[antlr-interest] Trouble parsing a language where '{' has too many meanings

[Felix Schmid]

Richard, thanks for the reply. However, why should this prevent the 
lexer from seeing

{
   Type = Hash
   ShortHelp = "A short comment"
   LongHelp = {
       Some other comment ending with a dot.
   }.

as a comment (where the input is

blubber {
   Type = Hash
   ShortHelp = "A short comment"
   LongHelp = {
       Some other comment ending with a dot.
   }.
   Items {
       FirstName {
           Type = String, ShortHelp = "Hallo"
           LongHelp = {
               Long Explanatory test spanning
               over multiple lines
           }.
       }
       LastName {
           Type = String
           Default = "Blah"
           ShortHelp = "(not so) interesting comment"
       }
   }
}

??

I think my problem is that I have to match the '{' in the parser rules 
because it can occur in so many situations.
What would help was a predicate in the lexer whose value I could set 
from a parser rule. Does ANTLR support this directly?



felix

Richard Clark wrote:
> Try changing the definition for ML_TEXT to put the closing element in
> a single string.
>
> ML_TEXT
>   :    '{'
>       ( options {greedy=false;} : . )*
>       '}.'
>   ;
>
> The lexer doesn't do backtracking, so under the old definition it
> would see  {...} and match it before seeing the "." Automatic error
> recovery would throw awayy the dot as unrecognized (and give an
> error.)
>
> Pulling the closing bracket and dot together '}.' means they'll be
> recognized as a unit.
>
> Run the following in ANTLRWorks' debugger to see it working:
>
> grammar multiBlock;
>
> top    : (block | comment)* ;
>
> comment    : ML_TEXT;
>
> block    : BLOCK ;
>
> ML_TEXT
>   :    '{'
>       ( options {greedy=false;} : . )*
>       '}.'
>   ;
>
> BLOCK    : '{' ('A'..'Z'|'a'..'z'|' ')* '}' ;
>
>
> ...Richard
>
> P.S. Remember that the first rule to match in the lexer wins.