[#ANTLR-95] ~T in lexer uses T as token not set of char in rule T

ANTLR v3

~T in lexer uses T as token not set of char in rule T

Created: 14/Mar/07 11:17 AM Updated: 14/Mar/07 01:24 PM

Component/s:

ANTLR Core

Affects Version/s:

3.0b6

Fix Version/s:

3.0b7

Description

« Hide

Jim Idle reports:

lexer grammar jim;

fragment EOLTOK
: ( '\n' | '\r' )
;

ALINE : ((~EOLTOK))+ ;

ANOTHER : ( (~('\n' | '\r')) + );

You will see that ALINE calculates the set using the token number assigned to EOLTOK, and so the set is:

{'\u0000'..'\u0003', '\u0005'..'\uFFFE'}

Whereas ANOTHER calculates the correct set as it uses the literals it finds and negates that.

All

Comments

Change History

Sort Order:

Ascending order - Click to sort in descending order

[ Permlink | « Hide ]

Terence Parr - 14/Mar/07 01:24 PM

Added unit tests to TestCharDFA. Had to fix ~T to go convert T to a set even if T refs U, another lexer rule.