search

LogmeinRescue / iOS-SDK

Rescue In-App Support iOS SDK
http://secure.logmeinrescue.com/welcome/webhelp/EN/SDKi/MobileSDK/c_riossdk_overview.html
Other
6 stars 7 forks source link

Observing flashlight availability #64

Closed cswagner closed 3 years ago

cswagner commented 3 years ago

At the beginning of a camera streaming session, the value for RescueCamera.sharedInstance.hasFlash is false and the value for RescueCamera.sharedInstance.flashState is .unavailable. Shortly after the session starts, these values switch to true and .off, respectively. However, for the corresponding AVCaptureDevice, it appears the torch is present and available from the beginning of the session (hasTorch and isTorchAvailable are both true).

What criteria are used to determine flashlight availability and what is the best practice for observing changes in its value?

Our current strategy is to poll RescueCamera.sharedInstance.hasFlash periodically. Is it possible to observe changes through key-value observation instead?

exceptioncatcher91 commented 3 years ago

Hi @cswagner,

When do you call RescueCamera.sharedInstance.enable(withAudio: <boolvalue>)? RescueCamera.sharedInstance.hasFlash returns false and RescueCamera.sharedInstance.flashState returns unavailable until you call this method. It may be useful if you see our example application here: https://github.com/LogmeinRescue/iOS-SDK/tree/master/Examples/RescueCameraExample You can get flashlight state changes via RescueCameraDelegate (func rescueCameraFlashTurnedOn(), func rescueCameraFlashTurnedOff()). These delegate methods are invoked when the torchMode property of the capture device changes. (RescueCamera.sharedInstance.flashState is also based on capture device’s torchMode.)



Daniel

cswagner commented 3 years ago

Thanks for sharing the example application, @exceptioncatcher91!

Like the example, we are calling RescueCamera.sharedInstance.enable(withAudio: <boolvalue>) in viewDidLoad, well before we show the button to toggle flash on/off. Right after calling that method, we still see RescueCamera.sharedInstance.hasFlash returning false.

Rather than observing whether or not flash is turned on/off, is there a way to observe flash availability in general?

exceptioncatcher91 commented 3 years ago

I looked into this problem a little bit deeper. Calling RescueCamera.sharedInstance.enable(withAudio: <boolvalue>) is not enough to get valid value via RescueCamera.sharedInstance.hasFlash, because value hasFlash is based on the avCaptureDevice's hasTorch property. When you would like to get the value of the hasFlash property, the avCaptureDevice is nil. Currently, there is no opportunity to observe that property. I need more time to decide how can i solve this problem for you, but is it going to be okay, if you get valid value immediately after you call RescueCamera.sharedInstance.enable(withAudio: <boolvalue>)? I think, if i can solve this problem in this way, then you do not need to observe hasFlash. Could you confirm this, please?

cswagner commented 3 years ago

We've removed polling the value of hasFlash and are, instead, checking it when streaming starts (via RescueCameraDelegate#rescueCameraStreamingStarted. This is sufficient for our use case.

Would it be possible to add handling flashlight availability to the example app? That might help shine light on any necessary API changes.

exceptioncatcher91 commented 3 years ago

Yes, its a good solution to check flashlight availability in RescueCameraDelegate#rescueCameraStreamingStarted. We extend our example application with flashlight handling. Thanks your remark!