Potential outgoing message corruption in managed websockets

[sync-by-unito]

When managed websockets are enabled we use array pooling to cut down on constant managed heap allocations for the byte arrays that store the sync messages Core wants us to send. However there is a chance that if a rented array is bigger than the buffer we need to store we will attempt to send the entire array, including stale data it contained previously.

[sync-by-unito]

➤ PM Bot commented:

Jira ticket: RNET-1175