] [PATCH] port-probe: fix crash in serial_probe_at_parse_response due to NULL response

[Ben Chan <benchan chromium org>]

I checked that |error| is set when |response| is NULL and response processors check for |error| first, so passing a NULL gchar* to response processors is fine. Uploaded a revised patch as you suggested.

On Thu, Jul 12, 2012 at 12:08 AM, Aleksander Morgado <aleksander lanedo com> wrote:

> Is it safe to assume that response processor(s) can handle NULL gchar*?
> I guess we need to check that in response processors as well.
>

Yes, there aren't so many response processors out there, as they're only
used during probing time (once the modem is created we use
mm_base_modem_at_command|sequence). The ones I checked will only assume
that response is valid if error is not set, so actually not a big deal
also to pass an empty string instead of NULL, but still I think we
should use NULL to indicate that error is set, so that we make a
difference between reporting an error and an empty AT reply.


--
Aleksander