--Boundary-02=_pCH4+Y9CVtksQyB
Content-Type: text/plain;
  charset="utf-8"
Content-Transfer-Encoding: quoted-printable
Content-Description: signed data
Content-Disposition: inline

On Friday 06 June 2003 03:39, Ian Reinhart Geiser wrote:
> On Thursday 05 June 2003 08:19 pm, Gav Wood wrote:
> > since it's a kded service, to restart it all you have to do is go into
> > kcontrol and under service manager select the lirc entry and press the
> > stop button, followed by the start button.
>
> Please reconsider the poor design choice of a kded module for the followi=
ng
> reasons:
> 	any deadlocks or hangs in your code will end anyones kde fun quickly
> 	any crashes in your module will likewise cause  pain dejour
> 	dcopservices are easier to manage (start/stop/kill -9 when they go nuts
> and eat cpu and memory)
> 	any memory issues go away on the dcopservices restart vs kded modules

what you say is true, but likewise if all developers heeded these warnings =
or=20
thought they had any real bearing on the choice of design then there would =
be=20
no kded shared library modules at all.

in fact, there are many kded modules in shared library form, so clearly the=
re=20
must be some consensus that for a simple "daemon-like" extension a kded=20
module is a good design choice, no?

in addition to this i would imagine that the code is so straightforward tha=
t=20
it's probably possible (with the help of a few assertions and guards) to=20
_prove_ that it will not crash or hang.

what exactly are kded modules supposed to be for anyway?

gav
=2D-=20
Gav Wood <gav@kde.org>

codito ergo non satis bibivi

--Boundary-02=_pCH4+Y9CVtksQyB
Content-Type: application/pgp-signature
Content-Description: signature

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.1 (GNU/Linux)

iD8DBQA+4HCp7nE5x1pIEBQRAj6eAJ9dSm0Jr4Y4UqamhAQvdMQMBCIoHgCeO6Kf
/n3uczQlS3yxJDuvbRe1S8Y=
=BxiB
-----END PGP SIGNATURE-----

--Boundary-02=_pCH4+Y9CVtksQyB--