[ADDED] Selecting renderer config files at install
[Added] Selecting renderer config files at install
OK I find the 2.6.4 (and early 3.0 builds from git) implementation just fine here - excluding "Streamium" allows UMS to identify my "Naim" renderer profile - Thanks SubJunk!
I wonder if there is a way to extend this slightly?
For example - when a renderer is detected, UMS could match with all the available profile matches. If only one is found (or none for the default renderer) we behave as now.
If more than one is found, we need a way of identifying which is correct. There could be several strategies here:
- Always prefer one of the list of conflicting servers. This is pretty much like the current implementation, at least functionally. Without configuration, I guess this would just be the first on the list. A configuration option (and because the numbers are small, a radio-button would probably do at the UI) would then allow that preference to be changed to a specific renderer.
- Use the IP address. This would enable multiple conflicting devices to be used on the same network, as long as the user had control of DHCP (or assigned a fixed IP)
- Use the server name by resolving in DNS. My Naim, for example, uses DHCP, but registers a host name with DNS based on its MAC address - so it is always identifiable.
I guess it is also almost feasible to do this before a renderer is seen on the network - UMS could statically scan all the match strings to look for conflicts, but it can't know exactly what headers are returned so I think this might not be reliable.
This would require a small administration interface of course, and some additional configuration to add the "stage 2" identification term.
Anyway - just floating the idea
.
I wonder if there is a way to extend this slightly?
For example - when a renderer is detected, UMS could match with all the available profile matches. If only one is found (or none for the default renderer) we behave as now.
If more than one is found, we need a way of identifying which is correct. There could be several strategies here:
- Always prefer one of the list of conflicting servers. This is pretty much like the current implementation, at least functionally. Without configuration, I guess this would just be the first on the list. A configuration option (and because the numbers are small, a radio-button would probably do at the UI) would then allow that preference to be changed to a specific renderer.
- Use the IP address. This would enable multiple conflicting devices to be used on the same network, as long as the user had control of DHCP (or assigned a fixed IP)
- Use the server name by resolving in DNS. My Naim, for example, uses DHCP, but registers a host name with DNS based on its MAC address - so it is always identifiable.
I guess it is also almost feasible to do this before a renderer is seen on the network - UMS could statically scan all the match strings to look for conflicts, but it can't know exactly what headers are returned so I think this might not be reliable.
This would require a small administration interface of course, and some additional configuration to add the "stage 2" identification term.
Anyway - just floating the idea
.Re: Selecting renderer config files at install
A little feedback on your idea floating around...
As you have already said, if you want to involve the IP address in the detecting proces, it needs to be a predictable value upfront (with or without using DHCP), so that will be the weak point for creating a generic solution.
The MAC address is a more solid option I think, but the setback is that it will only work for the devices sitting on your local network (try using arp -a from the command line; it shows the MAC addresses within your network). So that should not be too hard to program and add to UMS I guess (although pretty low level, there must be some java class available that can take care of it) ...
Only question remains, does it fit the objective of UMS and the majority of the users? So is it worth the effort?
Anyway I'm pretty pleased how things are working right now in version 3.0.0.
As you have already said, if you want to involve the IP address in the detecting proces, it needs to be a predictable value upfront (with or without using DHCP), so that will be the weak point for creating a generic solution.
The MAC address is a more solid option I think, but the setback is that it will only work for the devices sitting on your local network (try using arp -a from the command line; it shows the MAC addresses within your network). So that should not be too hard to program and add to UMS I guess (although pretty low level, there must be some java class available that can take care of it) ...
Only question remains, does it fit the objective of UMS and the majority of the users? So is it worth the effort?
Anyway I'm pretty pleased how things are working right now in version 3.0.0.
Re: Selecting renderer config files at install
THis sounds like a GREAT idea... now can you enable and disable on the fly without having to stop/start UMS to refresh renders? if so i can really see this as a great testing conf files tool. Switching between two versions of confs on the fly!
Please remember to Upload your debug logs!
Universal Media Server Wiki!
"...functionality before beautification..."