LRGen 8.3

LRGen 8.3 Standard. (1)  SLR(1) parser generator. (2)  SLR(1) lexer generator. (3)  Integrated symbol-table builder. (4)  Abstract syntax tree construction. (5)  Intermediate code generation. (6)  Compiler front end source code in C++. (7)  Grammars:  Calculator. Free download ...

LRGen 8.3 Professional. (1)  LALR/LR(1) parser generator. (2)  LALR/LR(1) lexer generator. (3)  Integrated symbol-table builder. (4)  Abstract syntax tree construction. (5)  Intermediate code generation. (6)  Compiler front end source code in C++. (7)  Better parser-table compression. (8)  Fast DFA lexer generation. (9)  Grammars:  Ada, C, C++, CICS, COBOL, dBASE, FORTRAN, Modula2, Pascal, PL/I, PL/M, SQL, Visual Basic, XPL, Zeus. License ... Contact us.

World's Fastest Lexers & Parsers *
LRGen creates the world's fastes lexers and parsers in C/C++.  We did the research and created the fastest algorithms, portable to many programming languages, but we think C/C++ gives you the best performance.  Our innovative 4-matrix-parser tables and optimization techniques also contribute to the high-performance of the language processors.  HTML, XML, C, C++, Java, SQL, COBOL and other languages can be processed with our generalized compiler front end.  [* Based on our experience]

LR Finite-State Machines
LRGen builds LALR/LR(1) parsers and DFA lexers as finite-state machines, a concept invented by Donald Knuth [1965] and made practical by Frank DeRemer [1969].  LRGen 8.3 offers a new option, fast DFA lexers.  Using the DFA option, a language processor built with LRGen, can process 10 MB of text (XML, HTML, C++, Java, etc.) in 0.609 seconds on a 3 GHz computer.  Here is a chart showing the different parts of the language processor and their execution times when reading a 72,000-line file.

Process	With DFA Lexer		With LR Lexer
Lexer	0.078 sec	12%	0.437 sec	45%
Parser	0.125 sec	21%	0.125 sec	13%
Symbol Table	0.124 sec	21%	0.124 sec	13%
AST Creation	0.126 sec	21%	0.126 sec	13%
Input/Output	0.155 sec	25%	0.155 sec	16%
TOTAL TIME	0.609 sec	100%	0.968 sec	100%
SPEED	118,000 lines/sec		74,000 lines/sec

Very Small Parsers
LRGen generates LR(1) parser tables that are very small in size.  The DFA lexer tables are less than 3,000 bytes for most languages.  Below is a table showing parser-table sizes (in bytes) for some languages.

Language	Grammar Rules	Parser Table Size
Ada	414	13,789
C	222	6,177
C++	307	10,967
COBOL	1,045	29,909
DB2	1,868	66,374
Lotus Script	767	28,400
Oracle SQL	1,321	44,116
PL/I	1,275	31,015
Visual Basic	478	18,113
XPL	109	2,465

World's Most Powerful BNF Notation
LRGen reads TBNF grammar notation, the most powerful language for defining translators, interpreters and compiler front ends.  We invented TBNF so that you don't have to write tons of code to get your project finished.  Abstract-syntax trees (not parse trees) are created automatically.  Intermediate code can be generated automatically.  Context sensitive language constructs can be handled automatically.

LBNF Grammar Notation
LRGen reads LBNF grammar notation, for generating lexers.  LBNF is Extended BNF grammar notation which is more readable than regular expressions and more powerful.  With LBNF it is easy to specify difficult input sequences such as nested comments. 

Generate Source Code In Any Language
LRGen grammars are language independent because you don't put source code in the grammar.  Therefore, LRGen can use a skeleton parser input file which tells it how to format the parser-table source code that it generates.  By changing the parser skeleton and lexer skeleton, you can change the generated source code to another programming language.  Users have been able to generate code in Pascal, C# (sharp), IBM Assembly Language and others.

Works With Microsoft Visual Studio C/C++
LRGen is a command-line program without a fancy IDE, however, it comes with project files for working with Visual Studio 2005.  We would like to have a fancy IDE for LRGen, but we're not there yet.  We have put our efforts into other things which the competition does not yet offer, such as performance and automating the whole development process.