sexpfmt

S-expressions are easy for machines to write, but generating formatted S-expressions can be painful. sexpfmt formats an input stream in a consistent way such that the output is both line-diffable and human-readable.

The formatting style used by sexpfmt is highly regular, unlike what many Lispers and Schemers prefer. Each indentation increments spaces by a fixed number of spaces (by default, 2).

(object 
  (object
    (name "croissant")
    (quantity 2))
  (object
    (name "latte")
    (quantity 1)
    (size "tall")))

The S-expression data format used is highly simplified compared to LISP's. There is no support for quote, quasiquote, unquote, or dot pair-builders. The character literal #\ (for space) is not supported either. Use #\space instead. There is also no support for #1234 = ... expressions to construct graphs.

Setup and Installation

• To build and install this tool, you will need Cargo and a Rust toolchain.

• Navigate to the root of this repository with a shell, then run:

  cargo install --path .
  

• To run tests, you will also need bash and Python3

Example Usage

$ cat my-file.sexp | sexpfmt > my-formatted-file.sexp
$ ./build/my-sexp-generator-program arg1 arg2 | sexpfmt >> formatted-logfile.sexp

For examples of sexpfmt's behavior, see the test directory.

TODO

• allow command line options to specify...
  • whether to print help and exit (e.g. -h or --help)
  • whether to normalize bookend tokens
  • the margin width and indent width.
  • file input, directly map file using OS API to handle very large files.
• preserve comments when parsing.
• better error reporting if mis-formatted file is fed as input.
  • e.g. we have almost everything needed to provide line numbers.
• consider whether to support more features like quote, quasiquote, unquote, pair building, etc.
  • explicit support for labels, e.g. (menu :version "0.1.2" :items (list ...))
• better documentation
• C API, binaries for easier integration into expect-testing in other languages.