Observable comparers for SortAndBind

[RolandPheasant]

Add IObservable<IComparer> option to SortAndBind to enable changing the sort order.

var comparer = SortExpressionComparer<Person>.Ascending(p => p.Name).ThenByAscending(p => p.Age);
var comparerObservable = new BehaviorSubject<IComparer<Person>>(comparer );

var sortAndBind = _cache.Connect()
            .SortAndBind(out var list, comparerObservable)
            .Subscribe()

This represents the completion of SortAndBind.

@dwcullop / @JakenVeina

I know there were some previous conversations about SubscribeSafe but I skipped over that. My question is why has the pattern of .SubscribeSafe(Observer.Create... been introduced and when should it be applied? I ask because I used SubscribeSafe(observer) in this PR

[JakenVeina]

My question is why has the pattern of .SubscribeSafe(Observer.Create... been introduced and when should it be applied?

Bottom line is it's recommended when doing subscriptions within operators.

Subscribes to the specified source, re-routing synchronous exceptions during invocation of the System.IObservable1.Subscribe(System.IObserver{0}) method to the observer's System.IObserver`1.OnError(System.Exception) channel. This method is typically used when writing query operators.

In a practical sense, I tried to write up a small test scenario that shows this in action, and after an hour, I gave up. There's a LOT of baked-in safeguards within RX that normally make .SubscribeSafe() redundant. I do remember this was part of the solution for #668. I also suspect that using .SubscribeSafe() over .Subscribe() has no observable effect on performance, because the meaningful difference between the two is that .SubscribeSafe() does 1 or 2 extra if checks, on top of the many that .Subscribe() already does. Good news is we can easily benchmark this, if we want to be sure, now that we have a few operators that use .SubscribeSafe() and also have benchmarks written for them.

[JakenVeina]

NB?

[RolandPheasant]

https://www.britannica.com/dictionary/NB

[JakenVeina]

TIL

[JakenVeina]

Wow, this diff is obnoxiously terrible in GitHub, BTW. The syntax-aware diff in VS is so much better. Never seen a case this bad before.

[github-actions]

This pull request has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.