It seems the new TCP/IP inventory fetching system in 1.19.1 has an issue in times of high ping and dies with the error "Inventory Fetch From Server Time Out (Retry/Cancel)" shortly after login.
Selecting retry does not cause the inventory load to restart as it should, even when ping times return to normal, pressing cancel stops the inventory load and causes a very high amount of perceived loss.
This TCP/IP based inventory load procedure seems to work until 400ms ping time, after that it seems to fail. Given that people in the Soutthern Hemisphere see 200-600ms ping as a normal occurrence this 400ms barrier is a real concern.
Can the time-out be increased to a more realistic number
Can the retry logic be fixed so that a retry foes occur.