--===============1713600194==
Content-Type: multipart/signed;
  boundary="nextPart4998447.a6BpxsQsZe";
  protocol="application/pgp-signature";
  micalg=pgp-sha1
Content-Transfer-Encoding: 7bit

--nextPart4998447.a6BpxsQsZe
Content-Type: text/plain;
  charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

Richard Dale wrote:
>On Sunday 26 March 2006 08:38, Andreas Silberstorff wrote:
>> And where DCOP was simple, D-Bus is designed to be interoperable
>>
>> > with other languages other than C++/Qt.
>
>DCOP is very easy to make interoperate with other languages. Athough
> some of the KDE interfaces use method overloading on the arg types,
> which is a bit more tricky.=20

Interestingly, D-Bus doesn't solve this problem either.

The specification and the library don't preclude overloading with=20
different parameters at all. It's actually up to each binding to decide=20
if and how they will support that. Some bindings might do automatic=20
type-conversion and call the same function; others may allow method=20
overloading by parameters (such as is the case of the Qt4 binding).

But, the consensus on the D-Bus mailing list and IRC channel is that=20
interfaces that are supposed to be interoperable shouldn't use that kind=20
of "fancy stuff". You'll still be able to use them for your own private,=20
non-interoperable communication.

>I hope it will be just as easy to interface with DBUS. I think it will
> be up to the developers of C++ apps to ensure they don't go overboard
> with nested custom structs or whatever with DBUS, and stick to things
> like simple string lists or hashes which can be easily represented in
> any of the above languages.

Indeed. D-Bus doesn't allow passing of arbitrary types. You can only send=20
a few, well-known types, plus a few compositions of those (arrays and=20
maps).

It also supports a type called "struct", but I have not gone deeper into=20
supporting it in the Qt4 binding. There are several limitations in using=20
it, including the fact that you can't generate code that receives=20
an "unnamed struct containing two ints and one string", for instance.

=46inally, if an application *really* needs to pass a complex type, it can=
=20
still do so. Just encode it with a QDataStream and send the QByteArray.=20
This will not be, of course, interoperable.

=2D-=20
Thiago Macieira  -  thiago (AT) macieira.info - thiago (AT) kde.org
  thiago.macieira (AT) trolltech.com     Trolltech AS
    GPG: 0x6EF45358                   |  Sandakerveien 116,
    E067 918B B660 DBD1 105C          |  NO-0402
    966C 33F5 F005 6EF4 5358          |  Oslo, Norway

--nextPart4998447.a6BpxsQsZe
Content-Type: application/pgp-signature

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2.2 (GNU/Linux)

iD8DBQBEJ8WXM/XwBW70U1gRAulwAJwO5SdVvJr8Ub8COPal01kGQou96QCfUs7g
cPie2KYTKxac/KB7mjXWQKA=
=mSc1
-----END PGP SIGNATURE-----

--nextPart4998447.a6BpxsQsZe--

--===============1713600194==
Content-Type: text/plain; charset="us-ascii"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

 
>> Visit http://mail.kde.org/mailman/listinfo/kde-devel#unsub to unsubscribe <<

--===============1713600194==--