On Mon, Aug 29, 2011 at 5:27 PM, Patrick Ohly <patrick.ohly(a)intel.com> wrote:
On Mo, 2011-08-29 at 15:12 +0200, Patrick Ohly wrote:
> > 1) In the case that we have no PnPInformation info we have...
> > deviceName = User-modifiable name
> > peerName = User-modifiable name
> Better leave the peerName unset. It's semantic will be "we know for sure
> that this device is a "<vendor>[ <product>]".
You decided to not implement it like this, did you?
That was an oversight. Attached, you'll find updated patch.
From your patch:
// This is the user-modifiable device name. Could be shown in GUIs, for
+ // This can be either the user-modifiable device name, vendor
+ // name, or product name (vendor + model). This depends on
+ // whether the device supports the Bluetooth Device ID profile
+ // and, if so, whether we have the model in the lookup table.
Was it simply easier to implement this way or do you prefer that
semantic of "peerName" in a template reported by the D-Bus server?
My concern is that if the D-Bus server always reports a value, the D-Bus
client won't be able to determine whether it has reliable information
about the manufacturer and model.
I agree that there is a need to be able for the client code to know
that they have reliable info or not. Setting peerName to empty is a
good way to do that.
I've also attached a n updated patch for the script. Just adds a
header with copyright and license info.