Generating Go Parser from C# Grammar: Wrong Types & Missing Base Structs

[RotkivDS]

Hi!
I just have a question about parser generation.
When I try to generate a C# parser with a Go target, the generated code is missing the base class.

java -jar antlr-4.13.2-complete.jar -Dlanguage=Go CSharpLexer.g4 CSharpParser.g4 CSharpPreprocessorParser.g4

The parser/lexer also uses the wrong types, such as Boolean instead of bool, and uses this instead of p or l in the lexer.

I've fixed this by editing the G4 files and adding new base structs written in Go to the G4 files. But I'm wondering if I did something wrong?

Here are the modified .g4 files:
CSharp_modifed.zip

[kaby76]

The csharp grammar does not have a Go port, which you can tell from in the targets list of the desc.xml. People sometimes write a non-functioning port, or break ports while extending other ports. This is why we have the desc.xml file. It's the only way to declare whether the port actually works or not.

The base classes for the grammar (CSharpLexerBase and CSharpParserBase) are written by hand, not generated. With @members, you don't declare the base class in the .g4 in the options { superClass=...; } statement. They are mutually exclusive.

But, in general, we don't use @members in the .g4's in this repo because it destroys any possibility of using the grammar in other ports. If every port required a different .g4, the result would be a maintainence nightmare in trying to keep everything in sync. The preferred method is to use "target agnostic format".

Antlr never developed an action grammar that use a higher-level language, e.g.,XQuery.