[mythtv] [mythtv-commits] Ticket #1598: AC3 passthrubroken with some alsa drivers (eg. snd_usb_audio)

[Mark Spieth]

>> just had a look at the patch.
>> I probably would have not changed the interfaces to all these functions 
>> and
>> used a comma separated alternate in the main device string to specify the
>> passthru device name.
>> ALSA:default,iec958:{ AES0 0x02 }
> I don't like this, it overloads the meaning of "audio device".
> If you do this, then before you know it every device type will
> have its own magic format. Issac's suggestion of additional
> constructors is better, you would still need to change the binary
> revision, but at least the interface is clear.
>
my argument is that its one device with 2 modes, the second of which is not 
generally used.
not sure why its magic. simple case where there is no difference between 
analog and dig passthru is still the same as before.
everything after the comma if it exists applies to the passthru mode. cant 
envisage anything simpler. the implementation of the audiooutputXXX classes 
stays the same as now. only the contructor interface does the "decoding" of 
the device string.
but fair enough.
additional constructors whilst maintaining the interface would still require 
equivalents in audiooutput. vmessy. less is better.

mark