Now does averaging when running iterations.
Realligned a bunch of code and continued to remove global variables.
Added CSV output along with printable.
Structures are now provided for print options so adding XML, HTML, ... should now be trivial.
Added --slave option if you are trying to syncronize around one server using the lock-directory option plus NFS.
Added options to allow you to run several concurrency runs in one call. aka --concurrency="1, 8, 64, 128"
Time collection is now done in a finer grain manner, and is linked to the sync call (doesn't count the cost of the blocked processes that are setup to run).
When InnoDB compares varchar field in ucs2 with given key using bin collation,
it calls my_strnncollsp_ucs2_bin() to perform comparison.
Because field length was lesser than length of key field should be padded
with trailing spaces in order to get correct result.
Because my_strnncollsp_ucs2_bin() was calling my_strnncollp_ucs2_bin(), which
doesn't pads field, wrong comparison result was returned. This results in
wrong result set.
my_strnncollsp_ucs2_bin() now compares fields like my_strnncollsp_ucs2 do,
but using binary collation.
- Added more comments.
- Added a RANGE_OPT_PARAM::remove_jump_scans flag that disables construction of index_merge
SEL_TREEs that represent unusable conditions like "key1part1<c1 OR key2part2<c2"
- make prune_partitions() function handle the case where range analysis produces a list of
index_merge trees (it turned out that this is possible, appropriate test case added).
- Other small fixes.
--engine="myisam,archive"
--number-of-rows has been removed for the time being and replaced with
option to allow you to just repeat multiple time all of the inserts you specifed with data. When number-of-rows-returns it will be instead used for averaging rows while growing concurrency.
Returned stats now has accurent row and query counts since it includes a multiple for the rows and queries added.
Parser is now a touch smarter and understands end of file issues if you have a delimiter at the end of the file.
1) Parsing now works
2) Options are safer, aka it doesn't kill the machine and it cleans up after itself
3) Option of --only-print added so that you can see what it does.
4) Tiny cleanup of the auto generate sql. A lot more needs to be done with this, for it to be very valuable. I suspect it doesn't work all that well.
5) Delimeter is now a single character. No good escaping going on.
6) You can now change which schema it is runninng against.
Now I think I can go make use of it!
Though I need to add support for a "only run this many inserts, divide by the number of clients connecting, to really test scaling"
