[antlr-interest] java.g - a couple of issues

[Davin McCall]

Thanks, but it actually seems like the line numbers are off if we leave 
the call to "newline()" _in_.

As far as I can tell, because we are no longer matching the newline 
character sequence in SL_COMMENT, it instead gets matched in the 
whitespace target, which includes a call to newline() anyway:

// Whitespace -- ignored
WS    :    (    ' '
        |    '\t'
        |    '\f'
            // handle newlines
        |    (    options {generateAmbigWarnings=false;}
            :    "\r\n"  // Evil DOS
            |    '\r'    // Macintosh
            |    '\n'    // Unix (the right way)
            )
            { newline(); }
        )+
        { _ttype = Token.SKIP; }
    ;

- Davin


Jon Schewe wrote:

>One problem with this is that if you remove the calls to newline(),
>you're line numbers will be off.  This makes error reporting rather
>difficult.
>
>On Fri, 2005-03-18 at 12:24 +1100, Davin McCall wrote:
>  
>
>>To clarify, we have changed the single line comment target "SL_COMMENT" 
>>as follows:
>>
>>(Original)
>>// Single-line comments
>>SL_COMMENT
>>    :    "//"
>>        (~('\n'|'\r'))* ('\n'|'\r'('\n')?)
>>        {$setType(Token.SKIP); newline();}
>>    ;
>>
>>(Modified)
>>// Single-line comments
>>SL_COMMENT
>>    :    "//"
>>        (~('\n'|'\r'))*
>>        {$setType(Token.SKIP); }
>>    ;
>>
>>And the typeArgumentBounds target from:
>>
>>typeArgumentBounds[boolean addImagNode]
>>    {boolean isUpperBounds = false;}
>>    :
>>        ( "extends"! {isUpperBounds=true;} | "super"! ) 
>>classOrInterfaceType[addImagNode]
>>        {
>>            if (isUpperBounds)
>>            {
>>                #typeArgumentBounds = 
>>#(#[TYPE_UPPER_BOUNDS,"TYPE_UPPER_BOUNDS"], #typeArgumentBounds);
>>            }
>>            else
>>            {
>>                #typeArgumentBounds = 
>>#(#[TYPE_LOWER_BOUNDS,"TYPE_LOWER_BOUNDS"], #typeArgumentBounds);
>>            }
>>        }
>>    ;
>>
>>To:
>>
>>typeArgumentBounds[boolean addImagNode]
>>    {boolean isUpperBounds = false;}
>>    :
>>        ( "extends"! {isUpperBounds=true;} | "super"! ) 
>>classOrInterfaceType[addImagNode] arraySpecOpt
>>        {
>>            if (isUpperBounds)
>>            {
>>                #typeArgumentBounds = 
>>#(#[TYPE_UPPER_BOUNDS,"TYPE_UPPER_BOUNDS"], #typeArgumentBounds);
>>            }
>>            else
>>            {
>>                #typeArgumentBounds = 
>>#(#[TYPE_LOWER_BOUNDS,"TYPE_LOWER_BOUNDS"], #typeArgumentBounds);
>>            }
>>        }
>>    ;
>>
>>
>>Davin
>>
>>
>>Davin McCall wrote:
>>
>>    
>>
>>>Hi,
>>>We're using Michael Studman's java 1.5 grammar as a base for our own 
>>>and have noticed the following issues. We've got fixes but I wanted to 
>>>both highlight the issues to the community and see if anyone else has 
>>>comments (especially in regards to whether the fixes are correct :-) ).
>>>
>>>1. Single line comments without a terminating newline at the end of a 
>>>file cause a parse error, due to the requirement that the comment is 
>>>followed by a line-end.
>>>
>>>It seems to be possible to just remove that requirement (and also the 
>>>call to newline()) and then everything works fine (ie. the match is 
>>>greedy anyway).
>>>
>>>2. Arrays aren't allowed as wildcard bounds (they should be).
>>>
>>>For instance, "? extends Thread[]" as a wildcard in some declaration 
>>>causes a parse error. A possible fix is to add "arraySpecOpt" in to 
>>>the "typeArgumentBounds" target.
>>>
>>>Davin
>>>
>>>
>>>
>>>      
>>>
>
>________________________________________________________________________
>Jon Schewe | http://mtu.net/~jpschewe
>GPG signature at http://mtu.net/~jpschewe/gpg.sig.html
>For I am convinced that neither death nor life, neither angels 
>nor demons, neither the present nor the future, nor any 
>powers, neither height nor depth, nor anything else in all 
>creation, will be able to separate us from the love of God that 
>is in Christ Jesus our Lord. - Romans 8:38-39
>
>
>
>
>
>  
>