NAME

Data::SExpression -- Parse Lisp S-Expressions into perl data structures.

SYNOPSIS

use Data::SExpression;

my $ds = Data::SExpression->new;

$ds->read("(foo bar baz)");          # [\*::foo, \*::bar, \*::baz]

my @sexps;
my $sexp;
while(1) {
    eval {
        ($sexp, $text) = $ds->read($text);
    };
    last if $@;
    push @sexps, $sexp;
}

$ds = Data::SExpression->new({fold_alists => 1});

$ds->read("((top . 4) (left . 5))");  # {\*::top => 4, \*::left => 5}

METHODS

new [\%args]

Returns a new Data::SExpression object. Possibly args are:

fold_lists

If true, fold lisp lists (e.g. "(1 2 3)") into Perl listrefs, e.g. [1, 2, 3]

Defaults to true.

fold_alists

If true, fold lisp alists into perl hashrefs. e.g.

"((fg . red) (bg . black) (weight . bold))"

would become

{
    \*fg       => \*red,
    \*bg       => \*black,
    \*weight   => \*bold
}

Alists will only be folded if they are a list of conses, all of which have scalars as both their car and cdr (See "scalarp" in Data::SExpression::Cons)

This option implies "fold_lists"

Defaults to false.

symbol_case

Can be "up", "down", or undef, to fold symbol case to uppercase, lowercase, or to leave as-is.

Defaults to leaving symbols alone.

use_symbol_class

If true, symbols become instances of Data::SExpression::Symbol instead of globrefs.

Defaults to false

fold_dashes

If true, dash characters in symbols (-) will be folded to the more perlish underscore, _. This is especially convenient when symbols are being converted to globrefs.

Defaults to false.

read STRING

Parse an SExpression from the start of STRING, or die if the parse fails.

In scalar context, returns the expression parsed as a perl data structure; In list context, also return the part of STRING left unparsed. This means you can read all the expressions in a string with:

my @sexps;
my $sexp;
while(1) {
    eval {
        ($sexp, $text) = $ds->read($text);
    };
    last if $@;
    push @sexps, $sexp;
}

This method converts Lisp SExpressions into perl data structures by the following rules:

Numbers and Strings become perl scalars

Lisp differentiates between the types; perl doesn't.

Symbols become globrefs in main::

This means they become something like \*main::foo, or \*::foo for short. To convert from a string to a symbol, you can use "qualify_to_ref" in Symbol, with "main" as the package.

But see "use_symbol_class" if you'd prefer to get back objects.

Conses become Data::SExpression::Cons objects

See Data::SExpression::Cons for how to deal with these. See also the fold_lists and fold_alists arguments to "new".

If fold_lists is false, the Lisp empty list () becomes the perl undef. With fold_lists, it turns into [] as you would expect.

Quotation is parsed as in scheme

This means that "'foo" is parsed like "(quote foo)", "`foo" like "(quasiquote foo)", and ",foo" like "(unquote foo)".

LISP-LIKE CONVENIENCE FUNCTIONS

These are all generic methods to make operating on cons's easier in perl. You can ask for any of these in the export list, e.g.

use Data::SExpression qw(cons consp);

cons CAR CDR

Convenience method for Data::SExpression::Cons->new(CAR, CDR)

consp THING

Returns true iff THING is a reference to a Data::SExpression::Cons

scalarp THING

Returns true iff THING is a scalar -- i.e. a string, symbol, or number

Data::SExpression::Parser callbacks

These are for internal use only, and are used to generate the data structures returned by "read".

new_cons CAR CDR

Returns a new cons with the given CAR and CDR

new_symbol NAME

Returns a new symbol with the given name

new_string CONTENT

Returns a new string with the given raw content

BUGS

None known, but there are probably a few. Please reports bugs via rt.cpan.org by sending mail to:

bug-Data-SExpression@rt.cpan.org

AUTHOR

Nelson Elhage <nelhage@mit.edu>

cpanm Data::SExpression

perl -MCPAN -e shell
install Data::SExpression