Fix bug when using option groups

into sinisa.nasamreza.org:/mnt/hdc/Sinisa/mysql-4.0

into mysql.sashanet.com:/reiser-data/mysql-4.0

Fix for --enable-local-infile
Portability fix for bison

into sinisa.nasamreza.org:/mnt/hdc/Sinisa/mysql-4.0

into co3064164-a.bitbike.com:/home/arjen/mysql-4.0

of one of our users which replicates from a live server. So I now have a
lot of confidence in it and will push. It passes all tests, so I if I broke
somebody's code in 3.23 merge this is your fault - you should have had a good test
case :-)

overrun sentry in my_vsnprintf() test
will not be pushed yet

Fix problem with HAVING and MAX() IS NOT NULL

Let's hope this did not break anything. Will not push this one until Monty
gets back

into mysql.sashanet.com:/reiser-data/mysql-4.0

fixed replication bug - flush_master_info() needs to be called every time we read an 
    event from master

this is will not be pushed yet

into serg.mysql.com:/usr/home/serg/Abk/mysql-4.0

into sinisa.nasamreza.org:/mnt/hdc/Sinisa/mysql-4.0

  Update
costarica.txt, costarica.gif, costarica.eps:
  new file

into sinisa.nasamreza.org:/mnt/hdc/Sinisa/mysql-4.0

accomodate EXPLAIN properly for all possible variant of UNION's
derived tables etc ...

This also fixes all reported (and unreported) bugs when EXPLAIN is
run on UNION's.

select_describe() now has a very dirty hack to avoid some problems
with transactional tables on table unlocking.

Code speedup is possible. For the moment all that was important was
that all tests are passed.

reference manual for O'Reilly.

  Script that takes the relative values in the 
  colspec tags and converts them to absolute values.

changeset - afaik, I can edit changesets with bk. So, rather than wait, I will
commit this now, finish up the other code, try to alter the changeset and then
push all the changes up.

into serg.mysql.com:/usr/home/serg/Abk/mysql-4.0