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

[slime-devel] XREF troubles
by Chris Capel 25 Nov '04

25 Nov '04

Whever I do a M-. on a method, up pops an XREF window. Nice. But there's a bit of a problem. It's in slime-mode. Slime-mode stomps on XREF's keys. Not nice. And this is a bit odd, because I was looking at the hook for XREF, and it disables slime-mode. But, since forever, I've had a lisp-mode hook that enables slime-mode, and I guess it gets run after the XREF hook. I think I read somewhere in the slime docs that you're supposed to set up your lisp-mode hook this way. So is this a bug, or do I need a different configuration? Chris Capel

2 2

[slime-devel] [PATCH] improvement to arglist
by Chris Capel 25 Nov '04

25 Nov '04

I consider the &whole keyword in a macro lambda list to be an implementation detail, and not something that needs to bother the person calling the macro (or something that needs to interfere with slime's automatic macro indention). I also consider &aux useless for the same purposes. So I changed "(defun arglist-to-string" in swank to remove "&whole var" and the rest of the lambda list after and including &aux from the lambda list. All in all, two extra functions, and a macro used by both; two new test cases for lambda list printing. Chris Capel

2 1

[slime-devel] [PATCH] improvement to repl history
by Chris Capel 25 Nov '04

25 Nov '04

I noticed that when you scroll through the history in the slime REPL, when you reach the end it doesn't clear the commandline, as bash and other shells do. I think this is rather inconvenient, so I patched it up. I made it so there's a blank space at the beginning, too, just because. No reason not to. I added a customization setting under Repl, that defaults to NIL. When it's true, the history will wrap around when the end is reached and you keep going. Chris Capel

2 1

[slime-devel] Fwd: Showing argument lists on return
by Svein Ove Aas 25 Nov '04

25 Nov '04

Ok, let's try that again, with me subscribed... ---------- Forwarded message ---------- From: Svein Ove Aas <sveina(a)gmail.com> Date: Tue, 23 Nov 2004 13:42:19 +0000 Subject: Showing argument lists on return To: slime-devel(a)common-lisp.net Cc: sveina(a)gmail.com Recently I've noticed I often press return-backspace-space in quick succession to keep the argument list display around. I put the following snippet in .emacs to fix it; I don't have time to make a proper patch of it now due to exams, but maybe some of you will find it handy. There are a number of obvious problems with it, and probably a few less obvious ones. Have fun! Also, I'm not subscribed to slime-devel, so please keep me CCed in any replies. ----- (define-key slime-mode-map (kbd "RET") 'slime-return) (defun slime-return () (interactive) (unwind-protect (if (slime-connected-p) (when (and slime-space-information-p (or (not (slime-busy-p)) ;; XXX should we enable this? ;; (not slime-use-sigint-for-interrupt)) )) (let ((names (slime-enclosing-operator-names))) (when names (slime-eval-async `(swank:arglist-for-echo-area (quote ,names)) (lambda (message) (if message (slime-message "%s" message))))))) (slime-message "Connect to a Lisp to see the argument list.")) (newline-and-indent))) (defun slime-repl-return () (interactive) (slime-check-connected) (assert (<= (point) slime-repl-input-end-mark)) (cond ((get-text-property (point) 'slime-repl-old-input) (slime-repl-grab-old-input)) (current-prefix-arg (slime-repl-send-input)) (slime-repl-read-mode ; bad style? (slime-repl-send-input t)) ((slime-input-complete-p slime-repl-input-start-mark slime-repl-input-end-mark) (slime-repl-send-input t)) (t (unwind-protect (when (and slime-space-information-p (slime-connected-p) (or (not (slime-busy-p)) ;; XXX should we enable this? ;; (not slime-use-sigint-for-interrupt)) )) (let ((names (slime-enclosing-operator-names))) (when names (slime-eval-async `(swank:arglist-for-echo-area (quote ,names)) (lambda (message) (if message (slime-message "%s" message))))))) (newline-and-indent)))))

2 1

[slime-devel] minor enhancements recommended!
by Surendra Singhi 25 Nov '04

25 Nov '04

Hi, I will suggest few minor changes which can be made, but will greatly simplify things for the users. I feel when there is an error *sldb * should show up in place of the repl buffer instead of replacing the code buffer. At present every time there is an error I have to replace the sldb buffer by code buffer and then repl by the sldb buffer. I am sure other users might also be facing this problem and so they must have written some short-cut keys to do this. But for xemac newbies(like me) it may be hard to write such short-cut keys and so this feature should be inbuilt. Second, I feel even when the sldb buffer is visible there should be a way of interpreting new function definitions(this maybe hard to implement). The reason why this feature is useful is that after looking at the error one changes the code and then executes it. But because repl is hung it the function doesn't executes. And so I am forced to go to sldb kill it, so that repl returns to normal state and then go back and interpret(execute) the function again. Moreover, when the sldb is killed the buffer positions revert back to what they were before the error occurred. This is annonying and I feel shound't happen. Thanks. -- Surendra Singhi www.public.asu.edu/~sksinghi

1 2

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

24 Nov '04

Index: slime/ChangeLog diff -u slime/ChangeLog:1.575 slime/ChangeLog:1.579 --- slime/ChangeLog:1.575 Sat Nov 20 21:49:09 2004 +++ slime/ChangeLog Wed Nov 24 21:29:39 2004 @@ -1,3 +1,51 @@ +2004-11-24 Helmut Eller <heller(a)common-lisp.net> + + * slime.el (slime-start-and-load): Use vanilla comint instead of + inf-lisp. Let's try that for a while. + (slime): Ask for the coding system when invoked with C-u C-u. + (slime-net-coding-system, slime-net-valid-coding-systems): Add + some alternatives for older Emacsen. + (slime-find-buffer-package): Skip quotes. Old code looks + sometimes like (in-package 'foo). + (slime-repl-mode-map): Inhibit C-c C-z. Avoids accidental loading + inf-lisp. + (slime-net-coding-system): Use find-coding-system in XEmacs. + coding-system-p means something different here. + (slime-repl-mode-map): XEmacs compatibilty: use (kbd "C-<up>") + instead of [C-up]. + + * swank.lisp (inspect-for-emacs-list): subseq on improper lists + breaks in Lispworks. Handle that case better. + + * swank-sbcl.lisp (inspect-for-emacs)[code-component]: Disassemble + code-components too. + + * swank-backend.lisp (import-swank-mop-symbols): Better error + message for the assertion. + + * swank-cmucl.lisp (debug-var-value): Return #:invalid or + #:unknown instead of :<not-available>. + (swank-compile-file): Load the fasl file only if load-p is true. + (inspect-for-emacs, inspect-alien-record, inspect-alien-pointer): + Add inspector support for some alien types. + + * swank-lispworks.lisp (emacs-connected): Set the sigint handler + only for single threaded operation. I.e. when + *communication-style* is nil. + + * swank-allegro.lisp (set-external-format): New function. Use LF + as eol mark. + (call-with-compilation-hooks): Trap compiler-notes too. + +2004-11-24 Luke Gorrie <luke(a)synap.se> + + * slime.el (slime-repl-mode-map): Add C-up and C-down to move + through history. Consistent with comint-mode. + (slime-repl-mode-map): Add slime-load-file on `C-c C-l' and + slime-compile-file on `C-c C-k'. This is mostly to override + unwanted inf-lisp bindings in lisp-mode-map. + (slime-load-file): Handle (buffer-file-name) being nil. + 2004-11-20 Helmut Eller <heller(a)common-lisp.net> * swank-sbcl.lisp (make-socket-io-stream): Add some #+sb-unicode. @@ -71,7 +119,7 @@ ACL's native inspector. (inspect-for-emacs (t), inspect-for-emacs (function)): Use it. - * wank.lisp (inspect-for-emacs array): Use row-major-aref instead + * swank.lisp (inspect-for-emacs array): Use row-major-aref instead of a displaced array. I hope that does the same. (inspect-for-emacs integer): Ignore errors in decode-universal-time. Negative values and, in SBCL, also small

1 0

[slime-devel] net-read error with slime -D "2004-11-19 19:00"
by Matthew Danish 24 Nov '04

24 Nov '04

The latest slime causes problems for me in Allegro 6.2 mlisp on Win32. I suspect the culprit is the new encoding stuff. I get a "net-read error: (end-of-file)" and an ugly debugger backtrace with lots of line noise in it. Any ideas? -- ;; Matthew Danish -- user: mrd domain: cmu.edu ;; OpenPGP public key: C24B6010 on keyring.debian.org

2 1

[slime-devel] [PATCH] macro indentation fix
by Chris Capel 24 Nov '04

24 Nov '04

My previous patch fixed the arglist code, but the function that returns macro indention was still broken. (In more ways than one, actually--it didn't treat &body and &rest equivalently. That problem turned out to be the actual problem with the indentation macro that led me on this adventure in the first place. Go figure.) This patch fixes that quite nicely, I believe. Chris Capel *** /home/chris/lisp-libs/slime/swank.lisp.temp 2004-11-24 04:58:59.724472328 -0600 --- /home/chris/lisp-libs/slime/swank.lisp 2004-11-24 05:01:38.509333360 -0600 *************** *** 3405,3411 **** (defun macro-indentation (arglist) (if (well-formed-list-p arglist) ! (position '&body (remove '&whole arglist)) nil)) (defun well-formed-list-p (list) --- 3405,3415 ---- (defun macro-indentation (arglist) (if (well-formed-list-p arglist) ! (let ((indent-list (remove-whole-from-arglist ! (remove-aux-from-arglist ! arglist)))) ! (or (position '&body indent-list) ! (position '&rest indent-list))) nil)) (defun well-formed-list-p (list)

2 3

[slime-devel] process filter bug
by Surendra Singhi 24 Nov '04

24 Nov '04

I am using xemacs and slime FAIRLY-STABLE version from cvs. At times when I execute some code on the *slime-repl*. It gives me the following error in the Xemacs *Warnings* buffer. (4) (error/warning) Error in process filter: (wrong-number-of-arguments nil 3) The *sldb [clisp/0]* is blank and hangs till i kill it by pressing q. -- Surendra Singhi www.public.asu.edu/~sksinghi

1 2

[slime-devel] Improvement to slime-maybe-start-lisp
by Chris Capel 24 Nov '04

24 Nov '04

Sorry, I don't have the resources to send a diff from here. When slime starts an inferior lisp process, it should probably check to make sure swank isn't already loaded (by an init file, or being in the default lisp image) before it calls load. This can take lots of extra time on a slow machine. So here's what I have it send the lisp process in slime-maybe-start-lisp: "(when (null (find-package :swank)) (load %S))\n" Chris Capel

1 1