Version 1.5

• License modified to GNU LGPL version 3
  The license agreement for this library has been changed to GNU LGPL,
  version 3. This change was made for compatibility reasons and most
  users will remain unaffected.
• Added Visual Basic.NET code generation
  A Visual Basic.NET source code generator has been contributed by
  Adrian Moore. It uses the same run-time library (DLL) as the C# .NET
  version.
  Bug #8199
• Added support for case-insensitive parsing
  A CASESENSITIVE grammar declaration has been added to make it
  possible to parse in case-insensitive mode. By default all parsing is
  case-sensitive. Thanks to Adrian Moore for providing a partial implementation.
  Bug #5060
• Added support for reusing parsers and tokenizers
  A reset() method has been added to both the Parser and the Tokenizer
  classes, allowing to reuse a parser with a new input stream. By reusing a
  single parser and tokenizer for various input files speed and memory
  gains can be made, especially for complex grammars.
  Bug #4500
• Added factory methods for creating various objects
  New factory methods for creating tokens, productions, tokenizers and
  analyzers have been added. These can be overridden in subclasses to
  provide more specific classes if desired.
• Improved the tokenizer processing speed
  The tokenizer speed has been improved by using a custom DFA
  implementation for regular expressions. The tokenization phase now
  runs some 20-30% faster in the general case.
  Bug #3603 &
  Bug #8202
• Changed to use native Java regular expressions
  Since the built-in regular expression library is not used as frequently
  after the implementation of DFA regular expression handling, the native
  Java regular expression classes are now used. This implies improved
  support for advanced regular expression syntax (at a performance
  penalty).
  Bug #3597 &
  Bug #17189
• Changed .NET parser namespace for C# 2.0
  The C# parser namespace has been modified from PerCederberg.Grammatica.Parser
  to PerCederberg.Grammatica.Runtime due to the new C# 2.0 compiler strictness.
  Bug #14302
• Changed .NET parser API for better integration
  The .NET parser API has been modified for almost all classes, introducing
  properties and indices instead of getter and setter methods. The old methods
  have been deprecated, but are still available to avoid breaking existing applications.
  Bug #8693
• Changed built-in regular expression API subtly
  Due to fixes required to be able to reuse tokenizers, the built-in regular
  expression class API had to be changed subtly. In essence, all the methods
  can now throw IOException which breaks a Java compile.
  Bug #4500
• Fixed possible divide by zero in profiling
  When profiling grammars with small input files, the time measured could
  previously be zero (0). This caused a division by zero exception when
  calculating the number of tokens parsed per second.
  Bug #7998
• Fixed parse error on some grammar comments
  Grammar comments ending with **/ previously caused a parse error.
  The comment token regular expression has now been corrected to fix this.
  Bug #12767
• Fixed premature EOF reported in some cases
  The tokenizer previously assumed EOF had been encountered when a full
  look-ahead buffer couldn't be read. Thanks to Jeremy M Stone for finding this.
  Bug #23818
• Fixed token match priority to use the grammar order
  String tokens were previously always checked and considered first, unless a
  regular expression token matched more characters. This has now been
  changed to depend on the grammar file ordering instead.
  Bug #13009