Fix for bug#9896, each fragment has a twin fragment (in 4.1,5.0), pages in this were not counted, giving incorrect memory consumption

  Bug#10504: additional test

into  mysql.com:/usr/home/bar/mysql-5.0.b10504

into lmy002.wdf.sap.corp:/home/georg/work/mysql/bugs/mysql-5.0-master

into  snake-hub.snake.net:/src/extern/MySQL/bk/mysql-5.0

into  snake-hub.snake.net:/src/extern/MySQL/bk/merge-4.1

  Delete: man/which.2

Brian's "he is too lazy to commit this so Eric is going to have to do it" environmental variable hack for prefix. This allows you to just set a prefix in your profile which will be used in all of your builds. AKA you can be even lazier(!) about calling BUILD scripts. Don't blame Eric, Brian borrowed his keyboard.

into  127.(none):/home/osku/mysql/5.0/clean

into moonbone.local:/work/12291-bug-5.0-mysql

into mysql.com:/M50/mysql-5.0

into  sanja.is.com.ua:/home/bell/mysql/bk/work-owner2-5.0

  Character set does not support traditional mode
ctype_utf8.result, ctype_utf8.test:
  adding test case.
password.c, mysql_com.h
  Changeing octet2hex availability from static to public.
item_strfunc.cc:
  Result string is now checked to be well-formed.
  Warning/error is generated, depending on sql_mode.

Optimizer did choose "Range checked for each record" for one of the tables.
For first few loops over that table it choose sequential access, on later
stage it choose to use index. Because table was previously initialized for 
sequential access, it skips intitialization for index access, and when
server tries to retrieve data error occurs.

QUICK_RANGE_SELECT::init() changes so if file already initialized for
sequential access, it calls ha_rnd_end() and initializes file for index
access.

into moonbone.local:/work/12887-bug-5.0-mysql

into moonbone.local:/work/12887-bug-5.0-mysql

into serg.mylan:/usr/home/serg/Abk/mysql-5.0

into moonbone.local:/work/12887-bug-5.0-mysql

into  mysql.com:/home/jimw/my/mysql-4.1-clean

into  selena.:H:/MYSQL/src/fin-mysql-5.0

into moonbone.local:/work/12887-bug-4.1-mysql

For queries with GROUP BY and without hidden GROUP BY fields DISTINCT is
optimized away becuase such queries produce result set without duplicates.
But ROLLUP can add rows which may be same to some rows and this fact was
ignored.

Added check so if ROLLUP is present DISTINCT can't be optimized away.

into moonbone.local:/work/11416-bug-5.0-mysql

into  mysql.com:/home/jimw/my/mysql-4.1-clean

into  selena.:H:/MYSQL/src/fin-mysql-5.0

into  mysql.com:/home/my/mysql-4.1

Ignored some generated files

into  sanja.is.com.ua:/home/bell/mysql/bk/work-owner2-5.0

into  mysql.com:/export/space/pekka/ndb/version/my50