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

[slime-devel] Bug in compiler note location resolving
by Hannu Koivisto 26 Jun '04

26 Jun '04

Greetings, With SLIME CVS HEAD as of 2004-03-28T23:15Z, Emacs 21.3 and sbcl 0.8.9.6, I noticed that an attempt to slime-load-system portable hemlock results to --------------------------------------------------------------- The value ", or \\" is not of type LIST. [Condition of type TYPE-ERROR] Restarts: 0: [RETRY] Retry performing #<ASDF:COMPILE-OP NIL {A203339}> on #<HEMLOCK-SYSTEM::HEMLOCK-CL-SOURCE-FILE "main" {98D2271}>. 1: [ACCEPT] Continue, treating #<ASDF:COMPILE-OP NIL {A203339}> on #<HEMLOCK-SYSTEM::HEMLOCK-CL-SOURCE-FILE "main" {98D2271}> as having been successful. 2: [ABORT] Return to SLIME toplevel. 3: [ABORT] Reduce debugger level (leaving debugger, returning to toplevel). 4: [TOPLEVEL] Restart at toplevel READ/EVAL/PRINT loop. Backtrace: 0: (NTHCDR 2 2 ", or \\")[:EXTERNAL] 1: (SWANK-BACKEND::SOURCE-PATH-SOURCE-POSITION (0 5 2 2) (NIL NIL (NIL NIL NIL (NIL NIL) (NIL (NIL "DISPLAY"))) "Invokes the editor, Hemlock. If X is supplied and is a symbol, the definition of X is put into a buffer, and that buffer is selected. If X is a pathname, the file specified by X is visited in a new buffer. If X is not supplied or Nil, the editor is entered in the same state as when last exited. When :init is supplied as t (the default), the file \\" NIL ", or \\" NIL " is loaded from the home directory, but the Lisp command line switch -hinit can be used to specify a different name. Any compiled version of the source is preferred when choosing the file to load. If the argument is non-nil and not t, then it should be a pathname that will be merged with the home directory." (NIL NIL (NIL "You are already in the editor, you bogon!")) (NIL ((NIL NIL) (NIL (NIL NIL (NIL NIL) (NIL NIL)))) (NIL (QUOTE NIL) (NIL (NIL NIL (NIL NIL) (NIL NIL NIL) (NIL (NIL NIL))) (NIL (QUOTE NIL) (NIL (QUOTE NIL) (NIL ((NIL NIL (NIL NIL)) (NIL ((NIL (NIL (NIL (QUOTE NIL) "Edit " (NIL NIL)))) (NIL (NIL (NIL NIL NIL) (NIL NIL))) (NIL NIL)) (NIL (NIL NIL)) (NIL (NIL (NIL NIL)) (NIL (QUOTE (NIL NIL))) (NIL) (NIL NIL) (NIL (NIL NIL))))) ((NIL (NIL NIL) (NIL NIL)) (NIL NIL NIL)) (NIL (NIL "~S is not a symbol or pathname. I can't edit it!" NIL)))) (NIL NIL) (NIL (NIL (NIL (QUOTE NIL) (NIL ((NIL (FUNCTION (NIL (NIL) (NIL NIL NIL))))) (NIL NIL) (NIL)))) (NIL NIL))))))) #<HASH-TABLE :TEST EQ :COUNT 85 {A9BC909}>) 2: (SWANK-BACKEND::SOURCE-PATH-STREAM-POSITION (20 5 2 2) #<FILE-STREAM for "file \"/home/azure/projects/software/hemlock/src/main.lisp\"" {AF82DA9}>) 3: (SWANK-BACKEND::SOURCE-PATH-FILE-POSITION (20 5 2 2) #P"/home/azure/projects/software/hemlock/src/main.lisp") 4: ((SWANK-BACKEND::RESOLVE-NOTE-LOCATION ((EQL NIL) PATHNAME T T T)) #<unused argument> #<unused argument> #<unused argument> #P"/home/azure/projects/software/hemlock/src/main.lisp" #<unused argument> (20 5 2 2) #<unused argument>) 5: (SWANK-BACKEND::COMPILER-NOTE-LOCATION #S(SB-C::COMPILER-ERROR-CONTEXT :ENCLOSING-SOURCE ("UNWIND-PROTECT" "FLET" "BLOCK" "MULTIPLE-VALUE-BIND" "MULTIPLE-VALUE-CALL" "BLOCK" "SB-C::%WITHIN-CLEANUP" "RETURN-FROM" "PROGN") :SOURCE (" (LET ((HEMLOCK-INTERNALS:*BEEP-FUNCTION* #'HEMLOCK-INTERNALS::HEMLOCK-BEEP) (HEMLOCK-INTERNALS::*GC-NOTIFY-BEFORE* #'HEMLOCK-INTERNALS::HEMLOCK-GC-NOTIFY-BEFORE) (HEMLOCK-INTERNALS::*GC-NOTIFY-AFTER* #'HEMLOCK-INTERNALS::HEMLOCK-GC-NOTIFY-AFTER) (*STANDARD-INPUT* HEMLOCK-INTERNALS::*ILLEGAL-READ-STREAM*) (*QUERY-IO* HEMLOCK-INTERNALS::*ILLEGAL-READ-STREAM*)) (COND ((NOT HEMLOCK-INTERNALS::*EDITOR-WINDOWED-INPUT*) (UNLESS HEMLOCK-INTERNALS::*EDITOR-HAS-BEEN-ENTERED* # # #) (CATCH 'HEMLOCK-INTERNALS::HEMLOCK-EXIT # # #)) (T (HEMLOCK-EXT:WITH-CLX-EVENT-HANDLING # # #))))") :ORIGINAL-SOURCE " (HEMLOCK-INTERNALS::SITE-WRAPPER-MACRO (UNLESS HEMLOCK-INTERNALS::*EDITOR-HAS-BEEN-ENTERED* (HEMLOCK-INTERNALS::%INIT-REDISPLAY HEMLOCK-INTERNALS::DISPLAY) (SETQ HEMLOCK-INTERNALS::*EDITOR-HAS-BEEN-ENTERED* T) (HEMLOCK-INTERFACE:INVOKE-HOOK (REVERSE HEMLOCK-INTERNALS::*AFTER-EDITOR-INITIALIZATIONS-FUNS*))) (CATCH 'HEMLOCK-INTERNALS::HEMLOCK-EXIT (CATCH 'HEMLOCK-INTERNALS::EDITOR-TOP-LEVEL-CATCHER (COND # # #)) (HEMLOCK-INTERFACE:INVOKE-HOOK HEMLOCK::ENTRY-HOOK) (UNWIND-PROTECT (LOOP #) (HEMLOCK-INTERFACE:INVOKE-HOOK HEMLOCK::EXIT-HOOK))))" :CONTEXT ((DEFUN HEMLOCK)) :FILE-NAME #P"/home/azure/projects/software/hemlock/src/main.lisp" :FILE-POSITION 9218 :ORIGINAL-SOURCE-PATH (2 2 5 20))) 6: (SWANK-BACKEND::SIGNAL-COMPILER-CONDITION #<SB-INT:SIMPLE-STYLE-WARNING {AF79091}> #S(SB-C::COMPILER-ERROR-CONTEXT :ENCLOSING-SOURCE ("UNWIND-PROTECT" "FLET" "BLOCK" "MULTIPLE-VALUE-BIND" "MULTIPLE-VALUE-CALL" "BLOCK" "SB-C::%WITHIN-CLEANUP" "RETURN-FROM" "PROGN") :SOURCE (" (LET ((HEMLOCK-INTERNALS:*BEEP-FUNCTION* #'HEMLOCK-INTERNALS::HEMLOCK-BEEP) (HEMLOCK-INTERNALS::*GC-NOTIFY-BEFORE* #'HEMLOCK-INTERNALS::HEMLOCK-GC-NOTIFY-BEFORE) (HEMLOCK-INTERNALS::*GC-NOTIFY-AFTER* #'HEMLOCK-INTERNALS::HEMLOCK-GC-NOTIFY-AFTER) (*STANDARD-INPUT* HEMLOCK-INTERNALS::*ILLEGAL-READ-STREAM*) (*QUERY-IO* HEMLOCK-INTERNALS::*ILLEGAL-READ-STREAM*)) (COND ((NOT HEMLOCK-INTERNALS::*EDITOR-WINDOWED-INPUT*) (UNLESS HEMLOCK-INTERNALS::*EDITOR-HAS-BEEN-ENTERED* # # #) (CATCH 'HEMLOCK-INTERNALS::HEMLOCK-EXIT # # #)) (T (HEMLOCK-EXT:WITH-CLX-EVENT-HANDLING # # #))))") :ORIGINAL-SOURCE " (HEMLOCK-INTERNALS::SITE-WRAPPER-MACRO (UNLESS HEMLOCK-INTERNALS::*EDITOR-HAS-BEEN-ENTERED* (HEMLOCK-INTERNALS::%INIT-REDISPLAY HEMLOCK-INTERNALS::DISPLAY) (SETQ HEMLOCK-INTERNALS::*EDITOR-HAS-BEEN-ENTERED* T) (HEMLOCK-INTERFACE:INVOKE-HOOK (REVERSE HEMLOCK-INTERNALS::*AFTER-EDITOR-INITIALIZATIONS-FUNS*))) (CATCH 'HEMLOCK-INTERNALS::HEMLOCK-EXIT (CATCH 'HEMLOCK-INTERNALS::EDITOR-TOP-LEVEL-CATCHER (COND # # #)) (HEMLOCK-INTERFACE:INVOKE-HOOK HEMLOCK::ENTRY-HOOK) (UNWIND-PROTECT (LOOP #) (HEMLOCK-INTERFACE:INVOKE-HOOK HEMLOCK::EXIT-HOOK))))" :CONTEXT ((DEFUN HEMLOCK)) :FILE-NAME #P"/home/azure/projects/software/hemlock/src/main.lisp" :FILE-POSITION 9218 :ORIGINAL-SOURCE-PATH (2 2 5 20))) 7: (SWANK-BACKEND::HANDLE-NOTIFICATION-CONDITION #<SB-INT:SIMPLE-STYLE-WARNING {AF79091}>) 8: (SIGNAL 1 #<SB-INT:SIMPLE-STYLE-WARNING {AF79091}>)[:EXTERNAL] 9: (SB-C::COMPILER-STYLE-WARNING-HANDLER 1 #<SB-INT:SIMPLE-STYLE-WARNING {AF79091}>)[:EXTERNAL] 10: (SIGNAL 1 #<SB-INT:SIMPLE-STYLE-WARNING {AF79091}>)[:EXTERNAL] 11: (WARN 5 SB-INT:SIMPLE-STYLE-WARNING)[:EXTERNAL] 12: (SB-C::NOTE-LEXICAL-BINDING 1 HEMLOCK-INTERNALS::*GC-NOTIFY-BEFORE*)[:EXTERNAL] . . . --------------------------------------------------------------- It seems that SLIME has read the docstring as several forms even though it really is just one form. You can reproduce the problem by putting the following code to a file and then invoking slime-compile-and-load-file on it. --------------------------------------------------------------- (defun cl-user::foo (&optional x &key (init t) (display "")) "Invokes the editor, Hemlock. If X is supplied and is a symbol, the definition of X is put into a buffer, and that buffer is selected. If X is a pathname, the file specified by X is visited in a new buffer. If X is not supplied or Nil, the editor is entered in the same state as when last exited. When :init is supplied as t (the default), the file \"hemlock-init.lisp\", or \".hemlock-init.lisp\" is loaded from the home directory, but the Lisp command line switch -hinit can be used to specify a different name. Any compiled version of the source is preferred when choosing the file to load. If the argument is non-nil and not t, then it should be a pathname that will be merged with the home directory." (when *in-the-editor* (error "You are already in the editor, you bogon!")) (let ((*in-the-editor* t) (display (unless *editor-has-been-entered* (maybe-load-hemlock-init init) ;; Device dependent initializaiton. (init-raw-io display)))) (catch 'editor-top-level-catcher (site-wrapper-macro (unless *editor-has-been-entered* ;; Make an initial window, and set up redisplay's internal ;; data structures. (%init-redisplay display) (setq *editor-has-been-entered* t) ;; Pick up user initializations to be done after initialization. (invoke-hook (reverse *after-editor-initializations-funs*))) (catch 'hemlock-exit (catch 'editor-top-level-catcher (cond ((and x (symbolp x)) (let* ((name (nstring-capitalize (concatenate 'simple-string "Edit " (string x)))) (buffer (or (getstring name *buffer-names*) (make-buffer name))) (*print-case* :downcase)) (delete-region (buffer-region buffer)) (with-output-to-mark (*standard-output* (buffer-point buffer)) (eval `(grindef ,x)) ; hackish, I know... (terpri) (change-to-buffer buffer) (buffer-start (buffer-point buffer))))) ((or (stringp x) (pathnamep x)) (find-file-command () x)) (x (error "~S is not a symbol or pathname. I can't edit it!" x)))) (invoke-hook entry-hook) (unwind-protect (loop (catch 'editor-top-level-catcher (handler-bind ((error #'(lambda (condition) (lisp-error-error-handler condition :internal)))) (invoke-hook abort-hook) (%command-loop)))) (invoke-hook exit-hook))))))) --------------------------------------------------------------- -- Hannu

4 4

[slime-devel] Patch for swank-allegro.lisp
by Bill_Clementson＠peoplesoft.com 01 Apr '04

01 Apr '04

Here is a patch - misplaced parens in the latest cvs version of swank-allegro.lisp. Bill Clementson *** c:/DOCUME~1/BC1886~1/LOCALS~1/Temp/swank-allegro.lisp Wed Mar 31 15:17:45 2004 --- c:/DOCUME~1/BC1886~1/LOCALS~1/Temp/swank-allegro-orig.lisp Wed Mar 31 15:17:45 2004 *************** *** 311,317 **** "yielding before sending" 0.1 (lambda () (mp:with-process-lock (mutex) ! (< (length (mailbox.queue mbox)) 10)))) (mp:with-process-lock (mutex) (setf (mailbox.queue mbox) (nconc (mailbox.queue mbox) (list message)))))) --- 311,317 ---- "yielding before sending" 0.1 (lambda () (mp:with-process-lock (mutex) ! (< (length (mailbox.queue mbox) 10))))) (mp:with-process-lock (mutex) (setf (mailbox.queue mbox) (nconc (mailbox.queue mbox) (list message))))))

2 1

[slime-devel] Daily ChangeLog diff
by Luke Gorrie 31 Mar '04

31 Mar '04

Index: slime/ChangeLog diff -u slime/ChangeLog:1.316 slime/ChangeLog:1.317 --- slime/ChangeLog:1.316 Mon Mar 29 12:08:17 2004 +++ slime/ChangeLog Tue Mar 30 18:12:43 2004 @@ -1,3 +1,28 @@ +2004-03-30 Helmut Eller <e9626484(a)stud3.tuwien.ac.at> + + * slime.el (slime-repl-result-face): New face. + (slime-inspector-mode-map): Add a binding for M-. + (compile-defun): Add test case for escaped double quotes inside a + string. + + * swank.lisp (ed-in-emacs): New allowed form for argument. + (pprint-eval-string-in-frame): Apply arguments in proper order. + + * swank-cmucl.lisp (method-dspec): Include the specializers. + (class-definitions): Renamed from struct-definitions. Try to + locate condition-classes and PCL classes (in the future). + (debug-function-arglist): Insert &optional, &key, &rest in the + right places. + (form-number-stream-position): Make it a separate function. + +2004-03-29 Lawrence Mitchell <wence(a)gmx.li> + + * swank.lisp (ed-in-emacs): New allowed form for argument. + + * slime.el (slime-ed): Deal with list form of argument. For a + list (FILENAME LINE [COLUMN]), visit the correct line and column + number. + 2004-03-29 Helmut Eller <e9626484(a)stud3.tuwien.ac.at> * swank-source-path-parser.lisp (cmucl-style-get-macro-character):

1 0

[slime-devel] New argument to ED-IN-EMACS.
by Lawrence Mitchell 31 Mar '04

31 Mar '04

While ANSI says that ED should only accept one optional argument, it'd be nice if SLIME's ED-IN-EMACS could deal with another, optional, position argument, making it more useful as a general editing tool, as opposed to being quite lisp specific. What are people's thoughts on something like this: Index: ChangeLog =================================================================== RCS file: /project/slime/cvsroot/slime/ChangeLog,v retrieving revision 1.316 diff -u -r1.316 ChangeLog --- ChangeLog 29 Mar 2004 17:08:17 -0000 1.316 +++ ChangeLog 29 Mar 2004 22:02:19 -0000 @@ -1,3 +1,11 @@ +2004-03-29 Lawrence Mitchell <wence(a)gmx.li> + + * swank.lisp (ed-in-emacs): New optional argument, WHERE. + Describes a LINE:COLUMN location in the file we're visiting. + + * slime.el (slime-ed): Deal with new WHERE argument. In the case + of a filename, visit the correct line and column number. + 2004-03-29 Helmut Eller <e9626484(a)stud3.tuwien.ac.at> * swank-source-path-parser.lisp (cmucl-style-get-macro-character): Index: swank.lisp =================================================================== RCS file: /project/slime/cvsroot/slime/swank.lisp,v retrieving revision 1.155 diff -u -r1.155 swank.lisp --- swank.lisp 27 Mar 2004 21:14:52 -0000 1.155 +++ swank.lisp 29 Mar 2004 21:56:36 -0000 @@ -1133,12 +1133,16 @@ (let ((*package* *buffer-package*)) (format nil "~{~S~^~%~}" values)))))) -(defslimefun ed-in-emacs (&optional what) +(defslimefun ed-in-emacs (&optional what where) "Edit WHAT in Emacs. -WHAT can be a filename (pathname or string) or function name (symbol)." +WHAT can be a filename (pathname or string) or function name (symbol). + +If WHERE is non-nil, it should be a string of the form +\"LINE[:COLUMN]\" telling us where in the file we should jump to." (send-oob-to-emacs `(:ed ,(if (pathnamep what) (canonicalize-filename what) - what)))) + what) + ,where))) ;;;; Compilation Commands. Index: slime.el =================================================================== RCS file: /project/slime/cvsroot/slime/slime.el,v retrieving revision 1.248 diff -u -r1.248 slime.el --- slime.el 29 Mar 2004 00:59:13 -0000 1.248 +++ slime.el 29 Mar 2004 21:59:26 -0000 @@ -1511,8 +1511,8 @@ (setf (slime-use-sigint-for-interrupt) t)) ((:%apply fn args) (apply (intern fn) args)) - ((:ed what) - (slime-ed what)) + ((:ed what where) + (slime-ed what where)) ((:debug-condition thread message) (apply 'ignore thread) ; stupid XEmacs warns about unused variable (message "%s" message))))) @@ -3637,7 +3637,7 @@ (defvar slime-ed-use-dedicated-frame t "*When non-nil, `slime-ed' will create and reuse a dedicated frame.") -(defun slime-ed (what) +(defun slime-ed (what where) "Edit WHAT, either a filename (string) or function name (symbol), or nil. This for use in the implementation of COMMON-LISP:ED." ;; Without `save-excursion' very strange things happen if you call @@ -3648,7 +3648,19 @@ (setq slime-ed-frame (new-frame))) (select-frame slime-ed-frame)) (cond ((stringp what) - (find-file (slime-from-lisp-filename what))) + (find-file (slime-from-lisp-filename what)) + (when (string-match "\\([0-9]+\\):?\\([0-9]+\\)?" where) + ;; Go to the correct line. + (goto-line (string-to-number (match-string 1 where))) + ;; Check to see if we've specified a column, and if so, + ;; try and reach it, but don't go off the end of the + ;; current line. + (let ((col (match-string 2 where))) + (and col (setq col (string-to-number col))) + (while (and col + (< (point) (point-at-eol)) + (/= (decf col) -1)) + (forward-char 1))))) ((and what (symbolp what)) (slime-edit-definition (symbol-name what))) (t nil)))) ; nothing in particular -- Lawrence Mitchell <wence(a)gmx.li>

2 2

[slime-devel] Daily ChangeLog diff
by Luke Gorrie 30 Mar '04

30 Mar '04

Index: slime/ChangeLog diff -u slime/ChangeLog:1.315 slime/ChangeLog:1.316 --- slime/ChangeLog:1.315 Sun Mar 28 19:59:58 2004 +++ slime/ChangeLog Mon Mar 29 12:08:17 2004 @@ -1,3 +1,9 @@ +2004-03-29 Helmut Eller <e9626484(a)stud3.tuwien.ac.at> + + * swank-source-path-parser.lisp (cmucl-style-get-macro-character): + New function. Workaround for bug(?) in SBCL. + (make-source-recording-readtable): Use it. + 2004-03-29 Luke Gorrie <luke(a)bluetail.com> * HACKING: Some small updates (more needed). @@ -27,7 +33,7 @@ * swank-lispworks.lisp (spawn): Remove CL symbols from mp:*process-initial-bindings*, to avoid the irritating behavior - for request execute in different threads. E.g., when someone + for requests executed in different threads. E.g., when someone tries to set *package*. * swank.lisp (*log-io*): New variable. Bind it to *terminal-io*

1 0

[slime-devel] Daily ChangeLog diff
by Luke Gorrie 29 Mar '04

29 Mar '04

Index: slime/ChangeLog diff -u slime/ChangeLog:1.314 slime/ChangeLog:1.315 --- slime/ChangeLog:1.314 Sat Mar 27 16:39:00 2004 +++ slime/ChangeLog Sun Mar 28 19:59:58 2004 @@ -1,3 +1,21 @@ +2004-03-29 Luke Gorrie <luke(a)bluetail.com> + + * HACKING: Some small updates (more needed). + + * slime.el (slime-inspector-buffer): Enter `slime-inspector-mode' + after `slime-mode'. This seems to give priority of keymap to the + inspector, so that it can override SPC. + (slime-easy-menu): Add slime-switch-to-output-buffer. + Enable SLIME menu in the REPL buffer. + (slime-symbol-name-at-point): Avoid mistaking the REPL prompt for + a symbol. + (slime-words-of-encouragement): A few new ones. + (slime-insert-xrefs): Removed the final newline from XREF + buffers. This helps to avoid unwanted scrolling. + + * doc/slime.texi: Added a section about user-interface + conventions and our relationship with inf-lisp. + 2004-03-27 Helmut Eller <e9626484(a)stud3.tuwien.ac.at> * slime.el (slime-changelog-date): Reinitialize it at load-time. @@ -36,6 +54,9 @@ * slime.el (slime-group-similar): Bugfix: return NIL if the input list is NIL. + (slime-inspector-buffer): Enter `slime-inspector-mode' after + `slime-mode'. This seems to give priority of keymap to the + inspector, so that it can override SPC. 2004-03-26 Bjørn Nordbø <bn(a)telenor.net>

1 0

[slime-devel] [PATCH] arglist printing in LW 4.1
by Bjørn Nordbø 29 Mar '04

29 Mar '04

This fixes arglist printing in Lispworks 4.1, and otherwise it should be harmless. I also added a comment in ChaneLog. Is this correct? -- Lisp can accomodate a wide range of psychological problems.

3 7

[slime-devel] Tab behavior in REPL buffers
by Peter Seibel 29 Mar '04

29 Mar '04

I haven't looked deeply into this but I was wondering if anyone else found the behavior of TAB in REPL buffers a bit annoying. I'm so used to using TAB to indent code that I'm constantly hitting it to try to line up my code in the REPL. Which usually results in a huge *Completions* typeout window popping up with a list of every symbol in existance because I'm usually sitting on some whitespace. I notice in lisp source buffers TAB is *not* bound to slime-complete-symbol. Should I just remap TAB myself or does anyone else find the current behavior annoying and perhaps worth changing? -Peter -- Peter Seibel peter(a)javamonkey.com Lisp is the red pill. -- John Fraser, comp.lang.lisp

4 4

Re: [slime-devel] Re: compile-log-show buffer
by Julian Stecklina 28 Mar '04

28 Mar '04

Luke Gorrie <luke(a)bluetail.com> writes: > Helmut Eller <e9626484(a)stud3.tuwien.ac.at> writes: > >> We byte-compile some functions when slime is loaded and some Emacs >> versions display the log buffer. Not sure what the right fix is. We >> could rewrite the functions so that they don't produce compiler >> messages (might be difficult) or perhaps there are variables to >> suppress the log buffer. Suggestions welcome. > > Fixed. I wrapped the `byte-compile' calls in `save-window-excursion'. Thanks. I'll try that. Regards, -- Julian Stecklina Signed and encrypted mail welcome. Key-Server: pgp.mit.edu Key-ID: 0xD65B2AB5 FA38 DCD3 00EC 97B8 6DD8 D7CC 35D8 8D0E D65B 2AB5 Any sufficiently complicated C or Fortran program contains an ad hoc informally-specified bug-ridden slow implementation of half of Common Lisp. - Greenspun's Tenth Rule of Programming

1 0

[slime-devel] Daily ChangeLog diff
by Luke Gorrie 28 Mar '04

28 Mar '04

Index: slime/ChangeLog diff -u slime/ChangeLog:1.312 slime/ChangeLog:1.314 --- slime/ChangeLog:1.312 Fri Mar 26 11:07:08 2004 +++ slime/ChangeLog Sat Mar 27 16:39:00 2004 @@ -1,3 +1,37 @@ +2004-03-27 Helmut Eller <e9626484(a)stud3.tuwien.ac.at> + + * slime.el (slime-changelog-date): Reinitialize it at load-time. + This avoids the need to restart Emacs (horror!) after an update. + + * swank-cmucl.lisp (debug-function-arglist): Properly reconstruct + the arglist from the debug-info. (Not complete yet.) + (arglist): Use it. + + * swank-lispworks.lisp (spawn): Remove CL symbols from + mp:*process-initial-bindings*, to avoid the irritating behavior + for request execute in different threads. E.g., when someone + tries to set *package*. + + * swank.lisp (*log-io*): New variable. Bind it to *terminal-io* + at load-time, so we can log to a non-redirected stream. + (disassemble-symbol): Allow generalized function names. + (apropos-symbols): Handle the PACKAGE argument properly to get + useful output for C-c P. + + * slime.el (slime-repl-indent-and-complete-symbol): New command. + Bound to TAB in the REPL mode. First try to indent the current + line then try to complete the symbol at point. + (slime-dispatch-event): Ignore a unused thread variable to keep + XEmacs' byte compiler quiet. + + * swank-sbcl.lisp (enable-sigio-on-fd): Use sb-posix::fcntl + instead of sb-posix:fcntl to avoid the ugly reader hack. SBCL + doesn't have package locks and even if they add locks in the + future sb-posix::fcntl will still be valid. + (getpid): Use defimplementation instead of defmethod. + (function-definitions): Take generalized function names ala '(setf + car)' as argument. + 2004-03-26 Luke Gorrie <luke(a)bluetail.com> * slime.el (slime-group-similar): Bugfix: return NIL if the input

1 0

[slime-devel] Updating SLIME
by Jan Rychter 27 Mar '04

27 Mar '04

How do you guys update SLIME? It seems to me that since slime-changelog-date is a defvar, it doesn't change when reloading (or re-evaluating) slime.el. So when the protocol changes, emacs still thinks it speaks the old version. Do you actually *gasp* quit your XEmacs or Emacs from time to time? --J. PS: oh, the irony. As I was writing the first incarnation of this E-mail, I have inserted the USB clipdrive which hosts my GPG key. That caused the machine to freeze instantly. USB in Linux is decidedly the crappiest piece of software that I have to live with.

2 1

[slime-devel] [PATCH] bugfix for SBCL
by Bjørn Nordbø 27 Mar '04

27 Mar '04

This patch fixes a bug in swank-sbcl.lisp. -- Lisp can accomodate a wide range of psychological problems.

2 1

[slime-devel] can't change *readtable* from repl
by Alan Ruttenberg 27 Mar '04

27 Mar '04

CL-USER> *readtable* #<READTABLE #x5322ECE> CL-USER> (copy-readtable) #<READTABLE #xD70F94E> CL-USER> (setq *readtable* *) #<READTABLE #xD70F94E> CL-USER> *readtable* #<READTABLE #x5322ECE> Noticed this because cl-interpol enabling function stopped working at some point. -Alan

2 2

[slime-devel] Re: [slime-cvs] CVS update: slime/HACKING
by Luke Gorrie 26 Mar '04

26 Mar '04

Helmut Eller <heller(a)common-lisp.net> writes: > +If you would like to send us improvements you can create a patch with > +C-x v = in the buffer or with 'diff -c'. You can also include a > +ChangeLog entry describing your change. Hey, I can't read 'diff -c'! But I'm willing to learn if you can cite some virtues :-)

5 5

[slime-devel] CVS broken: destructure-case failed
by Bjørn Nordbø 26 Mar '04

26 Mar '04

After updating to the current CVS version, I get the following error on slime-compile-defun. I have tried to find the error, but it seems to be in slime.el and I'm no good at elisp. - The error have probably been introduced in a patch during the last 24 hours or so. - It shows itself consistently on my setup; GNU Emacs 21.2.1, Lispworks 4.[13], Solaris. - Using the slime.el from yesterday (2004-03-25) works fine. This is the error message I get in Emacs: Compilation finished: 0 errors 0 warnings 0 notes [0.12 secs] error in process filter: cond: destructure-case failed: nil error in process filter: destructure-case failed: nil -- Lisp can accomodate a wide range of psychological problems.

2 1

[slime-devel] compile-log-show buffer
by Julian Stecklina 26 Mar '04

26 Mar '04

Hello, if I do (require 'slime) in XEmacs a *Compile-Log-Show* buffer opens. This is very annoying. At the moment I use (delete-other-window) in my init.el, but how can I find the real problem? Regards, -- Julian Stecklina Signed and encrypted mail welcome. Key-Server: pgp.mit.edu Key-ID: 0xD65B2AB5 FA38 DCD3 00EC 97B8 6DD8 D7CC 35D8 8D0E D65B 2AB5 Any sufficiently complicated C or Fortran program contains an ad hoc informally-specified bug-ridden slow implementation of half of Common Lisp. - Greenspun's Tenth Rule of Programming

3 2

Re: [slime-devel] compile-log-show buffer
by Julian Stecklina 26 Mar '04

26 Mar '04

Helmut Eller <e9626484(a)stud3.tuwien.ac.at> writes: > Julian Stecklina <der_julian(a)web.de> writes: > >> if I do (require 'slime) in XEmacs a *Compile-Log-Show* buffer >> opens. This is very annoying. At the moment I use >> (delete-other-window) in my init.el, but how can I find the real >> problem? > > We byte-compile some functions when slime is loaded and some Emacs > versions display the log buffer. Not sure what the right fix is. We > could rewrite the functions so that they don't produce compiler > messages (might be difficult) or perhaps there are variables to > suppress the log buffer. Suggestions welcome. The strange thing is, that this log buffer is empty... Regards, -- Julian Stecklina Signed and encrypted mail welcome. Key-Server: pgp.mit.edu Key-ID: 0xD65B2AB5 FA38 DCD3 00EC 97B8 6DD8 D7CC 35D8 8D0E D65B 2AB5 Any sufficiently complicated C or Fortran program contains an ad hoc informally-specified bug-ridden slow implementation of half of Common Lisp. - Greenspun's Tenth Rule of Programming

1 0

[slime-devel] Daily ChangeLog diff
by Luke Gorrie 26 Mar '04

26 Mar '04

Index: slime/ChangeLog diff -u slime/ChangeLog:1.304 slime/ChangeLog:1.308 --- slime/ChangeLog:1.304 Wed Mar 24 07:51:27 2004 +++ slime/ChangeLog Thu Mar 25 19:48:04 2004 @@ -1,5 +1,28 @@ +2004-03-26 Luke Gorrie <luke(a)bluetail.com> + + * slime.el (slime-merge-notes-for-display): New function to merge + together compiler notes that refer to the same location. This is + an optimization for when there are a lot of compiler notes: + `slime-merge-note-into-overlay' concat'd messages together one by + one in O(n^2) time/space, and became noticeably slow in practice + with ~100 notes or more. + (slime-tree-insert): This function is now automatically + byte-compiled (good speed gain). + +2004-03-25 Bjørn Nordbø <bn(a)telenor.net> + + * swank-lispworks.lisp: (create-socket, set-sigint-handler) + (who-references, who-binds, who-sets): Add backward compatibility + for LW 4.1. + (dspec-buffer-position): Fix inappropriate use of etypecase. + 2004-03-24 Luke Gorrie <luke(a)bluetail.com> + * swank-sbcl.lisp (getpid): Use sb-posix:getpid. + + * slime.el (slime-inspector-mode-map): Added SPC as extra binding + for slime-inspector-next (like info-mode). + * doc/slime.texi: Added completion style and configuration. 2004-03-23 Alan Shutko <ats(a)acm.org> @@ -38,7 +61,7 @@ New functions. (with-connection): Fix quoting bug and move upwards before first use. - (send-output-to-emacs): Add kludge for SBCL !-package names. + (guess-package-from-string): Add kludge for SBCL !-package names. (apropos-list-for-emacs): Lispworks apparently returns duplicates; remove them. (inspect-object): Princ the label to allow strings and symbols.

1 0

Re: [slime-devel] possible bug in SLIME/Lispworks
by Bjørn Nordbø 26 Mar '04

26 Mar '04

Helmut Eller wrote: > Bjørn Nordbø <bn(a)telenor.net> writes: > > > > How do I set a breakpoint for SLDB in a source fragment (ie. what > > do I use instead of break)? > > You could use INVOKE-DEBUGGER, which hopefully doesn't rebind the > debugger hook :-). Alternatively you can add a method to > ENV-INTERNALS:ENVIRONMENT-DISPLAY-NOTIFIER. E.g. you could add Another solution, I found, was to call (SWANK::SIMPLE-BREAK). As it is only used for debugging, I probably won't be condemned to heck for doing this.. However, I noticed your patch in the ChangeLog and now BREAK does what I want. Thank you very much! :-) > (defmethod env-internals:environment-display-notifier > (env &key restarts condition) > (swank::swank-debugger-hook condition nil)) > > to your ~/.swank.lisp. I added something like this in the CVS > version. It is also possible to override the Lispworks window > debugger by specializing the env argument. That is probably cool, but I wouldn't know as I have never used the Lispworks GUI environment. Bjørn

3 3

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

25 Mar '04

Index: slime/ChangeLog diff -u slime/ChangeLog:1.303 slime/ChangeLog:1.304 --- slime/ChangeLog:1.303 Tue Mar 23 16:38:48 2004 +++ slime/ChangeLog Wed Mar 24 07:51:27 2004 @@ -1,3 +1,7 @@ +2004-03-24 Luke Gorrie <luke(a)bluetail.com> + + * doc/slime.texi: Added completion style and configuration. + 2004-03-23 Alan Shutko <ats(a)acm.org> * swank-clisp.lisp (set-default-directory): New function.

1 0

[slime-devel] Manual labour
by Luke Gorrie 24 Mar '04

24 Mar '04

[Repost: I'm not sure that my via-Gmane post is getting through.] I've drafted a user manual in doc/slime.texi. It's also online at http://www.bluetail.com/~luke/misc/lisp/slime.ps.gz http://www.bluetail.com/~luke/misc/lisp/slime.pdf The PDF version is a bit ugly (from ps2pdf), because texi2pdf is choking on slime.texi for some reason. Feedback would be welcome, so please RTFM! Cheers, Luke

4 7

[slime-devel] Tab completion bug?
by Dirk Gerrits 24 Mar '04

24 Mar '04

Typing M, U, L, <TAB> in the SLIME REPL produces "multiple-value|-" where | represents the cursor/point. Is this a bug? If it'd produce "multiple-value-|" then "multiple-value-bind" and such would be just a tad easier to type. Regards, Dirk Gerrits

4 9

[slime-devel] getpid implementation on sbcl
by Christophe Rhodes 24 Mar '04

24 Mar '04

May be of interest. (Mostly because it removes the warning at startup, but also using the sb-posix supported function rather than the sb-unix internal one. Cheers, Christophe -- http://www-jcsu.jesus.cam.ac.uk/~csr21/ +44 1223 510 299/+44 7729 383 757 (set-pprint-dispatch 'number (lambda (s o) (declare (special b)) (format s b))) (defvar b "~&Just another Lisp hacker~%") (pprint #36rJesusCollegeCambridge)

1 0

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

24 Mar '04

Index: slime/ChangeLog diff -u slime/ChangeLog:1.302 slime/ChangeLog:1.303 --- slime/ChangeLog:1.302 Mon Mar 22 11:35:57 2004 +++ slime/ChangeLog Tue Mar 23 16:38:48 2004 @@ -1,3 +1,56 @@ +2004-03-23 Alan Shutko <ats(a)acm.org> + + * swank-clisp.lisp (set-default-directory): New function. + +2004-03-23 Helmut Eller <e9626484(a)stud3.tuwien.ac.at> + + * swank-allegro.lisp (send): Wait a bit if there are already many + message in the mailbox. + + * swank-clisp.lisp (xref-results): Use fspec-location instead of + the of fspec-source-locations. Reported by Alan Shutko. + (break): Be friendly to case-inverting readtables. + + * swank-lispworks.lisp (emacs-connected): Add default method to + environment-display-notifier. Reported by Bjørn Nordbø. + (set-default-directory, who-specializes): Implemented for + Lispworks. + (gfp): New function. + (describe-symbol-for-emacs, describe-definition): Distinguish + between ordinary and generic functions. + (call-with-debugging-environment): Unwind a few frames. Looks + better and avoids the problems with the real topframe. + (interesting-frame-p): Use Lispworks dbg:*print-xxx* variables to + decide which frames are interesting. + (frame-actual-args): New function. + (print-frame): Use it. + + * swank.lisp (open-streams, make-output-function): Capture the + connection not only the socket. This way the streams can be used + from unrelated threads. Reported by Alain Picard. + (create-connection): Factorized. Initialize the streams after the + connection is created. + (initialize-streams-for-connection, spawn-threads-for-connection): + New functions. + (with-connection): Fix quoting bug and move upwards before first + use. + (send-output-to-emacs): Add kludge for SBCL !-package names. + (apropos-list-for-emacs): Lispworks apparently returns duplicates; + remove them. + (inspect-object): Princ the label to allow strings and symbols. + (send-output-to-emacs): Deleted. + (defslimefun-unimplemented): Deleted. Was unused. + + * slime.el (slime-easy-menu): Add some more commands. + (slime-changelog-date): New variable. Initialized with the value + returned by the function of the same name. This detects + incompatible versions if Emacs has not been restarted after an + upgrade. + (slime-check-protocol-version, slime-init-output-buffer): Use it. + (slime-events-buffer, slime-log-event): Use fundamental mode + instead of lisp-mode to avoid excessive font-locking for messages + with lots of strings. + 2004-03-22 Luke Gorrie <luke(a)bluetail.com> * doc/slime.texi: New user manual.

1 0

[slime-devel] possible bug in SLIME/Lispworks
by Bjørn Nordbø 23 Mar '04

23 Mar '04

... or maybe I'm just doing something wrong, but every time a break is evaluated, the following happens: CL-USER> (break) Error: No applicable methods for #<STANDARD-GENERIC-FUNCTION ENVIRONMENT-INTERNALS:ENVIRONMENT-DISPLAY-NOTIFIER 102376E2> with args (#<ENVIRONMENT:TTY-ENVIRONMENT 10218C3F> :CONDITION #<SIMPLE-ERROR 100ACE73> :RESTARTS (#<RESTART CONTINUE 100ABE1B> #<RESTART CONTINUE 100F55F7> #<RESTART CONTINUE 100FF51F> #<RESTART CONTINUE 10107FEB> #<RESTART CONTINUE 100E9EDF> #<RESTART CONTINUE 100E04F3> #<RESTART CONTINUE 100E4427> #<RESTART CONTINUE 10CEFB8B> #<RESTART ABORT 10CEFBC3>)) 1 (continue) Call #<STANDARD-GENERIC-FUNCTION ENVIRONMENT-INTERNALS:ENVIRONMENT-DISPLAY-NOTIFIER 102376E2> again 2 Call #<STANDARD-GENERIC-FUNCTION ENVIRONMENT-INTERNALS:ENVIRONMENT-DISPLAY-NOTIFIER 102376E2> again 3 Call #<STANDARD-GENERIC-FUNCTION ENVIRONMENT-INTERNALS:ENVIRONMENT-DISPLAY-NOTIFIER 102376E2> again 4 Call #<STANDARD-GENERIC-FUNCTION ENVIRONMENT-INTERNALS:ENVIRONMENT-DISPLAY-NOTIFIER 102376E2> again 5 Call #<STANDARD-GENERIC-FUNCTION ENVIRONMENT-INTERNALS:ENVIRONMENT-DISPLAY-NOTIFIER 102376E2> again 6 Call #<STANDARD-GENERIC-FUNCTION ENVIRONMENT-INTERNALS:ENVIRONMENT-DISPLAY-NOTIFIER 102376E2> again 7 Call #<STANDARD-GENERIC-FUNCTION ENVIRONMENT-INTERNALS:ENVIRONMENT-DISPLAY-NOTIFIER 102376E2> again 8 Call #<STANDARD-GENERIC-FUNCTION ENVIRONMENT-INTERNALS:ENVIRONMENT-DISPLAY-NOTIFIER 102376E2> again 9 Return from break. 10 (abort) Return to SLIME toplevel. Type :b for backtrace, :c <option number> to proceed, or :? for other options CL-USER 6 : 9 > I have tried bothe with the latest SLIME and FAIRLY-STABLE and both Lispworks 4.2 and 4.3, but it breaks every time I try. I'm baffled, and any hints about what might be wrong is most welcome. I'm running Lispworks/Solaris and Emacs 21.2.1 btw. I have tried playing around in order to understand what goes wrong, an it seems to me like Lispworks is in error. However breaking works fine outside of SLIME, ie. from the *inferior-lisp*. Here's the backtrace in case anyone can see anything interesting. Call to ERROR Call to CERROR Call to (METHOD NO-APPLICABLE-METHOD (T)) Call to CLOS::APPLY-NO-APPLICABLE-METHOD Call to ENVIRONMENT:DISPLAY-NOTIFIER Call to INVOKE-DEBUGGER Call to ERROR Call to CERROR Call to (METHOD NO-APPLICABLE-METHOD (T)) Call to CLOS::APPLY-NO-APPLICABLE-METHOD Call to ENVIRONMENT:DISPLAY-NOTIFIER Call to INVOKE-DEBUGGER Call to ERROR Call to CERROR Call to (METHOD NO-APPLICABLE-METHOD (T)) Call to CLOS::APPLY-NO-APPLICABLE-METHOD Call to ENVIRONMENT:DISPLAY-NOTIFIER Call to INVOKE-DEBUGGER Call to ERROR Call to CERROR Call to (METHOD NO-APPLICABLE-METHOD (T)) Call to CLOS::APPLY-NO-APPLICABLE-METHOD Call to ENVIRONMENT:DISPLAY-NOTIFIER Call to INVOKE-DEBUGGER Call to ERROR Call to CERROR Call to (METHOD NO-APPLICABLE-METHOD (T)) Call to CLOS::APPLY-NO-APPLICABLE-METHOD Call to ENVIRONMENT:DISPLAY-NOTIFIER Call to INVOKE-DEBUGGER Call to ERROR Call to CERROR Call to (METHOD NO-APPLICABLE-METHOD (T)) Call to CLOS::APPLY-NO-APPLICABLE-METHOD Call to ENVIRONMENT:DISPLAY-NOTIFIER Call to INVOKE-DEBUGGER Call to ERROR Call to CERROR Call to (METHOD NO-APPLICABLE-METHOD (T)) Call to CLOS::APPLY-NO-APPLICABLE-METHOD Call to ENVIRONMENT:DISPLAY-NOTIFIER Call to INVOKE-DEBUGGER Call to ERROR Call to CERROR Call to (METHOD NO-APPLICABLE-METHOD (T)) Call to CLOS::APPLY-NO-APPLICABLE-METHOD Call to ENVIRONMENT:DISPLAY-NOTIFIER Call to INVOKE-DEBUGGER Call to BREAK Call to EVAL Call to SWANK::EVAL-REGION Call to SWANK:LISTENER-EVAL Call to EVAL Call to SWANK::EVAL-FOR-EMACS Call to FUNCALL Call to SWANK::READ-FROM-EMACS Call to (SUBFUNCTION 2 SWANK::HANDLE-REQUEST) Call to SWANK::CALL-WITH-REDIRECTED-IO Call to SWANK::HANDLE-REQUEST Call to (SUBFUNCTION 1 SWANK::DISPATCH-EVENT) Call to (SUBFUNCTION MP::PROCESS-SG-FUNCTION MP::INITIALIZE-PROCESS-STACK) CL-USER 7 : 9 > -- Lisp can accomodate a wide range of psychological problems.

1 1

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

slime-devel March 2004