On Wed, Jul 29, 2020 at 10:52:31PM +0200, Marcus Glocker wrote: Hello Marcus,
> Slightly adapted diff; Negative numbers can happen on controls. This looks really good to me, and a significant improvement on my original. I've tested everything I can think of and it all looks good, with one mild exception. UVC has separate controls for white_balance_temperature [int] and white_balance_temperature_auto [bool]. The former is only active if the latter is false. This probably makes devices about 30 seconds easier to develop, but it's not a nice UI for end users, so we've conflated them into one control. However, I now think we might need to display this conflated control slightly differently to others. For example consider this: $ video -d $ video -c brightness=128 contrast=128 saturation=128 gain=0 sharpness=128 white_balance_temperature=4000 That last line, I think, should probably be: white_balance_temperature=auto because although the device's white_balance_temperature is set to 4000K, we've turned auto white balance back on, so the device will choose whatever white_balance_temperature it feels like. If/when we get controls like zoom/exposure, they'll also probably want to be treated in the same way, so if there's a generic way of handling these in the code, it might be a nice bit of future proofing? However, this is not a huge deal: IMHO your diff is already good enough to go in tree! Laurie