| Date | Version | Comment |
|---|---|---|
| |
||
| 05.09.2010 | 1.3.5 | Support for lfc classes |
| | ||
| 08.08.2010 | 1.3.4 | Small fix for exception throw message in case of parse errors. |
| | ||
| 15.04.2009 | 1.3.3 |
Small fix in generated parser code. the _isReserved flag is reset to false
before starting new parse run ( in case of exception inside string handling ) |
| | ||
| 13.02.2009 | 1.3.2 | In some complex grammars with epsilon productions definitions, the first set calculation may lead to an infinite look. This has been fixed with thsi release. |
| | ||
| 13.02.2009 | 1.3.1 | In some complex grammars with epsilon productions definitions, the first set calculation may lead to an infinite look. This has been fixed with thsi release. |
| | ||
| 28.08.2007 | 1.3.0 |
Completed LALR parse analysis. As expected, analyse performance increases dramatically and generated parse table shrinks to about 15-20% ( based on generated Cego parser ) Still have to check out if generated code is correct and useful |
| | ||
| 27.08.2007 | 1.3.0 |
Started with implementation of LALR parse analysis. |
| | ||
| 12.08.2007 | 1.2.11 |
Added escape token for special regular expressions containing reserved control characters |
| | ||
| 11.08.2007 | 1.2.11 |
With GNU C compiler 3.4.6, a segmentation fault occured with large parse tables To avoid this problem, the option -y has been introduced to generate parser code with dynamic allocated parse tables |
| | ||
| 18.06.2007 | 1.2.10 |
Optimization for generated code done. Now parse table is loaded in a static array which leads to a dramatically improved startup performance Startup time has been observed 7 times faster for large parsers and parallel load ( Cego with 1640x132 x 30 threads : dynamic = 7 min startup, static = 1 minute startup ) |
| | ||
| 15.06.2007 | 1.2.9 |
Some optimization done for analysing performance Changed some call-by-value method parameters to call-by-reference ( Dragon::getJump, FirstHast::getFirstSet and Dragon::getClosure ) This lead to about 7% performance improvement |
| | ||
| 28.05.2007 | 1.2.9 |
Added Exception throw to nextChar and shiftToken methods |
| | ||
| 06.05.2007 | 1.2.8 |
Tests of Java code generation for Tracx parser finished Public release |
| | ||
| 28.04.2007 | 1.2.8 |
Fixes for Java code generation |
| | ||
| 24.01.2007 | 1.2.7 |
Catched missing grammar defintion in Main Clustered parser table load in generated C++ code for better compile performance Indent format corrections in generated code |
| | ||
| 14.01.2007 | 1.2.6 |
Minor corrections in error messages Public release |
| | ||
| 09.12.2006 | 1.2.5 |
Compiler Cleanup for Forte C++ ( Solaris ) |
| | ||
| 04.12.2006 | 1.2.4 |
Make dragon conform to base 1.0.4 |
| | ||
| 06.11.2006 | 1.2.3 |
Check of recursive usage of useless productions added. ( prod -> prod IDENTIFIER was nor detected as useless, if prod was not used in other productions ) |
| | ||
| 16.10.2006 | 1.2.2 | Changed '$'-Token to 'ENDTOKEN'-Token for Compiler compatibility |
| | ||
| 10.10.2006 | 1.2.1 | Some small documentation fixes have been done |
| | ||
| 09.10.2006 | 1.2.0 | New code runs. Java based parser basically tested. |
| | ||
| 07.10.2006 | 1.2.0 | Implementation of Java code generation done. Still untested. |
| | ||
| 03.10.2006 | 1.1.1 | Cleanup in the sample directory to make package ready for public download. Actually there are no open issues for dragon |
| | ||