问题
Using ANTLR4 and the Netbeans Platform I have created a grammar for which I have implemented syntax highlighting in my Netbeans Platform application. Everything works fine up to the point when I try to alter my code.
grammar MyRule;
my_rule : '(' my_rule ')'
| binary
| binary_hyst
| my_rule (AND my_rule)+
| my_rule (OR my_rule)+
;
binary : '(' binary ')'
| unary (EQ | NE) unary
;
binary_hyst : '(' binary_hyst ')'
| unary hyst? ( GT | LT | GTE | LTE ) unary hyst?
;
hyst : (HYST '(' unary ')');
unary : '(' unary ')'
| NUMBER
| STRING
| (MOV | DEC | INC) '(' unary ',' NUMBER ')'
| unary '*' unary
| unary '/' unary
| unary '+' unary
| unary '-' unary
;
// TOKENS
HYST : ('hyst');
MOV : ('mov');
INC : ('inc');
DEC : ('dec');
AND : ('&&');
OR : ('||');
EQ : ('==');
NE : ('!=');
GT : ('>');
LT : ('<');
GTE : ('>=');
LTE : ('<=');
NUMBER
: '-'? INT '.' [0-9]+ // 1.35, 0.3, -4.5
| '-'? INT // -3, 45
;
fragment INT : '0' | [1-9] [0-9]* ; // no leading zeros
STRING : '"' ( ESC | . )*? '"' ;
ID : [a-zA-Z_]+;
fragment ESC: '\\' [btnr"\\] ;
WS : [ \t\n\r]+ -> channel(HIDDEN) ;
This grammar allows code like this: ("some_string" >= 20)
. But when I try to alter my code (i.e. ("some_id" >= 20) && ("some_other_id < 10)
I receive an IllegalStateException:
java.lang.IllegalStateException: Lexer de.nordsys.catelogconfig.rulelanguagesupport.lexer.TriggerRuleNBLexer@181848
returned null token but lexerInput.readLength()=2
lexer-state: null
tokenStartOffset=20, readOffset=22, lookaheadOffset=23
Chars: "&\n" - these characters need to be tokenized.
Fix the lexer to not return null token in this state.
at org.netbeans.lib.lexer.LexerInputOperation.checkLexerInputFinished(LexerInputOperation.java:457)
at org.netbeans.lib.lexer.LexerInputOperation.nextToken(LexerInputOperation.java:217)
at org.netbeans.lib.lexer.inc.TokenListUpdater.relex(TokenListUpdater.java:627)
at
...
line 1:1 token recognition error at: '&\n'
This happens right when I type the first '&' for the '&&' token. It seems that Netbeans wants to parse my code and the lexer recognizes that this is not a valid token, which is correct. If I simply copy and paste a '&&' it works fine and Netbeans doesn't throw an exception. So this seems to be coming from me typing in incomplete tokens. How can I resolve this?
回答1:
You need to ensure that tokens are created for all possible input sequences. The most straightforward way to do this is add an ERR_CHAR
rule to the end of your lexer.
ERR_CHAR
: .
;
This allows the syntax highlighter to work, and defers reporting of the error to the parsing phase.
来源:https://stackoverflow.com/questions/26100737/ilegalstateexception-when-i-type-unfinished-code-into-a-jeditorpane-with-syntax