fixed zero index problem where _u8ResponseBufferLength is set.

[intelligentandgoodlooking]

Description

when a 1 word (2 byte) response is received _u8ResponseBufferLength is erroneously set to 0 instead of 1
similarly, a 16 word (32 byte) response is received, _u8ResponseBufferLength is set to 15 instead of 16.
This causes available() to return 1 less than it should.

Issues Resolved

It seems like not many people use the available() function.

Check List

General

• Code follows coding style defined in STYLE.md
• Doxygen comments are included inline with code
• No unnecessary whitespace; check with git diff --check before committing.

The following have been modified to reflect new features, if warranted

• README.md
• keywords.txt (use tabs as whitespace separators)
• library.properties
• examples/ - update or create new ones, as warranted

The following have NOT been modified

• doc/ - will be updated upon versioned release
• .ruby-gemset
• .ruby-version
• CHANGELOG.md - will be updated upon versioned release (HISTORY.md is deprecated)
• Gemfile
• LICENSE
• VERSION - will be updated upon versioned release

[PBudmark]

Yes,
the _u8ResponseBufferLength needs the stepup, but to avoid buffer overrun,
the assignment should be moved inside brackets at 'if (i < ku8MaxBufferSize)'
or else _u8ResponseBufferLength is increased even if no register is stored.

In the case with odd number of bytes, 'if (u8ModbusADU[2] % 2)',
_u8ResponseBufferLength = i + 1; should be done inside the brackets,
as i is already stepped from the for-statement.