diary at Telent Netowrks

Did I mention that people are asking for good Lisp/Emacs integration#

Thu, 09 Oct 2003 05:07:03 +0000

Did I mention that people are asking for good Lisp/Emacs integration? This is why I'm looking at SLIME for the cirCLe CD; the options available seem to be ILISP, ELI or SLIME, and SLIME. None of them actually work properly with right now, but the last has (a) real momentum behind it, (b) intentions to run on SBCL at some future date - unlike ELI, which will always be a Franz thing at heart, and (c) a refreshingly low amount of cruft.

And that's "work properly" which it doesn't, not "work at all", which it does. After writing the following entry I found a CVS root of Slime as it was ten days ago on Luke Gorrie's web site, so I spent a few hours hacking at it. By means of some really rather ugly copy/paste programming, I have increased by about 30% the amount of duplicated code in SLIME (ugly but inescapable; the most likely outcome of any serious refactoring I could do right now would be unresolvable merge conflicts) and got something for which the basic functionality runs in SBCL.

;;; Cursory testing has found that the following appear to work
;;; * Symbol completion.
;;; * Evaluation of forms with C-M-x.
;;; * Compilation of defuns with C-c C-c.
;;; * File compilation with C-c C-k (but error parsing is broken)
;;; * Disassembling the symbol at point with C-c M-d.
;;; * Describing symbol at point with C-c C-d.
;;; * Macroexpanding with C-c RET.
;;; Things that aren't done yet:
;;; * Cross-referencing.
;;; * Apropos.
;;; * Test suite has not been attempted yet
;;; * Basic debugger functionality, jumping to frames is not implemented yet.
;;; * Compiler warnings aren't currently sent to the Emacs stream