With the suppression of the created boolean, the mkdir()
operation was suppressed.
+ Improved exception management
+ and creation conditions

 - suppressed conn variable= use service.getConnection() instead
 - defined prepared queries
 - precompiled on the fly if used only once
 - incremented data base version to 490
 - systematic use of try-with-resources for ResultSet

- alid is now defined and set in Metadata 
- is it initialised to null
- there is a guard for the existence of # in the cell ids

This reduces the number of warning raised by findbugs

outputfilename == null instead of !=

- Suppressed useless catches
- Used Double.compare
- Suppressed unused Clonable

- relies on java.nio.charset.StandardCharsets for character encodings
- any mention of UTF relies on it
- Fixed default encoding assumptions (set to UTF-8 encoding everywhere)

- implemented properly clone() by calling super.clone()
- implemented old (deep) clone by copy constructors
- implemented copy construtors for toObjectAlignment and toEDOALAlignment
- suppressed all clone() from EDOAL
- adapted all the tests
- Deprecated toObjectAlignment() and toEDOALAlignment()

- Better exception (less generic Exception)
- more static marked final + protected static initialisations
- Extensions implement cloneable
- Prepared to use more SQL ProtectedStatement
- Fixed a couple of == comparison for strings

This is this that solves the problem mentioned in r2138

- updated path
- updated onepom
- updated clitest message with bash
- updated version of logback

- Upgraded to Thrift 0.10.0 (lib)
- Upgraded to HTTP Client 4.5.2 and HTTP Core 4.4.4 (lib)
- Upgraded to Jena 3.3.0 (lib)
- Updated relevant tests