Draft: new: Upload data in chunks,
adjusting them to size and time the previous request took, similar to
how data is downloaded, but having a generator class to store size and
time and calculating the next size to avoid complexity in scanner
.
Since the connection request sends keep_alive
(what isn't new, but
now has been checked with a custom server that process the uploaded
data), there isn't extra time opening a new connection.
Closes #40220 (closed)