Currently, as discovered in #327, when using ReadAsync or WriteAsync without first making a call to InitializeAsync and the initialization fails - the exception thrown is of type TaskCancelledException - not LibplctagException.
If I recall correctly, the intention was to throw a LibplctagException for the libplctag.NET async timeouts, and only to throw a TaskCancelledException if the supplied token is cancelled,
One issue with the attached patch is that there is now two separate timeouts - one for the Initialization, and a second one for the Read - so if there is a 10 second timeout, and a 9 second latency on operations - it will successfully complete in 2*9seconds=18seconds -> maybe the library consumer is expected it to timeout because 18 seconds is larger than the configured timeout.
Either way, this is an API-breaking change so will require some sort of version bump.
Currently, as discovered in #327, when using ReadAsync or WriteAsync without first making a call to InitializeAsync and the initialization fails - the exception thrown is of type TaskCancelledException - not LibplctagException.
If I recall correctly, the intention was to throw a LibplctagException for the libplctag.NET async timeouts, and only to throw a TaskCancelledException if the supplied token is cancelled,
One issue with the attached patch is that there is now two separate timeouts - one for the Initialization, and a second one for the Read - so if there is a 10 second timeout, and a 9 second latency on operations - it will successfully complete in 2*9seconds=18seconds -> maybe the library consumer is expected it to timeout because 18 seconds is larger than the configured timeout.
Either way, this is an API-breaking change so will require some sort of version bump.