[Bug 698837] v4l2: cache CIDs until the device is opened

[GStreamer (bugzilla.gnome.org)]

https://bugzilla.gnome.org/show_bug.cgi?id=698837
  GStreamer | gst-plugins-good | git

--- Comment #9 from Michael Olbrich <m.olbrich at pengutronix.de> 2013-05-21 08:19:57 UTC ---
(In reply to comment #8)
> Review of attachment 244836 [details]:
> 
> ::: sys/v4l2/gstv4l2object.c
> @@ +492,3 @@
>            G_PARAM_READWRITE | G_PARAM_STATIC_STRINGS));
> +
> +  g_object_class_install_property (gobject_class, PROP_EXTRA_CONTROLS,
> 
> Please add a gtk-doc documentation block above this to explain how it's meant
> to be used and what it does, also mention CID in the long description of
> g_param_spec_boxed()

Ok

> 
> ::: sys/v4l2/v4l2_calls.c
> @@ +276,3 @@
> +      case V4L2_CTRL_TYPE_INTEGER_MENU:
> +      case V4L2_CTRL_TYPE_BITMASK:
> +      case V4L2_CTRL_TYPE_BUTTON:{
> 
> Are these all "int" typed? Not long, int64, some unsigned variant of these?

Yes, these are all 32bit types. VIDIOC_S_CTRL only has a s32 for the value.
All other types require VIDIOC_S_EXT_CTRLS and that's a lot more complex and I
cannot test it.

> @@ +284,3 @@
> +            control.name[i] = '_';
> +        }
> +        GST_WARNING_OBJECT (e, "adding generic controls '%s'", control.name);
> 
> This should be GST_DEBUG_OBJECT(), right? It's not a warning

Of course. I changed it a few times while developing and commited the wrong
version.
Maybe GST_INFO_OBJECT so the available contols are easier to find?

> 
> @@ +855,3 @@
> +  if (!G_VALUE_HOLDS (value, G_TYPE_INT)) {
> +    GST_WARNING_OBJECT (v4l2object,
> +        "'int' value expected for control '%s'.", g_quark_to_string
> (field_id));
> 
> Return from this function here, otherwise you call g_value_get_int() on
> something that is not an int

of course

-- 
Configure bugmail: https://bugzilla.gnome.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug.
You are the assignee for the bug.