use opt_net_buffer_length instead of net_buffer_length

when UNIV_DEBUG is defined (introduced in ChangeSet@1.1850)

into mysql.com:/tmp/skr99/mysql-4.0

into deer.(none):/home/hf/work/mysql-4.0

configurations added to build the libraries with USE_TLS flag to be
used for libmysqld

into mysql.com:/tmp/skr99/mysql-4.0

Update of VC++ project files.

into mysql.com:/home/mysql_src/mysql-4.0

so we set SQL_BIG_SELECTS to 1.

Use %windir% instead of c:\winnt\ (Bug #3786)
Fixed wrong foreign key test in crash-me (Bug #3740)

into mysql.com:/dbdata/psergey/mysql-4.0-root

into mysql.com:/home/my/mysql-4.0

Ensured that all projects compile
Removed compiler warnings
Better setting of server_version variable.
Fix that make_win_src_distribution creates the privilege tables.

into mysql.com:/home/mysql_src/mysql-4.0

(fix by our Harrison Fisk): when one does SET GLOBAL SERVER_ID=x, we must set
server_id_supplied to 1.

work as expected": precision-losing conversion removed from 
sys_var_thd_ulonglong.

into gw.mysql.r18.ru:/usr/home/ram/work/4.0.b3756

  Fix a memory corruption bug: in 32-bit computers, every 4 billionth malloc outside innodb_additional_mem_pool_size was mistreated when freeing the memory; this could corrupt the InnoDB additional mem pool and could have caused crashes anywhere, also inside MySQL, or even database corruption! the bug exists also in 3.23 and 4.1; workaround: configure innodb_additional_mem_pool_size big enough

  Allocate the table name buffer from the mem heap of a foreign key object rather than pointing to the name buffer in the table object; even though this apparently caused no bugs in RENAME, DROP, ALTER TABLE, or elsewhere, this convention was very prone to memory allocation bugs

into mysql.com:/home/mysql_src/mysql-4.0

by adding sleep-and-retries (max 4 times) if MASTER_POS_WAIT() returns NULL
in sync_with_master and sync_slave_with_master.
The problem showed up only today, in MySQL 5.0 in rpl_server_id2.test,
but may affect 4.x as well, so fixing it here. Note that I am also fixing
5.0 too, with the same exact patch, because I don't want to leave 5.0 broken
until the next 4.0->4.1->5.0 merge.