slime-devel September 2004

slime-devel@common-lisp.net

46 participants
71 discussions

Re: [slime-devel] Emacs command for loading Slime and application
by Helmut Eller 28 Sep '04

28 Sep '04

Nicolas Neuss <Nicolas.Neuss(a)iwr.uni-heidelberg.de> writes: > 1. If Slime is not already running, the load command is not executed. > > 2. If Slime is already running, Femlisp is loaded, but the package is not > set. > > Can someone tell me how I can achieve my goal? Yes, this is bit more difficult because most of the commands work asynchronously. I changed slime-start-and-load a bit so that it should be usable for your needs. In general If you need commands of this sort, it's usually the easiest to write a small CL function for the job, so that Emacs can make a single RPC and doesn't need so much synchronization. > PS: Another problem is that executing (slime) apparently sets back some > variables, e.g. *READ-DEFAULT-FLOAT-FORMAT* which I have set to > DOUBLE-FLOAT when loading Femlisp. Why is that? Can't help you here without more details. Which Lisp are you using? Which communication style? Are threads involved? We don't set *READ-DEFAULT-FLOAT-FORMAT* explicit, but there's always the possibility for bug and that we used WITH-STANDARD-IO-SYNTAX in a wrong place. Helmut.

2 4

[slime-devel] Eval defun and compile defun failing on long functions
by Dan Pierson 28 Sep '04

28 Sep '04

Using a version of cmucl similar to 19a slime dies with the following in the *inferior lisp* buffer: ;; Connection to Emacs lost. ;; [Error in function LISP::ASSERT-ERROR: Short read: length=5467 pos=4093] Which looks rather like a packet lenth error of some sort -- 4093 seems suspicious to me. Does anyone have any ideas on how to look into this further? Dan Pierson

2 2

[slime-devel] (no subject)
by Pascal J.Bourguignon 28 Sep '04

28 Sep '04

In a Macintosh Programmer Workshop window: echo Hello Command-RETURN inserts: Hello in the window at the current position. In a Smalltalk Workspace window: 1 + 2 . A-d inserts 3 in the Workspace window at the current position. In an emacs lisp source buffer: (insert "\nHello\n") C-u C-x C-e inserts: Hello nil in the current buffer at the point. In a Common-Lisp source buffer with ilisp: (format t "~&Hello") C-z C-e inserts: Hello NIL in the current buffer at the point. In a Common-Lisp source buffer with slime: (format t "~&Hello") C-x C-e inserts nothing. Hides the result as soon as I type or move the mouse (like for example try to copy-and-paste it from the mini buffer, and I have to manipulate buffers, to reduce the size of my source, to hide it entirely to show the whole output that has been sent in some other buffer. Why so much hatred? I want both the output and the result of the function to be inserted at the point! How can that be done with slime? -- __Pascal Bourguignon__ http://www.informatimago.com/ Our enemies are innovative and resourceful, and so are we. They never stop thinking about new ways to harm our country and our people, and neither do we.

2 2

[slime-devel] Error: Attempting to change the lambda list of swank-compile-string
by John M. Adams 28 Sep '04

28 Sep '04

Dear Friends, Using fresh 1.0 cvs with ACL 7.0 beta I observe: ; Fast loading ; /home/jmadams/.slime/fasl/allegro-7.0.beta/swank-backend.fasl ; ... Error: Attempting to change the lambda list of #<STANDARD-GENERIC-FUNCTION SWANK-COMPILE-STRING> to (STRING &KEY BUFFER POSITION DIRECTORY) results in a confict with method #<STANDARD-METHOD SWANK-COMPILE-STRING (T)> Restart actions (select using :continue): 0: Remove this method 1: retry the load of /home/jmadams/.slime/fasl/allegro-7.0.beta/swank-backend.fasl 2: skip loading /home/jmadams/.slime/fasl/allegro-7.0.beta/swank-backend.fasl 3: retry the load of /home/jmadams/w/slime/swank-loader.lisp 4: skip loading /home/jmadams/w/slime/swank-loader.lisp 5: recompile /home/jmadams/w/slime/swank-loader.lisp 6: Return to Top Level (an "abort" restart). 7: Abort entirely from this (lisp) process. I see that others have reported similar problems, but I have seen no confirmed explanation. Has anyone actually got around this problem? I saw a suggestion that this might be caused by inadvertantly mixing different versions of swank.lisp and slime.el but I don't think that is the problem. I also saw that Helmut couldn't reproduce the problem, but I don't know what ACL version was being referred to. Thanks a lot,

2 1

[slime-devel] Re: SLIME docs
by Dan Pierson 28 Sep '04

28 Sep '04

Christian Lynbech wrote: > I wondered if the docs should mention usefull hacks to set up > functions to start slime with specialised lisp binaries/arguments (for > instance a special core image). I used to use something like these and rather miss them. The problem is that I find ,restart-inferior-lisp too useful to give up and it ignores all of this. Any suggestions for how to fix that? Dan Pierson

3 3

[slime-devel] Common Lisp Browser
by Rui Patrocínio 28 Sep '04

28 Sep '04

Hello, I've attached an extensible browser that I hope is useful and simple enough. It uses a lazy mechanism to expand the information required, so that it can be used to explore large systems. Currently it permits to explore class hierarchies and some xref information. I've a big hack that permits to see trace output in a tree-like fashion, but it's very ugly and implementation dependent and so it hasn't been included. A small interface in swank.lisp is needed, using the new API to MOP in the swank-backend: (defun class-direct-classes (class-name fn) (let ((class (find-class class-name nil))) (when class (mapcar #'(lambda (x) (symbol-name (class-name x))) (funcall fn class))))) (defslimefun mop (type symbol-name) "Return info about classes using mop. When type is: :subclasses - return the list of subclasses of class. :superclasses - return the list of superclasses of class." (let ((symbol (parse-symbol-or-lose symbol-name *buffer-package*))) (ecase type (:subclasses (class-direct-classes symbol #'swank-mop:class-direct-subclasses)) (:superclasses (class-direct-classes symbol #'swank-mop:class-direct-superclasses))))) Rui Patrocínio

2 1

[slime-devel] save SLIME session (and more)
by slime＠pelissero.de 27 Sep '04

27 Sep '04

In SLIME I've been missing the possibility to save my hacking sessions. I turn off my PC every night (my PC's ACPI doesn't work reliably) and the next day I have to spend half an hour to reload everything and figure out what package/system dependency I've screwed up the day before. Educative exercise, but admittedly annoying. The following patch adds this feature. It implements the "save" (or "freeze") shortcut and can interact nicely with Emacs desktop.el. User can type ",save" or click "REPL->Save session" and get a lisp image saved in the ~/.slime directory. To recover a previously saved session just do "M-x slime-recover-session" instead of "M-x slime". If the customisable variable slime-part-of-desktop is true, the saving and recovering is done automatically upon exit/start of Emacs (provided you use desktop.el). It has been tested on SBCL and CMUCL. In CLISP there seem to be some communications problems during the recovery, that I haven't been able to track down. Those are probably due to the fact that CLISP uses a different communication style than CMUCL and SBCL; just a guess. Unfortunately the following patch contains an unrelated modification I wanted to propose to the mailing list some days ago. The message got bounced by the mailing list server (I wasn't subscribed). Just to clarify what is what, my message went more or less like this: > I personally find a bit annoying the fact that the "compiler notes" > window opens up always at half the frame height, even if the list > to display is very short. I'd really appreciate if the window > popped up to a more modest size and expand as necessary (toggling > the tree branches). > > Something like the following patch would keep the window at most half > the frame's height, although starting at much smaller size. > > Index: slime.el > =================================================================== > RCS file: /project/slime/cvsroot/slime/slime.el,v > retrieving revision 1.390 > diff -c -r1.390 slime.el > *** slime.el 13 Aug 2004 21:23:53 -0000 1.390 > --- slime.el 27 Aug 2004 22:59:40 -0000 > *************** > *** 3322,3328 **** > (slime-tree-insert tree "") > (insert "\n"))) > (setq buffer-read-only t) > ! (goto-char (point-min))))) > > (defun slime-alistify (list key test) > "Partition the elements of LIST into an alist. KEY extracts the key > --- 3322,3329 ---- > (slime-tree-insert tree "") > (insert "\n"))) > (setq buffer-read-only t) > ! (goto-char (point-min)) > ! (fit-window-to-buffer nil (window-height))))) > > (defun slime-alistify (list key test) > "Partition the elements of LIST into an alist. KEY extracts the key > *************** > *** 3536,3542 **** > (backward-char 1) > (slime-tree-insert tree prefix) > (delete-char 1) > ! (goto-char start-mark))) > > > ;;;;; Adding a single compiler note > --- 3537,3544 ---- > (backward-char 1) > (slime-tree-insert tree prefix) > (delete-char 1) > ! (goto-char start-mark) > ! (fit-window-to-buffer nil (/ (frame-height) 2) (window-height)))) > > > ;;;;; Adding a single compiler note > Here is the whole lot: Index: slime.el =================================================================== RCS file: /project/slime/cvsroot/slime/slime.el,v retrieving revision 1.390 diff -u -r1.390 slime.el --- slime.el 13 Aug 2004 21:23:53 -0000 1.390 +++ slime.el 27 Sep 2004 17:18:44 -0000 @@ -69,6 +69,8 @@ (defun* slime-setup (&key autodoc typeout-frame) "Setup Emacs so that lisp-mode buffers always use SLIME." (add-hook 'lisp-mode-hook 'slime-lisp-mode-hook) + (add-hook 'desktop-save-hook 'slime-desktop-save-hook) + (add-hook 'desktop-after-read-hook 'slime-desktop-after-read-hook) (when typeout-frame (add-hook 'slime-connected-hook 'slime-ensure-typeout-frame)) (setq slime-use-autodoc-mode autodoc)) @@ -130,6 +132,12 @@ :type 'boolean :group 'slime-ui) +(defcustom slime-part-of-desktop t + "If non-nil, SLIME sessions get saved and restored together +with Emacs desktop." + :type 'boolean + :group 'slime-ui) + ;;;;; slime-lisp (defgroup slime-lisp nil @@ -742,6 +750,7 @@ [ "Send Input" slime-repl-return ,C ] [ "Close and Send Input " slime-repl-closing-return ,C ] [ "Interrupt Lisp process" slime-interrupt ,C ] + [ "Save session" slime-save-session ,C ] "--" [ "Previous Input" slime-repl-previous-input t ] [ "Next Input" slime-repl-previous-input t ] @@ -1199,6 +1208,36 @@ (add-hook 'slime-connected-hook hook) (slime)))) +(defun slime-guess-image-switch (lisp-program-name) + (cdr (assoc (file-name-sans-extension (file-name-nondirectory lisp-program-name)) + '(("lisp" . "-noinit -core") + ("sbcl" . "-userinit /dev/null -sysinit /dev/null -core") + ("clisp" . "-norc -M"))))) + +(defun slime-session-image-name (lisp-program-name) + (expand-file-name (concat "~/.slime/" + (file-name-sans-extension (file-name-nondirectory lisp-program-name)) + ".core"))) + +(defun slime-restore-session () + "Restart Slime on a previously saved Lisp image. It's user's +responsability to make sure that the saved image contains a swank +that is compatible with the version of Slime she is running. +That is, do a rm ~/.slime/*.core after upgrading Slime." + (interactive) + (unless (get-buffer-process (get-buffer "*inferior-lisp*")) + (let ((image-name (slime-session-image-name inferior-lisp-program))) + (if (file-exists-p image-name) + (let ((inferior-lisp-program (format "%s %s %s" + inferior-lisp-program + (slime-guess-image-switch inferior-lisp-program) + image-name))) + (call-interactively 'inferior-lisp) + (set-process-query-on-exit-flag (inferior-lisp-proc) + (not slime-kill-without-query-p)) + (slime-inferior-connect)) + (error "No session has been saved for %S" inferior-lisp-program))))) + ;;;;; Start inferior lisp ;;; ;;; Here is the protocol for starting SLIME: @@ -2966,6 +3005,10 @@ (slime-kill-all-buffers))) (:one-liner "Quit the lisp and close all SLIME buffers.")) +(defslime-repl-shortcut slime-repl-save-session ("save" "freeze") + (:handler #'slime-save-session) + (:one-liner "Save Lisp image and quit closing all SLIME buffers.")) + (defslime-repl-shortcut slime-repl-defparameter ("defparameter" "!") (:handler (lambda (name value) (interactive (list (slime-read-symbol-name "Name (symbol): " t) @@ -3322,7 +3365,8 @@ (slime-tree-insert tree "") (insert "\n"))) (setq buffer-read-only t) - (goto-char (point-min))))) + (goto-char (point-min)) + (fit-window-to-buffer nil (window-height))))) (defun slime-alistify (list key test) "Partition the elements of LIST into an alist. KEY extracts the key @@ -3536,7 +3580,8 @@ (backward-char 1) (slime-tree-insert tree prefix) (delete-char 1) - (goto-char start-mark))) + (goto-char start-mark) + (fit-window-to-buffer nil (/ (frame-height) 2) (window-height)))) ;;;;; Adding a single compiler note @@ -7855,6 +7900,31 @@ (defun sldb-xemacs-post-command-hook () (when (get-text-property (point) 'point-entered) (funcall (get-text-property (point) 'point-entered)))) + +(defun slime-save-session () + "Save and kill the current Lisp hacking session. To restore +the session use slime-restore-session instead of just the slime +command." + (interactive) + (when (slime-connected-p) + (let ((process (inferior-lisp-proc))) + (slime-eval-async `(swank:save-session-image-and-die ,(slime-session-image-name inferior-lisp-program))) + ;; wait untill inferior lisp dies + (while (eq 'run (process-status process)) + (sleep-for 1)) + (slime-kill-all-buffers)))) + +(defun slime-desktop-save-hook () + "Run upon saving the Emacs desktop. See desktop-save-hook." + (when (and slime-part-of-desktop + (get-buffer-process (get-buffer "*inferior-lisp*"))) + (slime-save-session))) + +(defun slime-desktop-after-read-hook () + (when (and slime-part-of-desktop + (file-exists-p (slime-session-image-name inferior-lisp-program)) + (y-or-n-p "Restore SLIME session? ")) + (slime-restore-session))) ;;;; Finishing up Index: swank.lisp =================================================================== RCS file: /project/slime/cvsroot/slime/swank.lisp,v retrieving revision 1.222 diff -u -r1.222 swank.lisp --- swank.lisp 13 Aug 2004 16:14:13 -0000 1.222 +++ swank.lisp 27 Sep 2004 17:19:26 -0000 @@ -2762,6 +2762,13 @@ (add-hook *pre-reply-hook* 'sync-indentation-to-emacs) +(defslimefun save-session-image-and-die (lisp-image) + "Save Lisp image in LISP-IMAGE file and die." + #+cmu (extensions:save-lisp lisp-image) + #+sbcl (sb-ext:save-lisp-and-die lisp-image) + #+clisp (ext:saveinitmem lisp-image) + #+clisp (ext:quit)) + ;;; Local Variables: ;;; eval: (font-lock-add-keywords 'lisp-mode '(("(\\(defslimefun\\)\\s +\\(\\(\\w\\|\\s_\\)+\\)" (1 font-lock-keyword-face) (2 font-lock-function-name-face)))) ;;; End: -- walter pelissero http://www.pelissero.de

1 0

[slime-devel] Daily ChangeLog diff
by Luke Gorrie 25 Sep '04

25 Sep '04

Index: slime/ChangeLog diff -u slime/ChangeLog:1.539 slime/ChangeLog:1.540 --- slime/ChangeLog:1.539 Thu Sep 23 23:39:47 2004 +++ slime/ChangeLog Fri Sep 24 00:24:17 2004 @@ -1,7 +1,8 @@ 2004-09-23 Helmut Eller <heller(a)common-lisp.net> * slime.el (slime-start-and-load): Take arguments so that the - function can be called non-interactively. + function can be called non-interactively. Only start SLIME is if + it is not running. (slime-recompile-bytecode): Don't warn about uses of cl-functions. (slime-reset): Kill all sldb buffers. (slime-goto-location-position): Fix syntax for Emacs 20. @@ -23,6 +24,7 @@ (inspector-content-for-emacs): Use print-part-to-string so that we see cycles in the data structure. (inspect-for-emacs): Minor beautifications. + (load-file-set-package): New function. * swank-cmucl.lisp (frame-package): Implemented. (inspect-for-emacs): Only include stuff that is actually stored in

1 0

[slime-devel] Daily ChangeLog diff
by Luke Gorrie 24 Sep '04

24 Sep '04

Index: slime/ChangeLog diff -u slime/ChangeLog:1.538 slime/ChangeLog:1.539 --- slime/ChangeLog:1.538 Sun Sep 19 10:04:57 2004 +++ slime/ChangeLog Thu Sep 23 23:39:47 2004 @@ -1,3 +1,34 @@ +2004-09-23 Helmut Eller <heller(a)common-lisp.net> + + * slime.el (slime-start-and-load): Take arguments so that the + function can be called non-interactively. + (slime-recompile-bytecode): Don't warn about uses of cl-functions. + (slime-reset): Kill all sldb buffers. + (slime-goto-location-position): Fix syntax for Emacs 20. + (sldb-mode-map): Add C-c C-d bindings. + (slime-open-inspector): Insert the type in the second line so that + we can make longer titles, e.g we should include the princed + version of the inspected object. + + * swank-backend.lisp (frame-package, label-value-line) + (label-value-line*): New functions. + + * swank.lisp (frame-locals-for-emacs): Bind *print-pretty* to + *sldb-pprint-frames* to get more compact lines and bind *package* + to frame-package to get shorter labels for variables. + (format-values-for-echo-area): Include the hex and octal + representation for integers. + (apply-macro-expander, disassemble-symbol): Use the buffer-package + for reading. + (inspector-content-for-emacs): Use print-part-to-string so that we + see cycles in the data structure. + (inspect-for-emacs): Minor beautifications. + + * swank-cmucl.lisp (frame-package): Implemented. + (inspect-for-emacs): Only include stuff that is actually stored in + the object itself (see objdef.lisp for exact object layout). + Include the disassembly for functions and code components. + 2004-09-19 Helmut Eller <e9626484(a)stud3.tuwien.ac.at> * swank-gray.lisp (stream-read-char): Treat empty strings as

1 0

[slime-devel] Re: SLIME docs - contributors
by Christophe Rhodes 23 Sep '04

23 Sep '04

Christian Lynbech <christian.lynbech(a)ericsson.com> writes: > In the Credits chapter, there is a footnote about the thanks for DanB > applying to his SBCL personality. > > I am unsure what the definition of "one of us" is but if that is being > listed in the hackers list above, that same footnote should perhaps > also apply to Christophe Rhodes and Martin Simmons. I believe that's referring to the changes Dan made to the SBCL end of things (the sb-introspect module of SBCL, for instance). Martin Simmons and I both have moderately substantial portions of code in SLIME proper: Martin improvements to the Lispworks backend, and me support for jumping to documentation from conditions in the debugger and from the compiler, enhancements to the hyperspec lookup, and so on. Cheers, Christophe

1 0

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

slime-devel September 2004