hi Dominique;
Bastian Winkler recently filed a bug[1] with a patch to implement
ClutterScriptable in ClutterInterval. the patch has been accepted, and
will be pushed to the 1.16 stable branch. I need to make a new 1.16
release soon(ish), so it'll be in the next stable release of Clutter.
ciao,
Emmanuele.
[1] https://bugzilla.gnome.org/show_bug.cgi?id=711573
> _______________________________________________
On 7 November 2013 14:14, Dominique Bureau <dbureau gmail com> wrote:
> Wow. Apologies for the funky characters in the post. Remember to never use
> Outlook on a Mac, it's evil.
>
> --
> Dominique Bureau
>
>
>
> On Wed, Nov 6, 2013 at 1:47 PM, Dominique Bureau <dbureau gmail com> wrote:
>>
>> Hello folks!
>>
>> I¹m currently looking into moving my scripted ClutterStates to
>> ClutterTransitions and found out that the ClutterInterval part of a
>> transition cannot be scripted (yet).
>>
>> So far, I¹ve found out that the ³initial² and ³final² properties are not
>> being handled properly in clutter-script-parser.c. While in
>> _clutter_script_parse_node(), the value is seen as a boxed type (as per
>> the object¹s property param spec), but the only boxed type value being
>> handled in _clutter_script_parse_node() is a ClutterColor.
>>
>> From there, I figured I got 2 options:
>>
>> * Implement ClutterScriptable interface in ClutterInterval
>> * Handle other boxed types in _clutter_script_parse_node()
>>
>>
>> In both cases, I¹m having issues related to the boxed type (which I am no
>> expert with).
>>
>> By implementing the scriptable interface, from parse_custom_node() I can
>> easily get the json node value and pass it off to
>> clutter_interval_set_*_value(), which in turn will handle the conversion
>> of the value into whatever we need. I¹m purposefully always returning
>> FALSE and not implementing the set_custom_property() to avoid other
>> problems that occur from letting ClutterScript handle the value (more on
>> that below).
>>
>> On the other hand, if I try to handle this in
>> _clutter_script_parse_node(), I¹m having a problem when the value is a
>> fundamental type. From what I can understand, I can¹t initialize a Gvalue
>> to be a plain abstract boxed type into which I would copy a fundamental
>> type.
>>
>>
>> I¹m having a bit of a feeling that using a g_param_spec_boxed property is
>> a bit hackish when dealing with fundamental types. At least, that¹s what I
>> understand from GLib¹s warning message when trying to pass a G_TYPE_DOUBLE
>> Gvalue through g_object_set_property():
>>
>> -----
>> (interval:19221): GLib-GObject-WARNING **: unable to set property
>> 'initial' of type 'GValue' from value of type Œgdouble'
>> -----
>>
>> Any hints/ideas?
>>
>> Thanks
>>
>> ‹
>> Dominique Bureau
>>
>>
>
>
> clutter-list mailing list
> clutter-list gnome org
> https://mail.gnome.org/mailman/listinfo/clutter-list
>
--
W: http://www.emmanuelebassi.name
B: http://blogs.gnome.org/ebassi/