get video input with torch

[mfulton26]

I see that getUserMedia can specify constraints but it doesn't seem to support constraining based on torch (even if I send in video: { torch: { exact: true } }; does it only support constraining on "image track" constraints and not "video track" constraints? I'm able to apply the torch constraint to a video track on a videoinput stream once I have one. I'm currently enumerating all devices and trying each one to find out which ones support torch which is a little slow and tedious. I'd much rather specify it as a constraint when calling getUserMedia or something. Please advise. Thank you.

[jan-ivar]

video: { torch: { exact: true } }

This should TypeError: "If CS contains a member that is a required constraint and whose name is not in the list of allowed required constraints for device selection, then reject p with a TypeError, and abort these steps."

torch is in mediacapture-image, which says: "None of the names of these photo capabilities and constrainable properties are in the list of allowed required constraints for device selection. Therefore, in getUserMedia(), these photo capabilities and constrainable properties can be constrained only with optional basic constraints and advanced constraints, but not with required constraints."

That said, video: { torch: true } should work.

... it doesn't seem to support constraining ... I'm able to apply the torch constraint to a video track ... once I have one

Is "it" Chrome? If so, please file a crbug to see if it's intentional, as "The User Agent SHOULD use the value of the computed fitness distance from the SelectSettings algorithm as an input to the selection algorithm".

[jan-ivar]

Closing as there does not appear to be a spec issue here. torch is in mediacapture-image.