Emacs Common Lisp is an implementation of Common Lisp, written in Emacs Lisp. It does not yet purport to conform to the ANSI standard since, among other things, CLOS, and pretty printing are missing. However, most other Common Lisp features like lexical closures, packages, readtables, multiple values, bignums, adjustable arrays, etc, are present. At this stage many bugs remain and error checking is sparse.

This implementation provides a Common Lisp environment, separate from Emacs Lisp, running in Emacs. It does not intend to extend Emacs Lisp with Common Lisp functionality; however, Common Lisp functions compile to byte code, so Emacs Lisp functions can call Common Lisp functions and vice versa.

All Emacs Lisp data can be passed unchanged to Common Lisp functions, except vectors, which are used to implement various Common Lisp types not present in Emacs Lisp. An Emacs Lisp vector should be converted to a Common Lisp vector by calling cl-vector.

To convert a Common Lisp vector back to Emacs Lisp, call el-vector. Common Lisp strings and bit vectors should be converted by el-string and el-bool-vector or el-bit-vector (depending on your Emacs flavour).

There is a mailing list for discussion about Emacs Common Lisp. Go to http://mailman.nocrew.org/cgi-bin/mailman/listinfo/emacs-cl to subscribe.

See INSTALL for usage instructions. See HOWTO for some hints on how to do common tasks.

Some algorithms and messages are from SBCL and Gareth McCaughan. Notes on the internals of the implementation follows:

Mapping from Emacs Lisp object types to Common Lisp object types:

EL type CL type bit-vector (XEmacs) simple-bit-vector bool-vector (GNU Emacs) simple-bit-vector character (XEmacs) character compiled-function compiled-function cons cons float single-float hash-table hash-table integer fixnum string simple-string subr compiled-function symbol symbol vector various, type in first element

Common Lisp objects represented by Emacs Lisp vectors:

bignum [BIGNUM <n0> <n1> ...] ratio [RATIO <numerator> <denominator>] complex [COMPLEX <realpart> <imagpart>] character [CHARACTER <code>] string [STRING <size> <string> <offset> <fill-pointer>] char-array [char-array <dimensions> <elements> <offset>] bit-vector [BIT-VECTOR <size> <elements> <offset> <fill-pointer>] bit-array [bit-array <dimensions> <elements> <offset>] simple-vector [SIMPLE-VECTOR <elt> ...] vector [VECTOR <size> <elements> <offset> <fill-pointer>] array [ARRAY <dimensions> <elements> <offset>] interpreted-function [INTERPRETED-FUNCTION <fn> <env> <name>] instance of class C [C <slot> ...]

Emacs feature wish list:

• Hash tables. Done in later versions.
• Weak hash tables. Done in later versions.
• A function that returns the address of an object, for implementing print-unreadable-object :identity t.
• A function that returns the amount of processor time used, for implementing get-internal-run-time.
• A way to portably embed information in byte-code objects.

There are problems with the cl:function macro when its output appears in compiled files:

• When applied to a byte-code function the result will be printed with the #[...] syntax. That works, but separate invokations of cl:function will result in distinct, though equal, code vectors.
• When's applied to a subr, the result will be printed with the #<subr ...> syntax, which will raise a syntax error when loaded.

In general, Emacs' byte compiler doesn't preserve object identity, which is a problem. Here's another case:

• This merges the two distinct strings into one object: (byte-compile `(lambda () (foo ,(copy-seq "abc") ,(copy-seq "abc"))))