You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The JSON interface is very verbose, and we're very concerned about performance, especially with a larger test run. The amount of information passed for a single test case is huge, and multiply that by hundreds, thousands or tens of thousands, and we'll see a lot of memory traffic, in-process with the IDE. It seems that the format could be a lot shorter, without each property having a full type descriptor passed with it. This would also make deserialization faster.
We don't see a need for the properties to be self-describing. If we need to make use of the data, we will already know the property ID, display name and value type. Describing it seems unnecessary, and leads to the verbosity.
Should Attributes include values such as "Hidden"? This feels like a UI concern, but in the reporting protocol.
We don't understand the use of the .NET type name for the ValueType field - what is the purpose of using "System.Collections.Generic.KeyValuePair`2[[System.String],[System.String]][]"? The type name doesn't help us - we can't create a new instance of the class to help deserialise, so it seems to be used as a key or identifier, and a very verbose one at that.
The text was updated successfully, but these errors were encountered:
Filing on behalf of @citizenmatt.
The text was updated successfully, but these errors were encountered: