option-value and usage in options-changed notifications

[maloel]

Prequel to DDS-related bulk-query changes:

• the rs2::options_list object now returns rs2::option_value rather than just the rs2_option
• rs2::option_value now holds a rs2_option_value const *
• rs2_option_value is a struct containing a type, id, and union of as_float etc.
• added rs2_option_type
• new APIs: rs2_get_option_value, rs2_delete_option_value, rs2_option_type_to_string, rs2_get_option_value_from_list
• on_options_changed notifications therefore now return the actual values seen so you don't need to query them again...

[Nir-Az]

While you are working on it, just adding that I see an exception when I stop debugging. (debug configuration)

[maloel]

While you are working on it, just adding that I see an exception when I stop debugging. (debug configuration)

I can take a look, but I don't understand the highlight you have or the meaning... is this in the private branch?

[OhadMeir]

While you are working on it, just adding that I see an exception when I stop debugging. (debug configuration)

I can take a look, but I don't understand the highlight you have or the meaning... is this in the private branch?

It also happened to me on the public branch before I completed my PR, I thought I handled all the cases.
The problem happened when closing the device, the callback iterated over a list that was cleared so the iterator was invalidated.

Nir showed me he had this error, I told him that you have a PR that changes the behavior, so don't know if it is still relevant.

[maloel]

when closing the device, the callback iterated over a list that was cleared

I'll try to reproduce. Technically I don't quite understand how... the device was destroyed while a callback was in progress? The join() wasn't blocking?
Or maybe the sensor was still alive, but its options were deleted?