- slime-devel - mailman3.common-lisp.net

[slime-devel] slime-indentation: what's cooking
by Nikodemus Siivola 19 May '11

19 May '11

I'm now probably going to lay off from making additions to slime-indentation for a while, eating my own dogfood instead of making more. I just wanted to give a heads-up on a few features in my github tree that are not in CVS yet: https://github.com/nikodemus/Slime/tree/slime-indentation 1. Support for named styles. This still needs cleanup, but there has been some progress. You may need to update your styles later if you start using this now, but probably not very much. 2. Per package indentation information. Currently Slime messes things up when sending indentation hints derived from macro-lambda lists if there are multiple macros with the same symbol-name in different packages, with different lambda-lists. I've made it so that things should -- mostly -- go right. 3. Delegating indentation specs. Specifying (as let) as the indentation of another symbol means to indent it as if it was a let -- and if the indentation of let changes, those delegating to it will automatically pick up the new indentation. This is especially important for styles: if the style you are defining inherits from another style that uses (as case) for case, ecase, typecase, etc, you only need to redefine the indentation of case to have everything change. 4. An attempt at deriving better macro-indentation specs, so that eg. (defmacro foo (x (&rest args) &body body) ...) gives us (foo (bar quux zot) (a b c d) (form1) (form2)) instead of (foo (bar quux zot) (a b c d) (form1) (form2)) If anyone wants to test-drive that branch, they're most welcome. Let me know if something doesn't work as expected. Cheers, -- nikodemus

1 0

[slime-devel] future of slime-indentation contrib
by Nikodemus Siivola 15 May '11

15 May '11

Currently slime-indentation contrib has a (patched?) copy of a 1995 vintage cl-indent.el inlined at the tail of slime-indentation.el. I propose the following: * Copy GNU Emacs' current cl-indent.el as contrib/slime-cl-indent.el * Make slime-indentation.el load this file. * Patch this file directly to improve indentation. * Those feeling energetic can send patches to Emacs/XEmacs. In the spirit of doing it rather than talking about it, I pushed a "slime-indentation" branch onto my git tree for review: https://github.com/nikodemus/Slime/tree/slime-indentation It incorporates Didier Verna's recent improvements, and fixes two pet hates of mine. (Distinguish "default" from "def.*" and indent PROG correctly.) I haven't lived with these changes for very long yet though, but so far they seem OK. Unless others feel uncomfortable with this approach, I would like to merge this stuff into the CVS. Cheers, -- Nikodemus

3 5

[slime-devel] MCLIDE 1.0b4
by Terje Norderhaug 13 May '11

13 May '11

FYI, a new version of the MCLIDE Mac Lisp IDE is now released. The swank client includes installers for most supported Lisp implementations. MCLIDE complements SLIME by making it easier to get started with Lisp for those that aren't already savvy emacs users. Here is the official announcement: -------------------------------------- MCLIDE is a free open source Macintosh IDE for Lisp implementations on any platform. It provides a double-clickable, stand-alone development environment that connects to a Lisp either locally or through the network. The new MCLIDE 1.0b4 can install most supported Lisp implementations from the user interface, and has rich interactive development and debugging tools that make it easier to get productive with Lisp: http://mclide.in-progress.com MCLIDE now also officially supports CLISP, in addition to Clojure, SBCL, Clozure CL, LispWorks and MCL. It can potentially be used with the same Lisp implementations as SLIME. MCLIDE provides a consistent developer experience regardless of the Lisp. The built-in programming editor has paren matching, automatic indentation and numerous shortcuts. It comes with plug-ins providing optional Lisp syntax styling and symbol completion. Use it on the same computer as your Lisp system, or target one or more Lisp implementations on other computers or operating systems. Follow the progress of MCLIDE at www.twitter.com/mclide --------------------------------------

1 0

[slime-devel] Indent IF*
by Gábor Melis 13 May '11

13 May '11

Hello Attached is a hack based on the code that indents LOOP (in contrib/slime-indentation.el) to indent IF*. I'm aware that few people hack on code with IF* but to them it's quite useful. You need to load the slime-indentation contrib and this file. It does this: (if* (eq t nil) then () () elseif (dsf) thenret x else (balbkj) (sdf)) Yes, that was the spec and the user doc. Cheers, Gabor

2 1

[slime-devel] Installing Slime on Windows XP
by David Hodge 12 May '11

12 May '11

Dear all, i am trying to install slime (the latest cvs version) on Windows XP. The beahviour below occurs with either clisp or sbcl. The first issue is that when I start slime, lisp always appears in the *inferior lisp* buffer. I can use slime-selector to open a repl. The second, more serious issue, is that when i do anything in the repl , the connection closes and i get a message error in process filter: slime-dispatch-event: Elisp destructure-case failed: (:write-string "2 " :repl-result) This was the result of entering (+ 1 1) in the repl. So, any advice as to where I should start looking fr the issue? Cheers

2 3

[slime-devel] Daily ChangeLog diff
by Helmut Eller 10 May '11

10 May '11

Index: slime/ChangeLog diff -u slime/ChangeLog:1.2188 slime/ChangeLog:1.2189 --- slime/ChangeLog:1.2188 Sun May 8 06:16:20 2011 +++ slime/ChangeLog Mon May 9 06:48:10 2011 @@ -1,3 +1,13 @@ +2011-05-09 Nikodemus Siivola <nikodemus(a)random-state.net> + + * swank-sbcl.lisp (swank-compile-string): Use :SOURCE-NAMESTRING + to provide the name of the actual source file, and + :ALLOW-OTHER-KEYS for compatibility with pre-1.0.48.1 versions. + This allows SBCL to get uninteresting redefinition muffling right. + (*trap-load-time-warnings*): Default to T to make the redefinition + notes more visible, especially now that redefinition muffling hides + the boring ones. + 2011-05-08 Nikodemus Siivola <nikodemus(a)random-state.net> * slime.el (slime-intern-indentation-spec): New function. Converts

1 0

[slime-devel] propose changes to `slime-marcoexpand-again' with patch
by MON KEY 09 May '11

09 May '11

Apropos the bug report here: https://bugs.launchpad.net/slime/+bug/777405 I'd like to propose the following changes to slime.el's macroexpansion functions: `slime-macroexpand-again' `slime-initialize-macroexpansion-buffer' `slime-create-macroexpansion-buffer' These changes make explicit which buffer should be current during calls to `slime-macroexpand-again' and are IMO in keeping with the intent of the command. The changes to `slime-initialize-macroexpansion-buffer' are w/r/t its optional arg BUFFER and provide additional checks to prevent `erase-buffer' and setting `buffer-undo-list' to nil when current-buffer is not "*slime-macroexpansion*". I've included the proposed changes in full below as well as a diff of slime.el from quicklisp 20110320 with slime.el from CVS 1.1364 ;;; ============================== (defun slime-macroexpand-again () "Reperform the last macroexpansion." (interactive) (slime-eval-async slime-eval-macroexpand-expression (slime-rcurry #'slime-initialize-macroexpansion-buffer (slime-buffer-name :macroexpansion)))) (defun slime-initialize-macroexpansion-buffer (expansion &optional buffer) (let ((expansion-buffer (or (and buffer (get-buffer-create buffer)) (slime-create-macroexpansion-buffer)))) (with-current-buffer expansion-buffer (pop-to-buffer (current-buffer)) (let ((inhibit-read-only t)) (if (eq (current-buffer) (get-buffer (slime-buffer-name :macroexpansion))) (progn ;; Get rid of undo information from previous expansions. (setq buffer-undo-list nil) (erase-buffer) (let ((buffer-undo-list t)) ; Make the initial insertion not be undoable. (insert expansion) (font-lock-fontify-buffer))) (let ((point-now (point))) (save-excursion (insert expansion) (font-lock-fontify-region (point-now) (point))))))))) (defun slime-create-macroexpansion-buffer () (let ((name (slime-buffer-name :macroexpansion))) (slime-with-popup-buffer (name :package t :connection t :mode 'lisp-mode) (slime-mode 1) (slime-macroexpansion-minor-mode 1) (setq font-lock-keywords-case-fold-search t) (get-buffer name)))) ;;; ============================== --- slime-cvs-1.1364.el 2011-05-09 14:46:38.000000000 -0400 +++ slime-20110320-cvs.el 2011-05-09 15:28:43.000000000 -0400 @@ -4988,22 +4973,30 @@ #'slime-initialize-macroexpansion-buffer))) (defun slime-macroexpand-again () - "Reperform the last macroexpansion." - (interactive) - (slime-eval-async slime-eval-macroexpand-expression - (slime-rcurry #'slime-initialize-macroexpansion-buffer - (current-buffer)))) + "Reperform the last macroexpansion." + (interactive) + (slime-eval-async slime-eval-macroexpand-expression + (slime-rcurry #'slime-initialize-macroexpansion-buffer + (slime-buffer-name :macroexpansion)))) (defun slime-initialize-macroexpansion-buffer (expansion &optional buffer) - (pop-to-buffer (or buffer (slime-create-macroexpansion-buffer))) - (setq buffer-undo-list nil) ; Get rid of undo information from - ; previous expansions. - (let ((inhibit-read-only t) - (buffer-undo-list t)) ; Make the initial insertion not be undoable. - (erase-buffer) - (insert expansion) - (goto-char (point-min)) - (font-lock-fontify-buffer))) + (let ((expansion-buffer + (or (and buffer (get-buffer-create buffer)) + (slime-create-macroexpansion-buffer)))) + (with-current-buffer expansion-buffer + (pop-to-buffer (current-buffer)) + (let ((inhibit-read-only t)) + (if (eq (current-buffer) (get-buffer (slime-buffer-name :macroexpansion))) + (progn + ;; Get rid of undo information from previous expansions. + (setq buffer-undo-list nil) + (erase-buffer) + (let ((buffer-undo-list t)) ; Make the initial insertion not be undoable. + (insert expansion) + (font-lock-fontify-buffer))) + (let ((point-now (point))) + (save-excursion (insert expansion) + (font-lock-fontify-region (point-now) (point))))))))) (defun slime-create-macroexpansion-buffer () (let ((name (slime-buffer-name :macroexpansion))) @@ -5012,7 +5005,7 @@ (slime-mode 1) (slime-macroexpansion-minor-mode 1) (setq font-lock-keywords-case-fold-search t) - (current-buffer)))) + (get-buffer name)))) (defun slime-eval-macroexpand-inplace (expander) "Substitute the sexp at point with its macroexpansion.

2 3

[slime-devel] Daily ChangeLog diff
by Helmut Eller 09 May '11

09 May '11

Index: slime/ChangeLog diff -u slime/ChangeLog:1.2187 slime/ChangeLog:1.2188 --- slime/ChangeLog:1.2187 Sat Apr 16 01:17:59 2011 +++ slime/ChangeLog Sun May 8 06:16:20 2011 @@ -1,3 +1,14 @@ +2011-05-08 Nikodemus Siivola <nikodemus(a)random-state.net> + + * slime.el (slime-intern-indentation-spec): New function. Converts + strings in an indentation spec to symbols. + (slime-handle-indentation-update): Intern the indentation spec using + slime-intern-indentation-spec, so that lisp side can pass complex indentation + specs. + * contrib/swank-indentation.lisp (application-indentation-hint): Downcase + symbol names -- any symbol in an indentation spec is pretty much guaranteed + to have a lowecase name on the Emacs side. + 2011-04-16 Stas Boukarev <stassats(a)gmail.com> * slime.el (slime-load-failed-fasl): New variable. Accepts `ask',

1 0

[slime-devel] Indentation question
by Nikodemus Siivola 08 May '11

08 May '11

I have a nasty habit of writing macros with multiple keyword clauses in them--like this: (deffoo foo (stuff) (:bar (a b) (cons a b)) (:zot (a b c) (fii a b c)) (:default 42) (:error (crap))) which of course tend to indent terribly (deffoo foo (stuff) (:bar (a b) (cons a b)) (:zot (a b c) (fii a b c)) (:default 42) (:error (crap))) I've occasionally added special indentation rules for these things in .emacs, but that's no fun at all. I've also sometimes used non-keyword symbols and added global definitions which tell slime how to indent things: (defmacro bar (lambda-list &body body) (error "oops, not in DEFFOO")) None of these is really satisfactory: hacking .emacs means users of a library that has macros with this style don't get the correct indentation in Slime out of the box. Using non-keyword symbols makes the code harder to read. (Keywords make it obvious that those are clauses or options of some sort.) Can someone suggest something better? Cheers, -- Nikodemus

7 10

[slime-devel] Can't locate module: SWANK-IO-PACKAGE::SWANK-PACKAGE-FU
by Sebastian Tennant 23 Apr '11

23 Apr '11

Hi Slimers, I'm running the latest stable SBCL: $ sbcl --version 1.0.47 the most recent version of SLIME shipped by quicklisp: $ ls -d ~/quicklisp/dists/quicklisp/software/slime-*-cvs /root/quicklisp/dists/quicklisp/software/slime-20110320-cvs and Emacs 23 shipped with Debian Squeeze: (emacs-version) => "GNU Emacs 23.2.1 (i486-pc-linux-gnu) of 2010-12-11 on raven, modified by Debian" When I try to connect to a running SBCL lisp image containing a SWANK server (same version) I immediately end up in the debugger: --8<---------------cut here---------------start------------->8--- Can't locate module: SWANK-IO-PACKAGE::SWANK-PACKAGE-FU [Condition of type SIMPLE-ERROR] Restarts: 0: [*ABORT] Return to SLIME's top level. 1: [TERMINATE-THREAD] Terminate this thread (#<THREAD "worker" RUNNING {CD36339}>) Backtrace: 0: (SWANK::MODULE-FILENAME SWANK-IO-PACKAGE::SWANK-PACKAGE-FU) 1: (SWANK:SWANK-REQUIRE ..) 2: (SB-INT:SIMPLE-EVAL-IN-LEXENV ..) 3: (EVAL ..) 4: (SWANK:EVAL-FOR-EMACS ..) 5: ((LAMBDA ())) 6: (SWANK-BACKEND::CALL-WITH-BREAK-HOOK #<FUNCTION SWANK:SWANK-DEBUGGER-HOOK> #<FUNCTION (LAMBDA #) {A557C25}>) 7: ((FLET SWANK-BACKEND:CALL-WITH-DEBUGGER-HOOK) #<FUNCTION SWANK:SWANK-DEBUGGER-HOOK> #<FUNCTION (LAMBDA #) {A557C25}>) 8: ((LAMBDA ())) 9: ((FLET #:WITHOUT-INTERRUPTS-BODY-[BLOCK353]358)) 10: ((FLET SB-THREAD::WITH-MUTEX-THUNK)) 11: ((FLET #:WITHOUT-INTERRUPTS-BODY-[CALL-WITH-MUTEX]300)) 12: (SB-THREAD::CALL-WITH-MUTEX ..) 13: (SB-THREAD::INITIAL-THREAD-FUNCTION) 14: ("foreign function: call_into_lisp") 15: ("foreign function: funcall0") 16: ("foreign function: new_thread_trampoline") 17: ("foreign function: #x4002D955") --more-- --8<---------------cut here---------------end--------------->8--- The lisp image uses a core built like this: (mapc #'ql:quickload '("SWANK" ;;; more packages here )) (swank-loader:dump-image "foo.core") sebyte -- Eamcs' AlsaPlayer - Music Without Jolts Lightweight, full-featured and mindful of your idyllic happiness. http://home.gna.org/eap

2 2