[antlr-interest] White space

[Alex Rodriguez]

John,

Yes. And it is infinitely better this way. Thank you, sir.

- Alex :)

On Wed, Sep 1, 2010 at 1:19 PM, John B. Brodie <jbb at acm.org> wrote:
> Greetings!
>
> On Wed, 2010-09-01 at 11:30 -0700, Alex Rodriguez wrote:
>> Greetings,
>>
>> Given a very simple grammar for a language that only has an 'if'
>> statement, I would like to be able to parse white space within literal
>> values. So far, this works (case 1):
>>
>> if(value=='white space'){doThis('arg')}
>>
>> But this doesn't work (case 2):
>>
>> if (value == 'white space') { doThis('arg') }
>
> because you have permitted blanks in an ID, the string "if " above is an
> ID under your rules and *NOT* the keyword 'if' followed by a blank.
>
>>
>> Note that case 2 is spaced for readability.
>>
>> Debugging case 2 in ANTLRWorks produces a MismatchedTokenException.
>>
>> What is the best way to handle both cases? Here is the grammar:
>
> move your literalValue rule into the lexer and take out the blank from
> ID.
>
> see attached for the way i would change your grammar to solve this
> issue.
>
> hope this helps...
>   -jbb
>
>
>>
>> grammar Lang;
>>
>> statement
>>       :
>>               'if' LPAREN ID EQ literalValue RPAREN '{' action '}'
>>       ;
>>
>> literalValue
>>       :
>>               '\'' ID '\''
>>       ;
>>
>> action
>>       :
>>               ID LPAREN literalValue RPAREN
>>       ;
>>
>> ID
>>       :
>>               ('a'..'z' | 'A'..'Z' | '0'..'9' | '@' | ':' | '_' | ' ' | '+')+
>>       ;
>>
>> LPAREN
>>       :
>>               '('
>>       ;
>>
>> RPAREN
>>       :
>>               ')'
>>       ;
>>
>> EQ
>>       :
>>               '=='
>>       ;
>>
>> WS
>>       :
>>               (' ' |'\t' |'\r' |'\n' )+ { $channel=HIDDEN; }
>>       ;
>
>