* Ctor refactoring

* Renames

* Renames

* Cleanup

* ProjectExternalErrorReporter ctor

* Refactoring

* Remove dependency on VS

* Aggregation improvements

* Use ordinal comparer for project names

* Construct ProjectExternalErrorReporter lazily

* Fix off by one

* Fix issues

* Cleanup style

* Remove BlueSquiggleForBuildDiagnostic

* TableDataSource refactoring

* Remove DiagnosticData.Workspace

* Fix workspace reference

* Simplify GetTrackingPoints

* Simplify deduplication

* EqualsModuloLocation

* Simplify navigate to

* Specialize table item

* Fix assert

* Feedback

* Dedup TryNavigateTo

* Couple of renames

fixed a crash where diagnostic source is removed but asked new snapshot from error list in two different snapshot

removed unused using and use ProjectGuid from editor rather than defining our own since we have new bits now

editor team has refactored thier table control APIs

when error list data source and diagnostic service gets out of sync, let error list itself request a fix up rather than waiting diagnostic service to do so.

due to async nature of error list and diagnostic that doesnt share versioning, intermidiate out of sync can happen.

there is API cleanup going on and they are removing ShimTable APIs.

current design of error list is that when solution is changed, we will momentally get out of sync but eventually catch up and make error list and solution in sync. during those out-of-sync period, it is possible for error list to show blank entries in the list. this change is to reduce those cases.

more correct way to do this is making error list and diagnostic service to share same version. that requires a bit of change in both side, so I am trying cheap way first. since at the end, two should get into sync eventually even without this change, cheap way probably is enough.