Skip to content

Latest commit

 

History

History
146 lines (116 loc) · 6.01 KB

CHANGELOG.md

File metadata and controls

146 lines (116 loc) · 6.01 KB
Raw

Changelog

This file is used to document any relevant changes done to UniCC.

[v1.8]

Released on Oct 28, 2023

  • Improved program module generator to support - and -settings
  • Updated Test suite
  • Target python updated to spaces instead of tabs and further PEP-8 improvements
  • Target javascript updated to more recent JS/ECMAScript6 standards
  • Target json fixed and updated

v1.7

Released on Jan 20, 2021

  • Improved error reporting on own parser
  • New #extends-directive added to extend grammars from others
  • Licensing terms changed from 3-clause BSD-license into MIT license
  • Minor function API changes regarding libphorward v1.0.0 compatibility
  • Distributed libphorward sources integrated into the lib/-directory
  • Auto-generated prototypes using pproto: make -f Makefile.gnu proto.h

v1.6

Released on: May 27, 2019

  • v1.6.2: Provide ERRSTYLE_IMPORTANT-flag to show up important warnings
  • v1.6.1: Removed memory leaking bug
  • Renamed top-level directive #!mode options to scannerless (default) and scanner.
  • !language-directive allows for identifier as language parameter.
  • Removed deprecated directives #parser, #description, #copyright and #version.
  • Removed old-style !model top-level directive.
  • Several internal code revisions.
  • Compiled and linked against libphorward 0.24.

v1.5

Released on: Nov 20, 2018

  • New targets for JavaScript and JSON.
  • Targets C and C++ improved to avoid memory leaking and handle malloc/realloc calls more securely.
  • All targets now supporting the UNICC_SUCCESS and UNICC_ERROR flags that can be set as current act (action) to stop parsing.
  • Started a grammar and templates test suite for better continous integration.
  • Compiled and linked against libphorward 0.23.
  • Removed outdated syntax-tree feature from C target.

v1.4

Released on: April 17, 2018

  • Compiled and linked against libphorward 0.22
  • C++ target forked from the C target startet
  • action_prefix value can be defined in target language templates to redefine the @-prefix for action code variables by any other desired prefix. This change has been done to make target language processing easier for languages where the @-sign is part of the target language syntax.

v1.3

Released on: December 21, 2017

  • v1.3.2: Python target 0.4
    • Raising parse error exceptions
    • Bugfix to avoid working with previous stack values
  • v1.3.1: Bugfix in the C target working with uninitialized values
  • Code generator does not emit action code when the action is empty.
  • New command-line parameter "-l" or "--language" to specify a target language via command-line. This is useful when a grammar only contains AST definitions, which allows to render and run it in any target.
  • Abstract syntax tree generation support:
    • Drafting an abstract syntax tree definition syntax for the grammar language, by using the operators "=" and ":=". See examples/expr.ast.par for an example working under both C and Python without any modifications.
    • C and Python parser targets extended to construct abstract syntax tree data structures dynamically.
    • These are first changes for an upcoming UniCC v2, where the targets shall be re-used with only few improvements.
  • Fixed several bugs in the Python target on insensitive mode.
  • Additionally allow ":" in case of "->" in grammar definitions.
  • Internal code revisions started, but interrupted for now.
  • Renamed folder "templates/" into "targets/".

v1.2

Released on: November 7, 2017

  • Started an (yet incomplete) new standard template to provide parser generation support for the Python programming language (targets/python.tlt).
  • Imported the source code of the C standard template into the UniCC repository for further developments, the previous repository gets closed. (targets/c.tlt)
  • Fixed bug in the lexical analyzer generation by linking against libphorward 0.20 that caused a massive overhead of deterministic finite automation states in scannerless mode due an arbitrary character set order. This bug did not produce invalid parsers, but horribly huge tables.
  • Imported the source code of min_lalr1 into the unicc repository, because it is the only place where it is used.
  • Several internal revisions
    • Changed internal names for files, functions and variables.
    • Replaced old-style function headers by more convenient ones.
    • Removed unused code.
  • Renewed manpage
  • New README.md and updated CHANGELOG.md

v1.1

Released on: September 9, 2016

  • Internal fixes to compiles against libphorward 0.18
  • Licensing terms changed from Artistic License 2.0 to 3-clause BSD-license

v1.0

Released on: June 29, 2012

  • Moved build toolchain to GNU autotools, cross-compiles well on Linux
  • Turned the entire regular expression handling to the new pregex_ptn structures of the Phorward Foundation Library.
  • Output of original regular expressions in the standard regex notation into the -tag of the parser description file.
  • New command-line option '-t' to print the output files to stdout instead of writing it into files.
  • Command-line option '-v' automatically switches '-s' on.
  • Added document type definition file 'unicc.dtd' and changed XML output.
  • The "#case insensitive strings" option was not correctly recognized due the revision of the regular pattern construction mechanism.