Field_bit::set_default() did not check the bit_len, hence used the undefined
bit_ptr, causing a crash. The patch adds a check that bit_len > 0 before
following the bit_ptr.

The reason of this valgrind's compaint is not a bug but rather a feature of bitwise ops:
for any value of the byte x
x | 1 -> 1,  and x & 0 -> 0.
x, being a null_byte part of record[1] can be left unassigned even after
ha_innobase::index_read_idx because the above and still be correct.
Addding a check memory upon the invocation of the function can detect this fact
long before record[1], old record, is eventually passed to my_write.

Fixed with initialization of record[1]'s null_bytes part in open_table_from_share.

into  mysql.com:/home/bk/MERGE/mysql-5.1-merge

into  mysql.com:/home/bk/MERGE/mysql-5.1-merge

into  mysql.com:/home/bk/MERGE/mysql-5.0-merge

into  romeo.(none):/home/bk/memcheck-mysql-5.1

Removing DBUG_DUMP printouts for valgrind builds since they trigger warnings.
Removing valgrind memory checks completely.
Removing bzero() of record when opening table that was added earlier.

into  mysql.com:/usr/home/bar/mysql-5.1-rpl

into  mysql.com:/usr/home/bar/mysql-5.0.b13926

Moving tests to 4.1 section

Moving tests into their new place into 4.1 tests section

Backporting from 5.0

into  romeo.(none):/home/bk/memcheck-mysql-5.1

Fix to correct behaviour of find_and_fetch_row() for tables that have primary keys stored
in storage engines that support the fast method to fetch rows given a primary key. The
method uses position() to retrieve the key for a given record and rnd_pos() to position
the internal "cursor" at the row. Rnd_pos() returns the found record in table->record[0],
so the record has to be moved to table->record[1] for further processing after calling
find_and_fetch_row().

into  amd64.:D:/src/mysql-5.1_bug23983

- When a shared library argument is supplied, it's checked for an OS
specific directory separator. The expected error is different
depending on the separator used. Create OS specific versions of these
tests.

  - Disconnect from transporter before starting to delete objects

into  mysql.com:/usr/home/bar/mysql-4.1-rpl

into  mysql.com:/usr/home/bar/mysql-5.1.b22646

into  mysql.com:/usr/home/bar/mysql-5.0.b22646

  
  open_table_from_share did not initialize table->record members. that was
  interpreted as the error by valgrind.
  
  Fixed with bzero-ing the members if compilation with -DHAVE_purify.

into  mysql.com:/usr/home/bar/mysql-5.0.b23619

into  mysql.com:/usr/home/bar/mysql-5.1.b23619

into  mysql.com:/usr/home/bar/mysql-5.0.b23619

into  romeo.(none):/home/bk/memcheck-mysql-5.1

Adding class to handle temporary allocate of memory during write of rows.
Eliminating cut'n-paste error between THD::binlog_{write,update,delete}_row().

into  mysql.com:/home/bk/MERGE/mysql-5.1-merge

into  mysql.com:/home/bk/MERGE/mysql-5.0-merge

into  mysql.com:/home/bk/MERGE/mysql-5.1-merge

into  mysql.com:/home/bk/MERGE/mysql-5.0-merge

into  mysql.com:/home/bk/MERGE/mysql-5.1-merge

into  mysql.com:/home/bk/MERGE/mysql-5.0-merge

into  mysql.com:/home/bk/MERGE/mysql-4.1-merge

into  dsl-hkibras-fe30f900-107.dhcp.inet.fi:/home/elkin/MySQL/TEAM/BARE/mysql-5.1-new-rpl

into  dsl-hkibras-fe30f900-107.dhcp.inet.fi:/home/elkin/MySQL/TEAM/BARE/5.0

mending windows test.