- MariaDB has different wording for a few error messages
- MySQL changed Extra='' to Extra=NULL for EXPLAIN outputs, MariaDB didnt
- The testsuite in storage/rocksdb/mysql-test needs paths to include
  files adjusted
- In SHOW COLUMNS output, Extra column is "NULL" in MariaDB vs '' in MySQL

Running MTR tests produces an error like this after the test run:

... 140562896741120 [ERROR] Invalid (old?) table or database name '.rocksdb'

There seems to be no way to have the MariaRocks to prevent these, so
the current way to fix them is to have --ignore-db-dirs in my.cnf

SHOW STATUS LIKE 'pattern' returned Rocksdb_XXX status variables
that had SHOW_FUNC type but didn't match the pattern (for example
Rocksdb_block_cache_add).

Among other things, this caused MTR to assume that each testcase
has damaged the execution environment.

The issue was a unitialized variable and then a typo in the condition
that checks if variable name matches the pattern.

In MariaDB, Field::make_sort_key stores NULL-indicator byte for
the field.
In MySQL, it doesn't, so MyRocks stores the NULL-indicator itself.

Switch to using Field::sort_string, which is the same as
Field::make_sort_key in MySQL.

rdb_cf_options.cc must be compiled with RTTI

- Add include/have_rocksdb.inc (TODO: is there any way to have this
  file somewhere under storage/rocksdb/mysql-test ?)
- Make rocksdb.test require have_partition.inc because it uses
  partitioned tables

- Off-by-one error in handler->index_flags() parameter caused
it to be called for the keypart that's after the last key part.

- RocksDB uses RTTI (e.g. it uses dynamic_cast)
- mysql-5.6 has RTTI enabled for the whole server, so there's no
  conflict.
- MariaDB doesn't have RTTI enabled
So the solution is to enable RTTI for
- RocksDB files
- Parts of MyRocks that interface with RocksDB features that require RTTI.

(If one just enables RTTI for all of MyRocks, they will get a
"undefined reference to `typeinfo for handler'" link error)

Also, ER_LOCK_WAIT_TIMEOUT error doesn't have a string parameter
in MariaDB.

Also provide handler::is_using_full_key

The original parameter comes from
commit a869c56d361bb44f46c0efeb11a8f03561676247
write/sync redo log before flushing binlog cache to file

we've already disabled some of that patch in
8cde1d44

cset
c2b8ced16e255ffdbcc1e638ac8dd4f398eb1a93
Implement START TRANSACTION WITH CONSISTENT INNODB SNAPSHOT

Should we start supporting it?

It is from "write/sync redo log before flushing binlog cache to file"
feature that MariaDB doesn't have (or need?)

adjust a few function definitions accordingly

Then Regex_list_handler is not needed and we don't have to backport
it right now.

And the code in ha_rocksdb.cc makes assumptions about hidden pk length
being 8, anyway.

(Squash with 4be8cae5)

"Add new function my_hash_const_element(), the const equivalent of
my_hash_element()" - comes from facebook/mysql-5.6,
7c869d34b9fa2262b941efd6363a260b7c37948f

Probably this is not the right way to do it, see the comment

(why are table reads/writes/updates collected in ha_statistics?
this structure was intended for table statistics that the engine
reports.. and now it is also used for engine-agnostic statistics
like rows_read, rows_requested, etc?)

Flashcache is available only in WebScaleSQL (not in MySQL or MariaDB).

They have been replaced with increment_statistics() calls made
by handler::ha_XXX() methods.

Issue #108: Index-only scans do not work for partitioned tables and extended keys
c1f1f0b64eaecd132fc6ec21663faf0510afacaf