On Sun, Jan 31, 2016 at 2:44 PM, Paul Davis <paul@linuxaudiosystems.com> wrote:

On Sun, Jan 31, 2016 at 6:33 AM, Kjetil Matheussen <k.s.matheussen@gmail.com> wrote:

I don't think Jack is the wrong solution for a DAW either. But Jack never got finished.
It has a wonderful API, but it shouldn't be a struggle for a program to create a jack client
if a jack server isn't running. (there were a lot of talk about this around 10 years ago,
but the end result never became as good as it should I think).

i am not sure what the problem is here. if the client doesn't specify anything, then the server will start automatically with the same parameters as it did last time. this has worked for years. no?

Well, I've never used it. It doesn't feel safe. There is no obvious place to

check that it does what it's supposed to.

I think the first program trying to create a client also should start the server. Not
just fork off a process, but actually run the server.
And if another program wants to create a jack client, it connects to the first client process,
which is the one running the server.

this seems a bit odd to me. if the first client is really just a client, why would it become the server?

If only one program produces sound, why would you also want to start a server?

what happens when the user closes it (or otherwise resets it)?

1. If it is the last client which is about to close, then the server closes as well.

2. If you close the client running the server while other clients are running as well,

then the server must be transferred to another client first, in a glitch-free and

spike-free manner, of course. This functionality would both feel natural

(the user doesn't have to think about the server concept at all), plus that it

would provide an enormous number of fun and interesting programming challenges

for the implementors of that functionality.

that's the whole point of using a control application that exists before (and almost certainly after) all other clients.

Furthermore, GUI should be built into libjack, so that you can call
jack_open_audio_driver_configuration_gui(), jack_open_audio_connection_configuration_gui(),
etc. inside your client.

I know there is something called libjackserver, but how many uses it?

libjackserver isn't what you think it is. you're thinking of the server API, used to start, stop and otherwise control a JACK server. it is the basis of jack2's dbus support, so basically everything using jack2 is using it. it is also present in jack1, and is also used internally there.

libjackserver is also used *all* the time.

Oh right, guess I knew that once though.

Does it do
all these things? How stable is it? In my opinion, there shouldn't
be any libjackserver, or jackd program, or qjackctl, only libjack.

this would simply lead *back* to jackd, since lots of people would want a program they could start in order to start jack, that would run before and after other clients. someone would write it, link it to libjack and it would become "jackd".

But my point is that you don't need the jackd program. Every client is also a potential

server (although the user doesn't know this), and since libjack provides functionalities for

configuring jack the same way for all clients, jackd is not needed. This way we can also

create formally specified error messages for the clients. Currently, if something goes

wrong, you have to dig around in the "messages" window in qjackctl, which may contain

some information that could help you make things work. It's really bad actually.