- From: Wanming Lin via GitHub <sysbot+gh@w3.org>
- Date: Tue, 24 Apr 2018 07:15:39 +0000
- To: public-device-apis-log@w3.org
@kereliuk, thanks for your comments!
> [2] Should this have an ID attached to it? Or can you not have multiple types of the same sensor?
Each sensor type is unique and necessary, I referred to the [`device.mojom.SensorType`](https://chromium.googlesource.com/chromium/src.git/+/master/services/device/public/mojom/sensor.mojom#10) defined in Chromium Mojo module, which are equivalent to sensor subclasses defined in each concrete sensor spec, e.g. The [Magnetometer ](https://w3c.github.io/magnetometer/#model)sensor type has two associated Sensor subclasses, Magnetometer and UncalibratedMagnetometer.
We can remove a concrete mock sensor by passing `sensorType`(acts as an unique ID) to the URI `/session/{session id}/sensor/{type}`.
So coming with two errors in my proposal, the `SensorType` in `MockSensorConfiguration` dictionary is unnecessary; the naming of "SensorType" enum is improper as it's different from [Sensor Type](https://w3c.github.io/sensors/#model-sensor-type) defined in Generic Sensor spec, how about "MockSensorType"?
>[3] Same as [2] and should this have the fields getSensorShouldFail and startShouldFail? Or are those not that useful in the use case for this and its best to keep it minimal?
Yep, they are useless in [3] and we can keep it minimal.
> Opens:
[2] I think if there is a choice here I would lean towards not having a flag, makes it easier for config especially from a WPT infra standpoint, namely for wpt.fyi; however, this isn't a huge deal if there is a required command line flag.
For defining a required command line flag or not, is it up to browser vendors? I mean, do we need to specify this in spec?
--
GitHub Notification of comment by Honry
Please view or discuss this issue at https://github.com/w3c/sensors/issues/363#issuecomment-383828411 using your GitHub account
Received on Tuesday, 24 April 2018 07:15:42 UTC