ANTLR4特定位置的特定字符

Question

I'm using ANTLR4 for a project but I've been tasked with something I'm not quite grasping. I have have a lexer rule tokenize something that starts with either and underscore or a letter, can't start with 'xml' and can contain periods, dashes, letters, digits and underscores. How would I make sure the beginning is a letter or an underscore and also be sure that it isn't 'xml'?

Answer 1

The grammar:

grammar Expr;
prog: word+ EOF;
word : STRING ;
STRING : ~[xml]STRINGCHAR+;
WS : [ \t\r\n]+ -> skip;
fragment
STRINGCHAR : [_a-z] ;

matches "boy" and "_girl" but not "xmlboy" by using the explicit exclusion ~ operator in the definition of STRING.