At Thu, 11 Jun 2015 23:43:53 -0400, Adam Goode wrote:
On Mon, Jun 8, 2015 at 7:23 AM, Takashi Iwai tiwai@suse.de wrote:
At Thu, 4 Jun 2015 21:04:50 -0400, Adam Goode wrote:
Expose the card number of seq clients. This allows interested userspace programs to discover the hardware device that backs a particular seq client. Before this change, the only way to get information about the hardware for a client was by using brittle heuristics.
Signed-off-by: Adam Goode agoode@google.com
The patch looks almost good. (One coding style fix about one-line if would be better, though.)
However, if we want to have more changes, I'd prefer the protocol version bump after all changes are merged. Are you going to submit any further changes?
thanks,
Takashi
Hi Takashi,
I fixed up the patch in v2. It fixes the style issue and delays updating protocol for now.
I think there might be another useful field to expose (the hardware midi device number), but I'm not sure yet. It is complicated since OPL3 does expose a client but does not use rawmidi. Otherwise, the rawmidi device number would be the obvious choice. I'm pretty sure I just need the card number to solve my problems in Chrome, but I will have to think about it.
IMO, we need to give the interface type (rawmidi, hwdep), too. Then the client can open the device and query more detailed information.
The only other interesting change to make right now would be a tstamp field (something like PCM), but it might be too complex for a quick fix. I do like the PCM timestamp mode selection, and would like to see such a thing in seq.
Yeah, this one would be more intrusive changes.
thanks,
Takashi