What's wrong with this simplest jflex code?

Question

I'm learning jflex, and wrote a simplest jflex code, which makes a single character #:

import com.intellij.lexer.FlexLexer;
import com.intellij.psi.tree.IElementType;

%%

%class PubspecLexer
%implements FlexLexer
%unicode
%type IElementType
%function advance
%debug

Comment = "#"

%%

{Comment} { System.out.println("Found comment!"); return PubTokenTypes.Comment; }
.                                        { return PubTokenTypes.BadCharacter; }

Then I generate a PubspecLexer class, and try it:

public static void main(String[] args) throws IOException {
    PubspecLexer lexer = new PubspecLexer(new StringReader("#!!!!!"));
    for (int i = 0; i < 3; i++) {
        IElementType token = lexer.advance();
        System.out.println(token);
    }
}

But it prints 3 nulls:

null
null
null

Why it neither return Comment nor BadCharacter?

Answer 1

It's not jflex problem, actually, it's because the idea-flex changes original usage.

When use jflex to write intellij-idea plugins, we are using a patched "JFlex.jar" and "idea-flex.skeleton", later defines the zzRefill method as:

private boolean zzRefill() throws java.io.IOException {
    return true;
}

Instead of original:

private boolean zzRefill() throws java.io.IOException {

  // ... ignore some code 

  /* finally: fill the buffer with new input */
  int numRead = zzReader.read(zzBuffer, zzEndRead,
                                          zzBuffer.length-zzEndRead);
  // ... ignore some code

  // numRead < 0
  return true;
}

Notice there is a zzReader in the code, and which holds the string #!!!!! I passed in. But in the idea-flex version, which is never used.

So to work with idea-flex version, I should use it like this:

public class MyLexer extends FlexAdapter {
    public MyLexer() {
        super(new PubspecLexer((Reader) null));
    }
}

Then:

public static void main(String[] args) {
    String input = "#!!!!!";
    MyLexer lexer = new MyLexer();
    lexer.start(input);
    for (int i = 0; i < 3; i++) {
        System.out.println(lexer.getTokenType());
        lexer.advance();
    }
}

Which prints:

match: --#--
action [19] { System.out.println("Found comment!"); return PubTokenTypes.Comment(); }
Found comment!
Pub:Comment
match: --!--
action [20] { return PubTokenTypes.BadCharacter(); }
Pub:BadCharacter
match: --!--
action [20] { return PubTokenTypes.BadCharacter(); }
Pub:BadCharacter