The SExp Library

Overview

The SExp Library supports the reading and writing of structured data using the S-expression syntax. It is a work in progress, and does not fully conform with any formal S-exp specification.

  • End-of-line comments begin with a semicolon (;)

  • An S-Expression is either an atomic token (boolean, number, symbol, or string), a quoted expression, or a list of S-Expressions enclosed in brackets.

  • Quoted expressions are formed by the single-quote character (') followed by an expression.

  • Lists are be delimited by matched pairs of () [] or {}, nested freely.

  • List items are separated with whitespace (space, tab, newlines, or carriage returns).

  • Symbols (or identifiers) begin with an initial character followed by zero or more subsequent characters, where an initial character is either a letter or one of the characters -+.@!$%&*/:<⇒?^_~ and a subsequent character is either an initial character, a decimal digit, or the character #.

  • Booleans are represented by the literals #f (false) and #t (true).

  • Numbers are either signed integers or floating-point numbers; the sign (if present) is one of "'+'," “-," or "~”.

  • Integers may be specified in decimal without any prefix, or in hexadecimal with the prefix "0x". In hex, the value is assumed to be unsigned, so -255 should be written "-0xff" rather than "0x-ff".

  • The format of a floating point number is described by the following regular expression:

    \[ \mathit{sign}^{?}\,\mathit{digit}^{+}\,\mathtt{.}\;\mathit{digit}^{+}\, ([\mathtt{eE}]\,\mathit{sign}^{?}\,\mathit{digit}^{+})^{?}\]

    Notably, “1.” and “.1” are invalid and “1” is parsed as an integer — floats must have a dot with digits on both sides.

  • Strings are enclosed in double quotes (").

    The difference between symbols and strings is that symbols are represented as Atom.atom types, so equality comparisons are more efficient.

The original version of the library was written by Damon Wang at the University of Chicago. It has since been modified and maintained by John Reppy.

Contents

structure SExp

Defines the tree representation of S-expression data.

structure SExpParser

Implements an S-Expression parser.

structure SExpPP

Implements an S-Expression pretty-printer.

structure SExpPrinter

Implements an S-Expression printer that produces condensed output without indentation or line breaks.

Usage

For SML/NJ, include $/sexp-lib.cm in your CM file.

For use in MLton, include $(SML_LIB)/smlnj-lib/SExp/sexp-lib.mlb in your MLB file.