yurloc commented on code in PR #5975:
URL:
https://github.com/apache/incubator-kie-drools/pull/5975#discussion_r1615897263
##########
drools-drl/drools-drl-parser/src/main/antlr4/org/drools/drl/parser/antlr4/DRLParser.g4:
##########
@@ -340,7 +340,8 @@ fromAccumulate := ACCUMULATE LEFT_PAREN lhsAnd
(COMMA|SEMICOLON)
) RIGHT_PAREN
*/
fromAccumulate : (DRL_ACCUMULATE|DRL_ACC) LPAREN lhsAndDef (COMMA|SEMI)
- ( DRL_INIT LPAREN initBlockStatements=chunk? RPAREN COMMA?
DRL_ACTION LPAREN actionBlockStatements=chunk? RPAREN COMMA? ( DRL_REVERSE
LPAREN reverseBlockStatements=chunk? RPAREN COMMA?)? DRL_RESULT LPAREN
resultBlockStatements=chunk RPAREN
+ ( DRL_INIT LPAREN initBlockStatements=chunk? RPAREN COMMA?
DRL_ACTION LPAREN actionBlockStatements=chunk? RPAREN COMMA? DRL_REVERSE LPAREN
reverseBlockStatements=chunk? RPAREN COMMA? DRL_RESULT LPAREN
resultBlockStatements=chunk RPAREN
+ | DRL_INIT LPAREN initBlockStatements=chunk? RPAREN COMMA?
DRL_ACTION LPAREN actionBlockStatements=chunk? RPAREN COMMA? DRL_RESULT LPAREN
resultBlockStatements=chunk RPAREN
Review Comment:
FYI, I've come up with a chunk rule version that can do parenthesis matching
and nesting and is not greedy:
```
chunk : .*? (LPAREN chunk RPAREN .*?)* ;
```
If you want to experiment, change the chunk rule like that and revert lines
343-344. BUT there seems to be a serious performance problem with this approach
as manifested by
`org.drools.testcoverage.regression.FusionAfterBeforeTest#testExpireEventsWhenSharingAllRules`
that times out with this chunk version.
I was curious if it's possible to make the chunk rule behave the same as the
nesting chunk method in the old parser. It's nice that it can be done but it
seems unusable for larger inputs. It's a reminder that we should be careful
about using nongreedy parser subrules as also suggested in the
[docs](https://github.com/antlr/antlr4/blob/master/doc/wildcard.md#nongreedy-parser-subrules):
> Nongreedy subrules should be used sparingly because they complicate the
recognition problem and sometimes make it tricky to decipher how the lexer will
match text.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
