* write case
* NOTE: There is an odd corner case here. It may occur that
* in a single iteration of the state machine, the read key is yielded
- * prior to the write key for the same level. This is undesireable
+ * prior to the write key for the same level. This is undesirable
* for quic, but it is ok, as the general implementation of every 3rd
- * party quic stack while prefering write keys before read, allows
+ * party quic stack while preferring write keys before read, allows
* for read before write if both keys are yielded in the same call
* to SSL_do_handshake, as the tls adaptation code for that quic stack
* can then cache keys until both are available, so we allow read before