how to recoganize a url pattern ?

[kurt-steiner]

I have tried this

grammar ExprParser;

url: Scheme '://' Host (':' Digit+)? ('/' Path?)? Query? Frag? Newline?;


Scheme
    : String
    ;

Host
    : '/'? Hostname
    ;

Hostname
    : String
    | '[' V6Host ']'
    ;

V6Host
    : '::'? (String | Digit+) ((':' | '::') (String | Digit+))*
    ;

String
    : NormalChars
    | Digit+
    ;

Path
    : String ('/' String)* '/'?
    ;

Query
    : '?' Search
    ;

Search
    : SearchParameter ('&' SearchParameter)*
    ;

Frag
    : '#' (String | Digit+)
    ;

SearchParameter
    : String ('=' (String | Digit+ | Hex))?
    ;

Hex
    : ('%' [a-fA-F0-9] [a-fA-F0-9])+
    ;


NormalChars
    : ([a-zA-Z~0-9] | Hex) ([a-zA-Z0-9.+-] | Hex)*
    ;
    
Digit      : [0-9];
Space      : ' ';
Newline    : '\r'? '\n';

and the start rule is url, with the input http://www.google.com, but I got this

what ??

[kaby76]

"token recognition error" says that it can't tokenize the input. You define the grammar almost entirely as a lexer grammar, but didn't take into account how Antlr lexers work:

• Always match the lexer rule to the longest string.
• If two or more lexer rules match the same string, the first lexer rule "wins."

So, either you need to write this as a collection of parser rules (see url grammar), or define one non-fragment lexer rule called "URL" and have all other lexer rules "fragment". "fragment" means "don't produce a token for this rule ever."

[kurt-steiner]

eh, this is my first time writing this grammar file.
and I try copying all this url.g4 to test this url: http://localhost/hello, but I found that

something wrong with my code ??
I copy all the content of url.g4 into the left textarea

[kaby76]

Look at the top-left where it says "Lexer Parser Sample". Click on "Lexer" to switch the left panel to the lexer grammar. Select all the text and delete the old lexer grammar as it has nothing to do with your "url.g4" combined grammar. Switch back to the "Parser" grammar and make sure it's OK. Then, click on the "run".

[kurt-steiner]

thanks, it works, finally