CEDET

Hi,

If a symbol in clojure can be:

list)

then I think you are out of luck.  How can you ever know if it is a 
paren or not?

If the symbol is instead represented like this in your text file:

list\)

then you will need tup write your own symbol lexer which would probably 
look like this:

"\\(\\w\\|[\\][()]\\)+"
       ^     ^---^-- chars to be quoted
       +--- symbol chars

you would then need to make sure your syntax table lists parens as 
parens, not as symbol constituents.

In this case, it is all about the lexer.  The grammar should be fine.

Good Luck
Eric

On 09/18/2012 08:14 AM, kototama kototama wrote:
> Hi,
>
>
> I'm trying to parse some Clojure code, and I have a rule like this in
> my grammar:
>
> LPAREN atom RPAREN
>
> (Basically here above instead of atom it should be list-content, but
> let's use this example to simplify).
>
> atom:
>    SYMBOL
> | ... other things ...
>
> The problem is ( and ) are valid characters within a symbol in
> Clojure. For instance this expression is valid (\)) and represents a
> list containing the ) character (the expression should be quoted to be
> valid but this does not matter for the example).
>
> I get a failure with the lexer because this expression:
>
> (list)
>
> returns
> ((LPAREN 1 . 2)
>   (SYMBOL 2 . 7))
>
> instead of LPAREN SYMBOL RPAREN since my regular expressions for
> lexing symbols allows the parenthesis.
>
>
> I'm a bit rusted with Bison so I'm not even sure if it's possible to
> achieve what I want with a LALR parser.
>
> Is there a solution?
>
> ------------------------------------------------------------------------------
> Live Security Virtual Conference
> Exclusive live event will cover all the ways today's security and
> threat landscape has changed and how IT managers can respond. Discussions
> will include endpoint security, mobile security and the latest in malware
> threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
> _______________________________________________
> cedet-semantic mailing list
> cedet-semantic@...
> https://lists.sourceforge.net/lists/listinfo/cedet-semantic
>

> then I think you are out of luck.  How can you ever know if it is a paren or
> not?

You are right. That's also the conclusion I reached while sleeping ;-).

The next lines are more related to Clojure than Semantic but I'll
write them for the archive:

The source of the Clojure reader can be found here:
https://github.com/clojure/clojure/blob/master/src/jvm/clojure/lang/LispReader.java

While symbolPat regex allows parenthesis in symbol names, it seems to
me it concretely never happens since the readDelimitedList function
parse such parenthesis block before the symbolPat is even used. But I
would need to run and test the code to confirm.

I also finally don't see how such symbols could be written in the
source code. Also the documentation states

"Symbols begin with a non-numeric character and can contain
alphanumeric characters and *, +, !, -, _, and ? (other characters
will be allowed eventually, but not all macro characters have been
determined)."
http://clojure.org/reader

Actually it's possible to create a symbol containing parenthesis by
calling the symbol function, ex: (symbol "mysy)m").
Such a symbol cannot be read back by the reader!

Using backslash does not work to express a symbol containing such a
character. Backslashes are only used for character literals.

So I'll restrict the lexer to the symbols syntax defined in the
documentation and this should be perfectly fine.

Thanks again.