Use non_sso_data() and sso_data() methods instead of direct access to
make sure data is accessed only when it is initialized.

In insert, it was wrongly assumed that if size < sso_capacity it means
that sso is used. It is possible that non_sso type is used and
size is < sso_capacity so explicit check is needed.

concurrent_hash_map: reimplement my_mask

Fix release build errors

my_mask is a variable which is reinitialized on each process
restart. Until now, v<> was used for this purpose. Because v<> is not
productized yet, reinitialize my_mask in runtime_initialize() method. Now,
runtime_initialize() MUST be called after proccess restart.

In future, when v<> is productized, we may switch back to using it.

and call it in every test before using hashmap

concurrent_hash_map: implement insert using tx

concurrent_hash_map: add test for concurrent inserts with the same key

Hash map minor refactoring

travis: limit the coverage build

Implement inserts which take rvalue ref uisng internal_insert.
This is more consistent with implementation of other inserts.

This approach simplifies the code and slightly improves performance.

Vector insert fix

and add tests for interators.

Fix a heterogeneous typo

Multithreaded concurrent hashmap API tests

This is the only possible use case for this function. Calling construct
with index other than size() would result in calling constructors
on already initialized memory.

insert_gap() method worked incorrectly. It used std::move_backward to move
elements in a vector, ignoring the fact that memory after size() was uninitialized.

This resulted in operator= being called on uninitialized elements instead of
copy/move constructor.

Fix this issue by reimplementing insert_gap as internal_insert. Use generic
implementation for iterators to implement all cases.

and change name of construct_range_copy to construct.

conctruct_range can be implemented using construct_range_copy by passing
std::make_move_iterator(iterator) as parameters.

Decrease number of concurrent threads in concurrent hashmap tests if run under
drd. It fix timeout issue during travis builds.
Varialbe "On_drd" is exposed by unittest library.

Add tests for all kinds of insert, erase and find, as defined in API.
 * Abstraction of class ConcurrentHashmapTestPrimitives is added to achieve
tests code as self documented scenarios.
 * Prints are not visible during ctest execution, but are helpful when
test binary is run manually.

string: modifiers

Make valgrind build separate travis step

It's done for build time optimization.

* Put builds as separate functions in run-build.sh
* Explicitly list, which builds should be run according to build type
* Add valgrind build type
* Add valgrind build as separate travis step

Add concurrent_hash_map_pmreorder_break_insert test