On 08/03/2010, Ray Rashif <schivmeister(a)gmail.com> wrote:
On 08/03/2010, Ray Rashif
<schivmeister(a)gmail.com> wrote:
> On 08/03/2010, Paul Davis <paul(a)linuxaudiosystems.com> wrote:
>> On Sun, Mar 7, 2010 at 10:17 AM, Philipp <hollunder(a)lavabit.com> wrote:
>>> Hi there.
>>> The distro I'm using installs a script with jack that lets it run as a
>>> daemon.
>>
>> I know that I personally do not consider this an appropriate way to
>> run JACK, and I am irritated by distributions that do this. That
>> doesn't mean I'm right.
>
> It runs as (1) a dummy user or (2) a normal, real user if configured.
> It will fail if there is no user defined. If I recall correctly, you
> do not consider running jackd as root an appropriate way to run jack.
>
> But if you say you do not condone automatically starting jackd on
> bootup or having a daemon for it, then I'm sure this functionality
> will be gladly removed.
After further testing, it appears JACK_PROMISCUOUS_SERVER no longer
works. Paul, is that intentional?
Google tells me that at least Debian has since abandoned the method,
but I cannot find any credible source to verify that claim (only one
jackaudio.org article mentioning the variable in an old 0.109 release
notes).
Some peculiarities:
export JACK_PROMISCUOUS_SERVER
umask 0000
/usr/bin/jackd $JACK_PARAMS
Will only allow root to run jackd. Take note that the variable is only
exported, but not set.
export JACK_PROMISCUOUS_SERVER=""
umask 0000
/usr/bin/jackd $JACK_PARAMS
Will only allow root to run jackd, but will _not_ allow any client to
connect, not even root's. Is that because jackd detects the variable
and will refuse any connection? But that doesn't make any sense.
If there aren't any other solution to allow for a multi-user jackd
daemon, then I'll request for the package above to remove the scripts.
I think in that case netjack would be more of a solution for a
multi-user DAW machine (with thin clients).
--
GPG/PGP ID: B42DDCAD