[alsa-devel] [PATCH 21/30] ALSA: oxfw: Change the way to name card

Takashi Iwai tiwai at suse.de
Tue Dec 2 07:41:19 CET 2014


At Tue, 02 Dec 2014 15:17:25 +0900,
Takashi Sakamoto wrote:
> 
> On Dec 2 2014 00:21, Takashi Iwai wrote:
> >>> As for the shortname, it is not an internal identifier but typically
> >>> displayed to identify the device for the user.  Something like
> >>> "FWSpeakers" might be recognizable, but I'd prefer the old "FireWire
> >>> Speakers".
> >>
> >> In this case (the old name), 'Speakers' is an identical name such as
> >> 'hw:Speakers'. IMHO, this is a bit strange because it's usual, not
> >> identical.
> >
> > You see now that you're breaking user's setup, don't you?
> > This is another reason why you shouldn't change the short name
> > string.
> 
> In this case, it depends on the estimated number of actual users for 
> this device.

No.  The number doesn't matter unless it's proved to be really no
regression.

> At least, snd-firewire-speakers cannot handle this device 
> correctly, while there're no bug reports, for 4 years. I think this is 
> an evidence of few users.

Did you get reports that the name hw:Speaker is bad, too?
If not, why do you insist to "fix" this at all?  It's just because you
feel so -- which is no objective measurement, either.

> ...although havind said so, it's not my favorite to spend more time to 
> such a small things, instead of more important things. I'll keep the 
> naming rule what it was in next patchset.

There is no more important things than fixing a regression.

Just fix it for the next patchset, or patches won't be accepted.
That's so simple.

I'm giving a strong word here because you might miss the point: in
such a rewrite, the primary item you have to care is to avoid
regressions.  Supporting other devices, the cleanness of the code, or
even the correctness of the code, is hence in the second place.


thanks,

Takashi


More information about the Alsa-devel mailing list