TCollection_ExtendedString, NCollection_String - merge classes for string management

There are several classes for string management with different capabilities but similar usage scenarios:
- TCollection_AsciiString. Main class for strings management in OCCT.
  NULL-terminated.
  Assumes fixed-length 1 byte long characters.
  Provides formatting tools (LeftAdjust/RightAdjust/LowerCase/Capitalize/UpperCase).
  Provides sub-string search tools.
  Does not aware of UTF-8, but can transparently store such strings.
- TCollection_HAsciiString. Wrapper for TCollection_AsciiString for smart pointer creation.
- TCollection_ExtendedString. This class has been originally designed to store wide (UNICODE) strings in OCCT, mostly used in OCAF persistence, but rarely outside.
  NULL-terminated.
  Assumes fixed-length 2 bytes long characters.
  Defines Standard_ExtCharacter as short, should be converted explicitly to wchar_t on Windows.
  Provides formatting tools (LeftAdjust/RightAdjust/LowerCase/Capitalize/UpperCase).
  Provides sub-string search tools.
  Expect string in UTF-16 encoding, but unaware of surrogate pairs.
  Has constructor from UTF-8 string (isMultiByte flag).
- TCollection_HExtendedString. Wrapper for TCollection_ExtendedString for smart pointer creation.
- NCollection_UtfString. Template class for UNICODE strings. Currently used in visualization for text rendering and interaction with FreeType.
  NULL-terminated.
  Assumes multibyte UNICODE characters (UTF8, UTF-16 or UTF32).
  Stores length (in UNICODE characters) and buffer size.
  Provides iterator NCollection_UtfIterator for efficient multibyte string processing (as alternative to accessing character by index).
  Provides methods to convert UNICODE string from/to system locale and between UNICODE variants (UTF-8/UTF-16/UTF-32).
  Does not provide string formatting capabilities.

It is considered to revise these classes to eliminate code duplication.

In addition, methods of class Resource_Unicode should be revised.