Re: [LAD] Failed to connect to session bus for device reservation
On Wed, 15 Feb 2017 07:45:49 -0800 (PST), Len Ovens wrote: >One of the biggest problems in linux audio is old information, linux >and the surrounding OS has changed. The biggest issue within this problem are wikis with links. When maintaining a Wiki, with too many links, nobody will check the content of all links, so we tend to not remove the work of others, because a link might be useful. Usually we only remove dead links, but not bad links. IMO the best approach is, wenn editing a Wiki, to avoid adding links, whenever possible. Another issue is self-responsibility. If a good documentation e.g. mentions /sys/bus/pci/drivers/ohci_hcd/unbind to unbind USB devices that share an IRQ with an audio interface, it shouldn't be that hard, to find out that the location changed to /sys/bus/pci/drivers/ohci-pci/unbind A problem could be to notice this or similar issues. At least I don't check such things all the times when making music. However, it's wise to not upgrade during a production, so there's no need to check everything. IOW when making music, it's not wise to upgrade Ardour, Guitarix etc., let alone to upgrade the kernel or something security related. I even wouldn't risk an openssl heartbleed alike fix, something absolutely necessarily for a lot of tasks, but completely irrelevant for a DAW. Btw. I'm a jack2, but jackd and not jackdbus user. Never change a winning team! Regards, Ralf ___ Linux-audio-dev mailing list Linux-audio-dev@lists.linuxaudio.org http://lists.linuxaudio.org/listinfo/linux-audio-dev
Re: [LAD] Failed to connect to session bus for device reservation
On Wed, 15 Feb 2017, Ralf Mattes wrote: Am Dienstag, 14. Februar 2017 18:58 CET, Len Ovensschrieb: No X11 -> No DBus Not true. I have done dbus-launch screen and used screen as my text only session manager with success. jack_control was able to start jackdbus, pulse was able to run and bridge to jack on a headless system. (it has been a while, I stopped because the system I was using was too memory strapped) Any of the terminals in the same screen instance will be able to communicate with any other. So it is possible. Yes, it is possible. But it also shows how little is known about dbus in the audio comunity (lack of documentation/quality of doxumentaion?). A naive (?) 'man jackd' won't even mention dbus. Want more ridicule? 'man jackdbus' : No manual entry for jackdbus See 'man 7 undocumented' for help when manual pages are not available. I can add more too. The tool for starting jackdbus, jack_control, has no docs. No man page, -h, --help do not work. running jack_control with no command gives a usage screen... and as it is a python script, the script itself is probably the best documentation there is (and still nor great). > It does mean learning more about your system... Hmm - did you? You suggest using dbus-launch even so the manpage of that program explicitly says "To start a D-Bus session within a text-mode session, do not use dbus-launch" and point to dbus-run-session ... Just enough to get things to work (and the experiment failed because I had only 0.3G ram, not because of dbus). However, having read the two man pages, I would still use dbus-launch with a text session manager like screen where I want to share one instance of dbus with a number of processes. Certainly I am no great sysadmin. I could also use more learning time on my system, but in general I have a system to use it, not learn about it's inner workings... So I learn only enough to get things to work, I copy lots of stuff others have done and ask questions. One of the biggest problems in linux audio is old information, linux and the surrounding OS has changed. I should actually try my test setup again with more memory and try both commands as I have such a box sitting here. -- Len Ovens www.ovenwerks.net ___ Linux-audio-dev mailing list Linux-audio-dev@lists.linuxaudio.org http://lists.linuxaudio.org/listinfo/linux-audio-dev
Re: [LAD] ?==?utf-8?q? Failed to connect to session bus for device reservation
On Wed, Feb 15, 2017 at 3:33 PM, Ralf Matteswrote: > > > Yes, it is possible. But it also shows how little is known about dbus in > the > audio comunity (lack of documentation/quality of doxumentaion?). > A naive (?) 'man jackd' won't even mention dbus. Want more ridicule? > 'man jackdbus' : > No manual entry for jackdbus > See 'man 7 undocumented' for help when manual pages are not available. > the presence of dbus support inside jackd has always been controversial. Jack1 does not (did not?) do this because I disagreed with integrating it directly into the jackd server. Jack2 does have dbus integration built in, but it uses the same manual page on most systems as jack1. it isn't intended that the user would ever need to be aware of or configure dbus interactions. the use case you're interested in is what we would call an edge case, and there's increasingly less tolerance for spending resources catering to these when there are so many non-edge features and functionality that are required by so many more people. ___ Linux-audio-dev mailing list Linux-audio-dev@lists.linuxaudio.org http://lists.linuxaudio.org/listinfo/linux-audio-dev
Re: [LAD] ?==?utf-8?q? Failed to connect to session bus for device reservation
Am Dienstag, 14. Februar 2017 18:58 CET, Len Ovensschrieb: > > > > No X11 -> No DBus > > Not true. I have done > dbus-launch screen > and used screen as my text only session manager with success. jack_control > was able to start jackdbus, pulse was able to run and bridge to jack on a > headless system. (it has been a while, I stopped because the system I was > using was too memory strapped) Any of the terminals in the same screen > instance will be able to communicate with any other. > > So it is possible. Yes, it is possible. But it also shows how little is known about dbus in the audio comunity (lack of documentation/quality of doxumentaion?). A naive (?) 'man jackd' won't even mention dbus. Want more ridicule? 'man jackdbus' : No manual entry for jackdbus See 'man 7 undocumented' for help when manual pages are not available. > It does mean learning more about your system... Hmm - did you? You suggest using dbus-launch even so the manpage of that program explicitly says "To start a D-Bus session within a text-mode session, do not use dbus-launch" and point to dbus-run-session ... Cheers, Ralf Mattes ___ Linux-audio-dev mailing list Linux-audio-dev@lists.linuxaudio.org http://lists.linuxaudio.org/listinfo/linux-audio-dev