re2c 2.0

On Monday, July 20, the release of re2c, a fast lexical analyzer generator, was released.
Major changes:

• Added support for Go language
  (enabled either with the --lang go option for re2c, or as a standalone re2go program).
  The documentation for C and Go is generated from the same text, but with different
  code examples. The code generation subsystem in re2c has been completely redesigned, which
  should make it easier to support new languages ​​in the future.

• Added alternative build system on CMake (thanks ligfx!).
  Attempts to translate re2c to CMake have been made for a long time, but before ligfx no one
  offered a complete solution.
  The old build system on Autotools continues to be supported and used,
  and in the foreseeable future there are no plans to abandon it (partly so as not to create
  problems for distribution developers, partly because the old build system
  more stable and concise than the new one).
  Both systems are equally continuously tested using Travis CI.

• Added the ability to set the interface code in configurations when using
  generic API (generic API). Previously, most APIs had to be specified in the form
  functions or function macros. Now they can be set in the form of arbitrary
  strings with named template parameters of the form @@{name} or just @@ (if
  there is only one parameter and there is no ambiguity). API style is configurable
  re2c:api:style (the functions value specifies the functional style, while the free-form value is arbitrary).

• Improved operation of the -c, --start-conditions option, which allows you to combine several
  interconnected lexers in one re2c block. Now you can use
  ordinary blocks on a par with conditional ones and set several unrelated conditional
  blocks in one file.
  Improved -r, --reuse options (reuse of code from one block
  in other blocks) in combination with the -c, --start-conditions and -f, --storable-state options
  (a stateful lexer that can be interrupted at an arbitrary location
  and continue execution later).

• Fixed a bug in the newly added algorithm for handling the end of input data
  (EOF rule), which in rare cases led to incorrect processing
  overlapping rules.

• Simplified bootstrap process. Previously, the build system tried to dynamically find already
  built re2c that could be used to rebuild itself.
  This led to incorrect dependencies (because the dependency graph turned out
  dynamic, which most build systems don't like).
  Now, in order to rebuild lexers, it is required explicitly
  configure the build system and set the RE2C_FOR_BUILD variable.

Thanks to everyone who contributed to this release!

Source: linux.org.ru