Features

A neat feature of ANTLR is that it will identify missing tokens for us, which can also be visualized in the parse tree GUI.

For example, we introduced two errors in our test.dfm pseudo-code, by removing the colon from the inherited object testFields, and the closing array bracket in OtherSortOptions:

...

inherited testFields TestFields
    Indexes = <>
    SortOptions = []
    OtherSortOptions = [test1, test2, test3
end

...

Let's see what happens if we output a parse tree using the -gui command:

grun DelphiDFM file -gui test.dfm

We even get the following output in our terminal:

line 42:27 token recognition error at: '[test1, test2, test3\n    end\n    inherited testMultipleInheritance: TestMultipleInheritance\n        inherited inhOne: InhOne\n            inherited inhTwo: InhTwo\n                inherited inhThree: InhThree\n                    object deepObj: DeepObj\n                        Height = 10\n                        Width = 20\n                        Top = 30\n                    end\n                end\n            end\n        end\n    end\nend'
line 39:25 missing ':' at 'TestFields'
line 57:3 mismatched input '<EOF>' expecting {'<', ARRAY, ID, FLOAT, INT, STRING}

This becomes highly useful since we have multpiple ways of identifying issues. Note that errors also get printed when using commands such as grun DelphiDFM file -tokens test.dfm where we output all parsed tokens from a file, just like we did previously.

Important

If the grammar file is modified (in our case DelphiDFM.g4), it's always necessary to regenerate and recompile from scratch to avoid stale .class files:

rm *.java *.class *.tokens *.interp

antlr4 DelphiDFM.g4

javac DelphiDFM*.java

Conclusion

The aim of these first two chapters, was to provide the necessary context for building a language grammar. And additionally, to offer a quick way of parsing desired code, by using the created ANTLR-specific grammar.

We learned the general concepts needed to write a language grammar for parsing with ANTLRv4. Then we applied the concepts to write our own DFM grammar, which we used to successfully parse the test.dfm pseudo-code. We also learned how to check all tokens of a file, and decomposed them to understand how they are built. Then we checked and represented all identified tokens in a visual parse tree. Finally, we have taken a look at some built-in error reporting features of ANTLRv4.