December 2004 - slime-devel - mailman3.common-lisp.net

[slime-devel] comma shortcut doesn't work always
by Edi Weitz 31 Jan '06

31 Jan '06

Another minor glitch that I encountered is that the comma shortcut in the REPL doesn't always work. I can consistently reproduce it (GNU Emacs 21.3.50 and also 21.3 release) like this: 1. Make sure point is at the beginning of a REPL input line. 2. Split Emacs frame into two windows, one of them being the REPL the other one containing something else. 3. Hide REPL window with C-x 0. 4. Now type M-x slime. 5. Type the comma. At this point I'm not asked for a shortcut command in the minibuffer but the comma is simply entered into the REPL instead. Anybody else seeing this? Cheers, Edi.

10 25

Re: [slime-devel] Re: CLISP SLIME implementation breaks in CVS
by Bill_Clementson＠peoplesoft.com 21 Jan '06

21 Jan '06

Luke Gorrie <luke(a)bluetail.com> writes: > Bill_Clementson(a)peoplesoft.com writes: > > > My previous SLIME installation was a CVS install from 4/18. I've just taken > > the latest CVS install and the CLISP SLIME implementation (under Windows) > > breaks. Here is the error: > > > > COERCE: #<MACRO #<COMPILED-CLOSURE SWANK::DEFSLIMEFUN>> cannot be coerced > > to type FUNCTION > > [Condition of type SIMPLE-TYPE-ERROR] > > I can reproduce this but I've no idea what is causing it. It /looks/ > like someone is trying to call defslimefun as a function rather than a > macro, but I don't see where it could happen -- the macro is declared > at the top of the only file that refers to it. > > I did fix a different but similar problem that I got under CLISP. It > seems like I had to swap the order of a couple of macros, one of which > expanded into into code that calls the other, but that surely can't be > the reason. > > > Backtrace: > > 0: EVAL frame for form (SWANK:CREATE-SWANK-SERVER 4007) > > Mine actually gives an empty backtrace (!). I fixed the Elisp-side to > handle that incase it helps someone more CLISP-savvy find the problem. Has anybody had any luck in finding the source of this problem? CLISP still breaks for me in the latest CVS. I have had a look at all of the DEFSLIMEFUN's that were modified since 4/18 (the previous CVS version that I had which was working ok with CLISP) and haven't found any obvious clues. Any suggestions on how I might debug this issue as it occurs at startup during the slime-connect processing in Emacs (after CLISP has started up successfully)? - Bill

5 6

[slime-devel] Problems with LWW FLI
by Edi Weitz 29 Mar '05

29 Mar '05

Hi! Has anyone else problems with SLIME when using Lispworks on Windows (XP) and its FLI? I've run into a (reproducable but complicated) situation where SLIME simply hung while there were no problems from the LW IDE or from the command line. I'm sitting at my Linux laptop now but I can try to create a simple test case - I just wanted to check if this is a known problem. Cheers, Edi.

4 11

[slime-devel] eval-in-emacs and 'window'-debugger
by Antonio Menezes Leitao 10 Feb '05

10 Feb '05

Hi everybody, I noticed that slime provides a function eval-in-emacs that sends a form to the emacs side where it is evaluated. However, I would like to get the results on the lisp side but I always get NIL. Is there a different swank function for that? (in the emacs-lisp-interface from Franz I used a lep::eval-in-emacs that did precisely what I needed). This is very useful to provide refactoring operations (that operate on the Lisp side) that need to interact with the user using emacs lisp operations such as y-or-n-p or read-from-minibuffer. I know that it might be difficult to find a general scheme to translate from emacs lisp values to common lisp values but all I need is symbols, strings, numbers and such. ----------------- A few years ago I developed a 'window'-debugger and 'window'-inspector for Ilisp+CMUCL similar to what I used in the Explorer Lisp Machines. I recently adapted it to Slime+Whatever-Common-Lisp-slime-supports. If anyone is interested I can post it. However, I would like to suggest two changes to Slime in order to improve its portability: 1. The function swank::inspected-parts should provide a recognizable value for unbound slots. Here is one example in SBCL CL-USER> (defclass c () ((a))) #<STANDARD-CLASS C> CL-USER> (swank::inspected-parts (make-instance 'c)) "The object is a STANDARD-OBJECT of type C. " (("A" . #:INSPECT-UNBOUND-OBJECT-3530)) and in CLISP CL-USER> (swank::inspected-parts (make-instance 'c)) "#<COMMON-LISP-USER::C #x20426856> standard object type: COMMON-LISP-USER::C" (("A" . INSPECT-TMP-PACKAGE-5413::|#<unbound>|)) I know that the returned values are recognizable (e.g., by being the value of the variable sb-impl::*inspect-unbound-object-marker* or something similar) but given the fact that it depends on the Common Lisp being used, it would be preferable if Slime provided a function that recognizes such value. If it already does, my apologies. 2. The function swank::frame-locals doesn't distinguish valid locals from invalid ones. I know that not all Common Lisps make that distinction but for the ones that make, it would be useful to get that information (for Common Lisps that don't provide such information a default could be chosen). Currently, Slime puts a :not-available as value which, IMHO, is error-prone. It would be nicer to have an extra field for each local that describes its status or (as in 1.) a slime function that recognizes such special value or as a third (and not so good) alternative, swank::frame-locals should not produce invalid locals. Thanks in advance, António Leitão.

2 2

[slime-devel] Arglist while in SLIME debugger
by Lynn Quam 28 Jan '05

28 Jan '05

I find it somewhat annoying that I cannot see the arglist information for functions while in the SLIME debugger. Is it possible to change SLIME-BUSY-P so that SLIME-AUTODOC-MODE and SLIME-SPACE will continue to work while inside the SLIME debugger? To implement this, the SLIME-BUSY-P needs to know the rex-continuations for the forms that cause the debugger to be entered. The SLIME debugger needs to be able to figure out the rex-continuation of the form that caused the debugger to be entered and add it to the list in the global variable *SLDB-REX-CONTINUATIONS*. When the debugger exits, it needs to remove it from the list. SLIME-BUSY-P would become something like: (defvar *sldb-rex-continuations* nil) (defun slime-busy-p () "True if Lisp has outstanding requests." (let ((sldb-continuations *sldb-rex-continuations*)) (dolist (contin (slime-rex-continuations) nil) (if (not (memq contin sldb-continuations)) (return t)))))

2 3

[slime-devel] Patch to make slime-insert-arglist more useful
by Matthias Koeppe 19 Jan '05

19 Jan '05

Please find below a patch that makes C-c C-s (slime-insert-arglist) much more useful. Instead of inserting a formal argument list, it inserts a template for a function call; this makes a difference for functions with optional and keyword arguments. After inserting the arguments, point is left on the first argument, like ILISP's C-u M-x arglist-lisp does; this allows editing the arguments easily. Index: slime.el =================================================================== RCS file: /project/slime/cvsroot/slime/slime.el,v retrieving revision 1.443 diff -u -p -u -r1.443 slime.el --- slime.el 16 Dec 2004 22:24:41 -0000 1.443 +++ slime.el 17 Dec 2004 13:24:07 -0000 @@ -4227,7 +4227,17 @@ more than one space." "Insert the argument list for NAME behind the symbol point is currently looking at." (interactive (list (slime-read-symbol-name "Arglist of: "))) - (insert (slime-eval `(swank:arglist-for-insertion ',name)))) + (let ((arglist (slime-eval `(swank:arglist-for-insertion ',name)))) + (cond + ((eq arglist :not-available) + (error "Arglist not available")) + ((string-match "^(" arglist) + (insert " ") + (save-excursion + (insert (substring arglist 1)))) + (t + (save-excursion + (insert arglist)))))) (defun slime-get-arglist (symbol-name) "Return the argument list for SYMBOL-NAME." Index: swank.lisp =================================================================== RCS file: /project/slime/cvsroot/slime/swank.lisp,v retrieving revision 1.272 diff -u -p -u -r1.272 swank.lisp --- swank.lisp 16 Dec 2004 21:16:50 -0000 1.272 +++ swank.lisp 17 Dec 2004 13:25:18 -0000 @@ -1143,17 +1143,108 @@ pretty printing of (function foo) as #'f (*print-length* 10) (*print-circle* t)) (format nil "~A => ~A" sym (symbol-value sym))))))) +(defun decode-keyword-arg (arg) + "Decode a keyword item of formal argument list. +Return three values: keyword, argument name, default arg." + (cond + ((symbolp arg) + (values (intern (symbol-name arg) keyword-package) + arg + nil)) + ((and (consp arg) + (consp (car arg))) + (values (caar arg) + (cadar arg) + (cadr arg))) + ((consp arg) + (values (intern (symbol-name (car arg)) keyword-package) + (car arg) + (cadr arg))) + (t + (error "Bad keyword item of formal argument list")))) +(progn + (assert (equal (multiple-value-list (decode-keyword-arg 'x)) '(:x x nil))) + (assert (equal (multiple-value-list (decode-keyword-arg '(x t))) '(:x x t))) + (assert (equal (multiple-value-list (decode-keyword-arg '((:x y)))) '(:x y nil))) + (assert (equal (multiple-value-list (decode-keyword-arg '((:x y) t))) '(:x y t)))) +(defun decode-optional-arg (arg) + "Decode an optional item of a formal argument list. +Return two values: argument name, default arg." + (etypecase arg + (symbol + (values arg nil)) + (list + (values (car arg) (cadr arg))))) +(progn + (assert (equal (multiple-value-list (decode-optional-arg 'x)) '(x nil))) + (assert (equal (multiple-value-list (decode-optional-arg '(x t))) '(x t)))) + +(defun arglist-to-template-string (arglist package) + "Print the list ARGLIST for insertion as a template for a function call." + (setq arglist (clean-arglist arglist)) + (etypecase arglist + (null "()") + (cons + (with-output-to-string (*standard-output*) + (with-standard-io-syntax + (let ((*package* package) (*print-case* :downcase) + (*print-pretty* t) (*print-circle* nil) (*print-readably* nil) + (*print-level* 10) (*print-length* 20)) + (pprint-logical-block (nil nil :prefix "(" :suffix ")") + (let ((mode nil)) + (loop + (let ((arg (pop arglist))) + (typecase arg + ((member &key &optional &rest &body) + (setq mode arg)) + (string + (princ arg) + (unless (null arglist) + (write-char #\space))) + (t + (case mode + (&key + (multiple-value-bind (keyword arg-symbol default-arg) + (decode-keyword-arg arg) + (declare (ignore default-arg)) + (write keyword) + (write-char #\space) + (princ arg-symbol))) + (&optional + (multiple-value-bind (arg-symbol default-arg) + (decode-optional-arg arg) + (declare (ignore default-arg)) + (write-char #\[) + (princ arg-symbol) + (write-char #\]))) + (&body + (pprint-newline :mandatory) + (princ arg) + (princ "...")) + (&rest + (princ arg) + (princ "...")) + (otherwise + (princ arg))) + (unless (null arglist) + (write-char #\space))))) + (when (null arglist) (return)) + (pprint-newline :fill)))))))))) + +(defun test-print-template (list string) + (string= (arglist-to-template-string list (find-package :swank)) string)) + (defslimefun arglist-for-insertion (name) (with-buffer-syntax () (cond ((valid-operator-name-p name) (let ((arglist (arglist (parse-symbol name)))) (etypecase arglist ((member :not-available) - " <not available>") + :not-available) (list - (arglist-to-string arglist *buffer-package*))))) + (arglist-to-template-string arglist *buffer-package*))))) (t - " <not available>")))) + :not-available)))) ;;;; Evaluation -- Matthias Köppe -- http://www.math.uni-magdeburg.de/~mkoeppe

2 2

[slime-devel] More completion shortcuts
by Edi Weitz 18 Jan '05

18 Jan '05

OK, here's another little completion shortcut that I think would be nice to have. YMMV, of course. Good night, Edi.

3 4

[slime-devel] *read-default-float-format* with C-c C-k, , load-system, etc.
by Dirk Gerrits 17 Jan '05

17 Jan '05

Hi, I'm not 100% sure this is a SLIME question, an ASDF question, or a Common Lisp question, but I'll ask it here anyway. In my program I'd like to have *read-default-float-format* be long-float instead of the default single-float. To accomplish this, I added (eval-when (:compile-toplevel :load-toplevel :execute) (setq *read-default-float-format* 'long-float)) to package.lisp, which should be the first file to be compiled/loaded by ASDF. Now if I do a ,load-system in the SLIME REPL, everything is fine: *read-default-float-format* evaluates to long-float during compilation and also does in the REPL. However, if I open one of my files, and add (defun foo () 50.0), do a C-M-x to compile it, and then call (foo) in the REPL, I get 50.0f0 instead of 50.0d0. C-c C-k doesn't help either, and C-x C-e on the 50.0 also produces a single-float. If I then save the file and do another ,load-system, then (foo) in the REPL /will/ return 50.0d0, as expected. How can I get C-M-x, C-c C-k and C-x C-e to use long-floats too? Am I doing something wrong that causes them to behave differently from ,load-system? I'm using SLIME and Emacs from CVS, and have set inferior-lisp-program and slime-multiprocessing to "sbcl" and t, respectively. Kind regards, Dirk Gerrits

4 13

[slime-devel] def* indent
by Travis Cross 11 Jan '05

11 Jan '05

I have a macro I use quite frequently that has a definition: (defmacro defpage (name title url &body body) ...) Even when I set lisp-prefix-match-indentation to nil (*see below), slime/emacs insists on indenting a call to defpage as (defpage test-page "Test Page" "/url/to" (div (div "the rest of the forms"))) If I rename defpage to something else, like 'mkpage', the indenting works as expected: (mkpage test-page "Test Page "/url/to" (div (div "the rest of the forms"))) I poked around in the slime source code, but I really wasn't able to make heads or tails of what is causing this unexpected (for me) behavior. If what I'm noticing is a 'feature,' what's the easiest way for me to work around this? Much appreciated, -- Travis (SBCL 0.8.15 / Slime CVS / GNU Emacs 21.3.1 / Gentoo) * though interestingly, when I use other test macros, it is evident that turning off lisp-prefix-match-indentation does have some effect. The macro (defmacro deftest (x y z) ...) will be indented: (deftest x y z) with lisp-prefix-match-indent on, but (deftest x y z) with the switch off. It seems that with lisp-prefix-match-indentation set to nil, slime/emacs completely ignores &body clues in the definition. If I redefine deftest as: (defmacro deftest (x y &body z) ...) it will still be indented incorrectly (as far as I can tell).

4 4

[slime-devel] Patch: slime-inspector-previous-inspectable-object
by Matthias Koeppe 11 Jan '05

11 Jan '05

The patch below adds a command `slime-inspector-previous-inspectable-object' and binds it to S-TAB in slime-inspector-mode. Moreover, the command `slime-inspector-next-inspectable-object' now accepts a prefix argument. The patch also makes wrapping around more reliable. I have adapted the code from the Emacs function `widget-move', which is used for similar functionality in Emacs Customize buffers. Index: slime.el =================================================================== RCS file: /project/slime/cvsroot/slime/slime.el,v retrieving revision 1.443 diff -u -p -u -r1.443 slime.el --- slime.el 16 Dec 2004 22:24:41 -0000 1.443 +++ slime.el 21 Dec 2004 14:55:34 -0000 @@ -6882,18 +6892,58 @@ Optionally set point to POINT." (set-window-configuration slime-saved-window-config) (kill-buffer (current-buffer))) -(defun slime-inspector-next-inspectable-object () - "sets the point to the next inspectable object" - (interactive) - (let ((pos (if (get-text-property (point) 'slime-part-number) - ;; we're in a part - (next-single-property-change - (or (next-single-property-change (point) 'slime-part-number) (point-min)) - 'slime-part-number) - ;; go to the next part or wrap around - (or (next-single-property-change (point) 'slime-part-number) - (next-single-property-change (point-min) 'slime-part-number))))) - (when pos (goto-char pos)))) +(defun slime-inspector-next-inspectable-object (arg) + "Move point to the next inspectable object. +With optional ARG, move across that many objects. +If ARG is negative, move backwards." + (interactive "p") + (or (bobp) (> arg 0) (backward-char)) + (let ((wrapped 0) + (number arg) + (old (get-text-property (point) 'slime-part-number)) + new) + ;; Forward. + (while (> arg 0) + (cond ((eobp) + (goto-char (point-min)) + (setq wrapped (1+ wrapped))) + (t + (goto-char (or (next-single-property-change (point) 'slime-part-number) + (point-max))))) + (and (= wrapped 2) + (eq arg number) + (error "No inspectable objects")) + (let ((new (get-text-property (point) 'slime-part-number))) + (when new + (unless (eq new old) + (setq arg (1- arg)) + (setq old new))))) + ;; Backward. + (while (< arg 0) + (cond ((bobp) + (goto-char (point-max)) + (setq wrapped (1+ wrapped))) + (t + (goto-char (or (previous-single-property-change (point) 'slime-part-number) + (point-min))))) + (and (= wrapped 2) + (eq arg number) + (error "No inspectable objects")) + (let ((new (get-text-property (point) 'slime-part-number))) + (when new + (unless (eq new old) + (setq arg (1+ arg)))))) + (let ((new (get-text-property (point) 'slime-part-number))) + (while (eq (get-text-property (point) 'slime-part-number) new) + (backward-char))) + (forward-char))) + +(defun slime-inspector-previous-inspectable-object (arg) + "Move point to the previous inspectable object. +With optional ARG, move across that many objects. +If ARG is negative, move forwards." + (interactive "p") + (slime-inspector-next-inspectable-object (- arg))) (defun slime-inspector-describe () (interactive) @@ -6909,6 +6959,7 @@ Optionally set point to POINT." ("d" 'slime-inspector-describe) ("q" 'slime-inspector-quit) ("\C-i" 'slime-inspector-next-inspectable-object) + ([(shift tab)] 'slime-inspector-previous-inspectable-object) ("\M-." 'slime-edit-definition)) -- Matthias Koeppe -- http://www.math.uni-magdeburg.de/~mkoeppe

2 2