Re: [java-gnome-hackers] Arrays in signal handlers

[Guillaume M. <res...@gm...>]

2012/2/2 Andrew Cowie <an...@op...>:
> Guillaume has hit an interesting engineering problem.
I did :)

> There is a signal which returns an array of GFile. That doesn't seem
> like a problem, except that on the C side of the signal marshaling code
> we don't have anything to handle arrays.
To get the JNI signature, to call the right method, a C function is
called for each argument of the method returning the correct string.
We currently handle a various number of cases (integer, strings,
objects etc...) but we do not handle array.

The called C function is "bindings_java_typeToSignature". So the
problem is to know if a gpointer is actually an array. We can't do
that with our current "switch...case" (I don't think so). So we should
probably do something else.

> My best guess is to write a custom handler C side that takes the
> gpointer and turns it into a GList* of GFile. But we would still need
> code to handle that in the signal marshaller.
A custom handler could be the beginning of a fix but we will still
need to handle arrays or lists.

We could use a G_TYPE_STRV which is an array of strings but in the
marshaller this type would be G_TYPE_BOXED so we will be hiting
another problem. How can we know that a G_TYPE_BOXED is actually a
G_TYPE_STRV?
GList* would be tricky, I never saw any signal using a GList* as parameter.

> Guillaume could give more detail, I'm sure.
I hope I was understandable enough ;)

-- 
Guillaume Mazoyer - http://respawner.fr/