The problem was an ab-use of last_rkey_length.
Formerly we saved the packed key length (of the search key)
in this element. But in certain cases it got replaced by
the (packed) result key length.
Now we use a new element of MI_INFO to save the packed key
length of the search key.
This is a replacement for the original patch given by Ingo. This one comes
from Monty.
The problem is that merge files now use unix style pathnames on all
platforms. The merge file open code was not properly converting those
pathnames back to non-unix when necessary.
Changed the creation of the .MRG file so that only the table name
is written when the MyISAM table is in the same database as the
MERGE table, a relative path is used in other cases in mysqld,
and possibly an absolute path is used in an embedded server.
No test case is added as the external behaviour is unchanged.
Only the file names within the .MRG file are changed.
Use 'mysqltest' as test database instead of test_$1 or test1,test2 to not accidently delete an important database
Safety fix for mailformed MERGE files
New records_in_range() interface (similar to read_range())
Macros for faster bitmap handling
Simplify read_range() code (#WL1786)
New general key_cmp() function to compare keys
New (simpler) internal timestamp handling.
More debuging to heap tables.
Small cleanups to multi-table-delete
false -> 0 and true -> 1 (We should use TRUE and FALSE)
Some simple optimzations, more comments and indentation changes.
Add ` around database in 'use database' in binary log.
Moved max_error_count and max_warning_count to variables struct.
Removed SHOW_WARNS_COUNT and SHOW_ERRORS_COUNT calls.
Changed string functions to use character set of first string argument as default return characterset
(Each string function can change the above assumption if needed)
Made a some new buffers thread specific and changeable.
Resize of key_buffer.
AUTO_COMMIT -> AUTOCOMMIT
Fixed mutex bug in DROP DATABASE
Fixed bug when using auto_increment as second part of a key where first part could include NULL.
Split handler->extra() to extra() and extra_opt() to be able to support thread specific buffers.
Don't write message to error log when slave reconnects becasue of timeout.
Fixed possible update problem when using DELETE/UPDATE on small tables
(In some cases we used index even if table scanning would be better)
A lot of minior code cleanups
