Allow != and == of Time objects with different sources

[sloretz]

rclcpp::Time forbids comparisons between objects with different clock types. It raises an exception when that happens. This makes sense for <, <=, etc, but I think == and != could be defined by saying Time objects with different clock types are never equal to each other.

I think this might simplify some of the logic for handling Time with different time sources here: locusrobotics/fuse#307

@ros2/team thoughts?

[alsora]

I think many users who compare times with different clocks are doing that by mistake, so having an exception in all operators make sense.

For the few use-cases where you really need to handle different clocks, it should be sufficient to use a custom comparison function that checks the clock type first, rather than modifying the core library.

[wjwwood]

I agree with @alsora. However, I would be ok with a different comparison function (like a named one, e.g. .clock_and_time_equal() or something) being added as a method on Time to reduce duplication across code bases.

[fujitatomoya]

I am with @alsora , probably throwing exception is better for user application.

[sloretz]

All feedback so far has pointed towards not allowing any kind of comparison between Time points with different clock types. Internally I received feedback that's pointed towards Time points for different clock types should have different types so that enforcement can happen at compile time. Given that direction, I'll close this PR.