May 2011 - slime-devel - mailman3.common-lisp.net

[slime-devel] Standard output no longer appears in REPL
by Steven E. Harris 25 Jun '11

25 Jun '11

It's been a couple of months since I last used SLIME. Today I noticed that when I evaluate expressions in the REPL that write to *standard-output*, the output appears in the *inferior-lisp* buffer, but not in the REPL as it used to. I'm using GNU Emacs on Windows atop the CLISP Cygwin build. SLIME is up to date with CVS. Is there some new configuration required to restore the old behavior with output streams feeding into the REPL output? -- Steven E. Harris

5 12

[slime-devel] slime-indentation: common-lisp-style
by Nikodemus Siivola 09 Jun '11

09 Jun '11

I've added a tentative implementation of indentation styles to slime-indentation on my Github tree: https://github.com/nikodemus/Slime/tree/slime-indentation With this, you can do ;; -*- common-lisp-style: stylename -*- which will then set the indentation specs and buffer-local variables specified for that style in common-lisp-style-alist. (Or you can do M-x set-common-lisp-style as well, of course.) The big deal is that with this you can have eg. IF ident with 2 spaces in one file, and with 4 in another -- as fits the project style. Without this I need to have multiple emacses started with different command-line options which is a huge pain -- and I expect I'm not the only one. Here's a sample style: (push '("eccentric" (variables (lisp-lambda-list-keyword-alignment t) (lisp-lambda-list-keyword-parameter-alignment t) (lisp-loop-indent-subclauses nil) (lisp-loop-indent-forms-like-keywords t)) (indentation (if (4 2 2)))) common-lisp-style-alist) Things I'd like to sort out before proposing this for Slime CVS: - tests - support for inheriting styles - define-common-lisp-style so we're not married to the alist spec - get someone with more emacs-fu to review this ...the last in particular I would appreciate help with. I've no idea if the things I'm doing with buffer-local variables are kosher or not. In particular, I feel a bit icky about (add-hook 'lisp-mode-hook (lambda () (hack-local-variables))) which I needed to do because (lisp-mode) kills all local variables. Cheers, -- nikodemus

1 1

[slime-devel] Daily ChangeLog diff
by Helmut Eller 27 May '11

27 May '11

Index: slime/ChangeLog diff -u slime/ChangeLog:1.2194 slime/ChangeLog:1.2195 --- slime/ChangeLog:1.2194 Mon May 23 23:27:55 2011 +++ slime/ChangeLog Fri May 27 00:45:44 2011 @@ -1,3 +1,11 @@ +2011-05-27 Helmut Eller <heller(a)common-lisp.net> + + Fix "wrong number of args" problem with slime-inspector-quit. + Reported by Anton Kovalenko. + + * slime.el (slime-popup-buffer-quit-function) + (slime-quit-threads-buffer): Drop the kill-buffer-p argument. + 2011-05-24 Helmut Eller <heller(a)common-lisp.net> * slime.el (destructure-case): Fix "`_' not left unused" warnings.

1 0

[slime-devel] slime-inspector-quit: wrong number of args
by Anton Kovalenko 27 May '11

27 May '11

Hi, The following commit 2011-05-22 Helmut Eller <heller(a)common-lisp.net> [...] (slime-inspector-quit): Drop the unused kill-buffer arg from the lambda-list. causes problem with closing inspector buffers using "q", because slime-popup-buffer-quit-function expects the callback (specified by variable with the same name) to accept kill-buffer-p argument. P.S. Unfortunately, the snapshot of 2011-05-22 got into quicklisp collection. It would be probably useful to notify Zach Beane directly when this bug is fixed. -- Regards, Anton Kovalenko +7(916)345-34-02 | Elektrostal' MO, Russia

2 1

[slime-devel] Daily ChangeLog diff
by Helmut Eller 26 May '11

26 May '11

Index: slime/contrib/ChangeLog diff -u slime/contrib/ChangeLog:1.460 slime/contrib/ChangeLog:1.461 --- slime/contrib/ChangeLog:1.460 Sat May 21 10:01:48 2011 +++ slime/contrib/ChangeLog Wed May 25 17:27:14 2011 @@ -1,3 +1,8 @@ +2011-05-26 Stas Boukarev <stassats(a)gmail.com> + + * slime-package-fu.el (slime-add-export): Add :export to the end + of a defpackage form when there's no :export present. + 2011-05-21 Helmut Eller <heller(a)common-lisp.net> * slime-snapshot.el: License is GPL.

1 0

[slime-devel] Add hook to swank::dispatch-event
by Paulo Madeira 25 May '11

25 May '11

What do you think about adding a hook to swank::dispatch-event, much alike the mechanism in slime-dispatch-event? The purpose would be to support any set of commands that the contribs add, or even the user, through .swank.lisp or any other means. Right now, you can't send any command to Emacs that isn't in dispatch-event's destructuring-case. Paulo Madeira

4 8

[slime-devel] Daily ChangeLog diff
by Helmut Eller 24 May '11

24 May '11

Index: slime/ChangeLog diff -u slime/ChangeLog:1.2191 slime/ChangeLog:1.2194 --- slime/ChangeLog:1.2191 Sun May 22 00:18:52 2011 +++ slime/ChangeLog Mon May 23 23:27:55 2011 @@ -1,3 +1,25 @@ +2011-05-24 Helmut Eller <heller(a)common-lisp.net> + + * slime.el (destructure-case): Fix "`_' not left unused" warnings. + Insert an (ignore) form for empty bodies; this inhibts the + compiler transformation that leads to the spurios warning. This + has no runtime cost in compiled code, since the compiler + recognizes (ignore) and emits no code for it. + (slime-flatten-tree): Deleted. No longer used. + +2011-05-23 Stas Boukarev <stassats(a)gmail.com> + + * swank.lisp (sdlb-print-condition): New (or old new) + function. It was bound by P in sldb-mode, but for some reason it + was lost. + +2011-05-23 Helmut Eller <heller(a)common-lisp.net> + + * slime.el (slime-apropos-label-properties): Deleted. Just + auto-load apropos-mode. + (slime-print-apropos): Use the variable apropos-label-face + which seems to exist in all relevant Emacsen. + 2011-05-22 Helmut Eller <heller(a)common-lisp.net> Turn on lexical-binding for slime.el.

1 0

[slime-devel] Better way to get pretty print output with ABCL?
by Mark Evenson 23 May '11

23 May '11

SLIME wraps in ABCL in a manner that doesn't use the TOP-LEVEL:REPL function so one doesn't get pretty printed output by default. If I define SWANK::PRESENT-REPL-RESULTS-WITH-PPRINT as below, and bind it as the value of SWANK::*SEND-REPL-RESULTS-FUNCTION*, the the pretty printer is used for ABCL output under SLIME. But I can't just stick this in the "~/.swank.lisp" file, as the value of swank::*send-repl-results-function* gets reset when contrib/swank-presentations.lisp is loaded making me think "I'm doing it wrong." Is the ABCL/SLIME implementation somehow wrong in not using the TOP-LEVEL:REPL routines for evaluation? How does one configure contrib contributions if not in "~/.swank.lisp"? Does ABCL need to be fixed here, or SLIME, or both? (defun present-repl-results-with-pprint (values) ;; Override a function in swank.lisp, so that ;; presentations are associated with every REPL result. (labels ((pprint-value (value) (let ((pprint-results (with-output-to-string (s) (pprint value s)))) (subseq pprint-results 1))) ;; Strip preceding newline of pprint (send (value) (let ((id (and *record-repl-results* (save-presented-object value)))) (send-to-emacs `(:presentation-start ,id :repl-result)) (send-to-emacs `(:write-string ,(pprint-value value) :repl-result)) (send-to-emacs `(:presentation-end ,id :repl-result)) (send-to-emacs `(:write-string ,(string #\Newline) :repl-result))))) (fresh-line) (finish-output) (if (null values) (send-to-emacs `(:write-string "; No value" :repl-result)) (mapc #'send values)))) -- "A screaming comes across the sky. It has happened before, but there is nothing to compare to it now."

1 0

[slime-devel] SLIME protocol version vs SWANK protocol version
by Hugo Duncan 23 May '11

23 May '11

Hi, I am writing a swank server for clojure, and for various reasons, not least of which the frequent updates to SLIME, the versions of slime.el used to connect are usually somewhat old. The clojure swank server could easily support multiple versions of the slime protocol, making it less painful to set up for the many emacs newbies in the clojure world, but to do so would require knowledge of the slime protocol version implemented in slime.el. Would a change in the swank:connection-info message to include slime-protocol-version be a possibility? (swank:connection-info "20110501") This obviously doesn't allow retroactive support of old slime versions, but would provide a path to doing so in the future. -- Hugo Duncan

2 2

[slime-devel] abcl as an inferior lisp program
by Matus Kmit 22 May '11

22 May '11

hi all i am new to this mailing list as well as to slime. upto now i managed to successfully setup and use slime with clisp and sbcl implementations. recently i started to work with abcl, which i already got to work on windows via cygwin + emacs + slime. i am however failing to do so in linux. i build abcl from source using ant and end up having an executable wrapper "abcl" file, kind of shell script which starts java and loads the respective abcl jar file and main class. so when i type "abcl" in linux shell, the abcl REPL opens up (running on JVM). i tried to set up .emacs config file in order to load abcl as an inferior lisp program analogically to what i did on windows using the "abcl.bat" file. i am getting the following error when starting slime in emacs: "let: SIGPIPE raised on process inferior-lisp; closed it" my .emacs file looks like this: (setq inferior-lisp-program "/Documents/abcl-src-0.25.0/abcl") ; your Lisp system (add-to-list 'load-path "~/Documents/slime") ; your SLIME directory (require 'slime) (slime-setup) i am suspecting the problem being abcl a shell script... any help would be greatly appreciated. thank you simplynitaai

2 1

[slime-devel] Trace output not redirected to the *slime-trace* buffer
by Sergio Pozzetti 19 May '11

19 May '11

Hi all, Newbie to SLIME, but I got it working with CCL. I'm trying to find a way to redirect the trace output to a different buffer. Currently, and irrespective of having the *slime-trace* buffer open and having tried M-x slime-redirect-trace-output, all the trace output comes up in the REPL. Thoughts? Thanks -- Sergio

2 2

[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