Especially in the compiled model, we currently generate lambdas which aren't fully qualified:
activityTagPacketFrameRow.TypeMapping = SqlServerLongTypeMapping.Default.Clone(
comparer: new ValueComparer<long>(
(long v1, long v2) => v1 == v2,
(long v) => v.GetHashCode(),
(long v) => v),
keyComparer: new ValueComparer<long>(
(long v1, long v2) => v1 == v2,
(long v) => v.GetHashCode(),
(long v) => v),
providerValueComparer: new ValueComparer<long>(
(long v1, long v2) => v1 == v2,
(long v) => v.GetHashCode(),
(long v) => v));
For better compilation, it is highly recommended to specify the return value of the lambdas, to avoid the need for inference:
// current
(long v1, long v2) => v1 == v2;
// preferred
bool (long v1, long v2) => v1 == v2;
Raised by @jaredpar
Especially in the compiled model, we currently generate lambdas which aren't fully qualified:
For better compilation, it is highly recommended to specify the return value of the lambdas, to avoid the need for inference:
Raised by @jaredpar