people have asked for a "match any of these but only once each" construct. We
only have * or + loops. I need it myself now in antlr. I have
prequelConstructs
: ( ^(o+=OPTIONS .*)
| ^(i+=IMPORT .*)
| ^(t+=TOKENS .*)
)*
{BasicSemanticChecks.checkNumPrequels(gtype, $o, $i, $t);}
;
and then i check the count in support code. Ugly but gets a good message.
Here is current output
error(20): U.g:10:0: repeated grammar prequel spec (option, token, or import);
please merge
error(20): U.g:9:0: repeated grammar prequel spec (option, token, or import);
please merge
error(48): U.g:3:10: illegal option foo
error(26): U.g:6:1: token names must start with an uppercase letter: f
error(48): U.g:10:10: illegal option x
error(48): U.g:13:10: illegal option blech
error(48): U.g:17:16: illegal option ick
error(48): U.g:19:16: illegal option x
from
parser grammar U;
options { foo=bar; k=*; backtrack=true;}
tokens {
f='fkj';
S = 'a';
}
tokens { A; }
options { x=y; }
a
options { blech=bar; greedy=true; }
: ID
;
b : ( options { ick=bar; greedy=true; } : ID )+ ;
c : ID<blue> ID<x=y> ;
Should I think about a (...)^[1] construct or something for v4?
Ter
_______________________________________________
antlr-dev mailing list
[email protected]
http://www.antlr.org/mailman/listinfo/antlr-dev
