This code is my translation of the Antlr parser from the Oracle Workbench library written by Mike Lischke.
-
The translation removes almost all the code except for parsing. A few unused methods in MySQLSymbolInfo.cs were kept because they might be useful in some applications. The code was cleaned and reorganized using CodeMaid and Resharper.
-
The code source tree was flatten into one directory to make it simplier to navigate.
-
The Antlr grammars are processed using Antlr4BuildTasks, which tracks changes in the grammar, generates the parser and lexer, and cleans all files that are generated from the tool using the -depend option.
-
The error reporting using the "Code Completion" routines were completely replaced with my own lookahead predictor because Mike's code does not work correctly in some situations. For further information, see my blog entry describing the new algorithm.
-
As far as I could tell, the only tree walkers used were for code completion for error reporting, and not for static semantic analysis. Since the error message handler code was replaced, the tree walkers and the generation of the base classes were removed.
--Ken Domino