A communication packet can also be a binlog event sent from the master to the slave.
To be sent by master dump and accepted by slave io thread both have to have
the value of max_allowed_packet bigger than one that client connection had.

In the patch there is the MAX possible replicatio header size estimation for events
in binlog that embedded user query. Only these events of query_log_event type, i.e
just plain queries, require attention. 


sql/log_event.h:
  MAX value of the header of the replication packet, i.e the value of exceeding
  the query string part.
sql/slave.cc:
  Private value of max allowed packet slave io, sql threads can accept from master.
  The value is increased by the MAX possible size of replication event header.
  
  Note, that my_net_init is redundant for slave io because the thread uses instance of
  NET struct, embedded into MYSQL, as a client to server. We have left old code and even
  increment, redundantly as well, thd->net.max_packet_size just for the sake of
  consistency. TODO: eliminate my_net_init from execution path of slave io.
sql/sql_repl.cc:
  Increasing the private max allowed packet that the dump thread reads from master
  binlog and sends to net. Note, that happened prior this writting to binlog is
  safe from the current artifact because it is done by steps that each chunk can not
  exceed @@global.max_allowed_packet.
mysql-test/r/rpl_packet.result:
  BitKeeper file /home/elkin/MySQL/TEAM/FIXES/4.1/bug19402-max_allowed/mysql-test/r/rpl_packet.result
mysql-test/t/rpl_packet-master.opt:
  BitKeeper file /net/koti/usr_rh9/home/elkin.rh9/MySQL/TEAM/FIXES/4.1/bug19402-max_allowed/mysql-test/t/rpl_packet-master.opt
mysql-test/t/rpl_packet-slave.opt:
  BitKeeper file /net/koti/usr_rh9/home/elkin.rh9/MySQL/TEAM/FIXES/4.1/bug19402-max_allowed/mysql-test/t/rpl_packet-slave.opt
mysql-test/t/rpl_packet.test:
  test examines repliation of a max-sized query when database name is set to have
  NAME_LEN bytes. This makes replication header size the maximum.