-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1


> - Could you give more information about the aim of this inclusion in th=
e
>   kernel, the advantages / drawbacks, etc. The reasons you gave here ar=
e
>   not that useful to me... more, I don't really understand what LD_PREL=
OAD
>   has to do with aRts and soundcard access. Your brief explanation are =
a
> bit confused, or I'm completely wrong about what you meant. From what I
> understood, I see something like : userland -> kernel -> userland ->
> kernel, and it doesn't make any sense to me...

Hi,=20

I'm not sure if I get it completely right, so correct me if I'm wrong.=20

First, usually only one application can access the oss device for playing=
=20
sound. Arts is capable of mixing the soundoutput of multiple sources=20
(applications). You can ie. hear the incoming message on ICQ while listen=
ing=20
to mp3s. There is a way to have non arts enabled applications (wich can o=
nly=20
access oss devices) emit sound through arts. You can do artsdsp appxyz an=
d=20
appxyz can access some virtual oss device for output. In fact it's not=20
accessing it but a library intercepts the oss stuff and pipes it to arts.=
=20
Maybe some conversion takes place underway. I wouldn't know. The library =
is=20
loaded with LD_PRELOAD. That means as far as I understood it, load the=20
library before the application links to the library it uses. Having somet=
hing=20
like that in the kernel would enable virtual dsp devices and you wouldn't=
=20
have to use the artsdsp app or the LD_PRELOAD mechanism.  Sound is now=20
arriving from the application (userland) to the kernel (new module) back =
to=20
userland (artsd) back to the kernel (artsd accesses the real oss device)=20
right through your soundcard out to your speakers.=20

Was that at least close to what's really happening ??? ;) Hope so ;)

I've got another question myself. The new module runs in kernelspace righ=
t.=20
What happens if it tries to access arts and arts died for some reason. Mo=
st=20
applications I know have serious problems with that. I'm worried about no=
t=20
only a blocked application but a blocked kernel...=20

Or does that module operate artsfailureproof?

Wouldn't the best thing be something like arts within the kernel? Sounds =
like=20
the ultimate solution to me. Then again... I'm not a kernel developer at=20
all...

Thanks,

Richard
- --=20
- ---
http://www.richardstevens.de

Unix IS user friendly, it is just selective about who his friends are.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iEYEARECAAYFAjwC5Y4ACgkQWQvEMJfcXlTFSgCcDKDWerTRgGFeEzMiP31OIB+Y
r4QAoJkTq1Mz0JroFS6Axef9NN9Qqqn5
=3DnjvU
-----END PGP SIGNATURE-----
 
>> Visit http://mail.kde.org/mailman/listinfo/kde-devel#unsub to unsubscribe <<