A handful of csprojs were referencing version 8.0.0. Since that version
targets net8.0, I have no idea why it didn't cause the build to fail.
Instead, this framework mismatch led to a mysterious runtime crash with
the following error, despite the dll in question being present in the
build dir:
Unhandled exception. System.IO.FileNotFoundException: Could not load file or assembly 'System.Diagnostics.DiagnosticSource, Version=8.0.0.0, Culture=neutral, PublicKeyToken=cc7b13ffcd2ddd51'.
The system cannot find the file specified.
It is worth noting that this crash would only happen for exes published
as part of the DataStreamSuite.sln with the "--self-contained" flag. For
example, the Bts.DataStreamSuite.Server exe would always fail if
published via:
dotnet publish --configuration Release --sc --framework net7.0 --runtime linux-x64 <DataStreamSuite.sln>
The former is the base class for the latter, so this will widen the umbrella to catch a few corner cases
that would otherwise been handled by the generic "Exception" catch block. One corner case I know
of is when an op is canceled via "CancellationToken.ThrowIfCancellationRequested". In that case,
an OperationCanceledException is thrown (not a TaskCanceledException).
https://stackoverflow.com/questions/34359129/operationcanceledexception-vs-taskcanceledexception-when-task-is-canceled
Previously, animations were expressed through polymorphism (derived from the
base class CellAnimation), but the new impl uses composition. This 'OfType' call
was leftover from the polymorphism implementation, an attempt to linq cast
elements of base class (CellAnimation) to BackgroundAnimation.
This matches the namespace used in spark_websocket csproj. Eventually,
I'd like to use conventional capitalization (TitleCase namespaces/types,
etc.), but let's shelf that until we have enough time/energy to do a
more comprehensive refactoring.
Also add 200ms throttle. Contracts shouldn't change that often, so we don't need
lightening speed here. Furthermore, such changes may trigger expensive cpu ops.
Note: this change also replaces refs to ThreadedBlackboard w/ IBus (the
former is somewhat tethered to the dispatcher thread, particularly by
its timer_callback).
This shouldn't happen often, but if it does, we should know about it.
PS: We recently fixed a bug caused by inequalities between
VolSurfaceExpirations in vs input, and in SystemContractData. User
inputs to the vs grid would be immediately reverted, as the exps
associated with the inputs were not found in the SystemContractData.
This bug would have been easier to track down with this logging in
place.