February 2006 - slime-devel - mailman3.common-lisp.net

[slime-devel] problem with utf-8
by Nicolas Lamirault 11 Feb '06

11 Feb '06

hello, i must read a file encoded in utf-8 if i read this file from sbcl started in an xterm, i haven't got any error. i try to make this operation in my SLIME environment, but i have an error. so i put this in my emacs's configuration : (setq slime-net-coding-system 'utf-8-unix) but after this i can't start slime this is the error : Debugger entered--Lisp error: (cl-assertion-failed default-enable-multibyte-characters) signal(cl-assertion-failed (default-enable-multibyte-characters)) (or default-enable-multibyte-characters (signal (quote cl-assertion-failed) (list ...))) (progn (or default-enable-multibyte-characters (signal ... ...)) nil) (assert default-enable-multibyte-characters) (progn (assert default-enable-multibyte-characters)) (if (and (second props) (boundp ...)) (progn (assert default-enable-multibyte-characters))) (when (and (second props) (boundp ...)) (assert default-enable-multibyte-characters)) (let ((props ...)) (unless props (error "Invalid slime-net-coding-system: %s. %s" coding-system ...)) (when (and ... ...) (assert default-enable-multibyte-characters)) t) slime-check-coding-system(utf-8-unix) (let ((args ...)) (slime-check-coding-system coding-system) (when (or ... ...) (let ... ... ...))) (catch (quote --cl-block-slime-start--) (let (...) (slime-check-coding-system coding-system) (when ... ...))) (cl-block-wrapper (catch (quote --cl-block-slime-start--) (let ... ... ...))) (block slime-start (let (...) (slime-check-coding-system coding-system) (when ... ...))) (let* ((program ...) (program-args ...) (coding-system ...) (init ...) (name ...) (buffer ...)) (let (...) (while --cl-keys-- ...)) (block slime-start (let ... ... ...))) slime-start(:program "/usr/bin/sbcl" :program-args nil) apply(slime-start (:program "/usr/bin/sbcl" :program-args nil)) slime() call-interactively(slime) execute-extended-command(nil) call-interactively(execute-extended-command) have you got an idea for this problem ? thanks -- Nicolas Lamirault

3 2

[slime-devel] Not possible to inspect method slots in Lispworks
by Luke Crook 10 Feb '06

10 Feb '06

In the latest version of SLIME from CVS for Lispworks, I receive the following error when I attempt to inspect the values for a slot in an object: >> The slot #<STANDARD-EFFECTIVE-SLOT-DEFINITION A-METHOD 21C8CD0C> is missing from #<A-CLASS 21C8D174> (of class #<STANDARD-CLASS A-CLASS 21C4734C>), when reading the value. << This is for the following: CL-USER> (defclass a-class () ((a-method :accessor a-method))) #<STANDARD-CLASS A-CLASS 21C4734C> CL-USER> (make-instance 'a-class) #<A-CLASS 21C8D174> This used to work in the SLIME CVS version of perhaps a month or so ago. -Luke

3 3

[slime-devel] macro-indentation should ignore &optional
by darkness 10 Feb '06

10 Feb '06

Greetings. I've got a macro like: (defmacro tbf (&optional bar &body baz) ...) With SLIME from CVS, this gets indented as: (tbf (:arg value) (form1) (form2)) Whereas I think the correct indentation should be: (tbf (:arg value) (form1) (form2)) The problem (as I see it) is that swank:macro-indent counts &optional as an argument, rather than as a lambda list keyword. Below is a simple patch against SLIME CVS. =================================================================== RCS file: /project/slime/cvsroot/slime/swank.lisp,v retrieving revision 1.360 diff -u -r1.360 swank.lisp --- swank.lisp 6 Feb 2006 18:42:09 -0000 1.360 +++ swank.lisp 9 Feb 2006 17:16:26 -0000 @@ -4369,7 +4369,7 @@ (defun macro-indentation (arglist) (if (well-formed-list-p arglist) - (position '&body (clean-arglist arglist)) + (position '&body (remove '&optional (clean-arglist arglist))) nil)) (defun well-formed-list-p (list)

2 1

[slime-devel] Emacs freezing bug fix
by Timofei Shatrov 10 Feb '06

10 Feb '06

I found a fix for a bug that plagued me for a long time. Whenever I was connected to Internet, Slime was consistently hanging Emacs cold. The bug apparently stems from make-network-process function of Emacs which demonstrates the same behavior. Anyway, there is a fix: replace all occurences of "127.0.0.1" in slime.el with "localhost" - surprisingly this solves the problem. -- Best regards, Timofei Shatrov mailto:grue@mail.ru

2 1

RE: [slime-devel] Re: Slime inserting an extra parenthesis
by Bradford W Miller 08 Feb '06

08 Feb '06

I smell a project: http://www.rci.rutgers.edu/~cfs/472_html/Planning/PlanRecog.html Unfortunately, it's close enough to my day job I'd have IP problems doing the work. But I'd be happy to kibitz. -----Original Message----- From: slime-devel-bounces(a)common-lisp.net [mailto:slime-devel-bounces@common-lisp.net] On Behalf Of edi(a)agharta.de Sent: Tuesday, February 07, 2006 6:49 AM To: jan(a)rychter.com Cc: slime-devel(a)common-lisp.net Subject: Re: [slime-devel] Re: Slime inserting an extra parenthesis On Tue, 07 Feb 2006 10:27:40 +0100, Jan Rychter <jan(a)rychter.com> wrote: > Right, well -- but shouldn't completion be smart enough to know if > the parentheses should be closed or not? At the moment I can't really imagine how SLIME should be able to always do The Right Thing[TM] unless it can read your mind. But I think patches are always welcome. _______________________________________________ slime-devel site list slime-devel(a)common-lisp.net http://common-lisp.net/mailman/listinfo/slime-devel

1 0

[edi@agharta.de: Re: [slime-devel] Re: Slime inserting an extra parenthesis]
by Jeff Cunningham 08 Feb '06

08 Feb '06

> I don't get the described behavior in gnu-emacs. Must be an xemacs > thing. >... > Actually, I just noticed that in this case M-q does manage to do the > 'right thing' with the parenthesis. Try that. That was probably meant for the list. I don't have a problem with SLIME's handling of parentheses, the OP has. Cheers, Edi. Sorry, Edi, I thought I was mailing this to the list. -Jeff

1 0

[slime-devel] Slime inserting an extra parenthesis
by Jan Rychter 07 Feb '06

07 Feb '06

I often use insert-parentheses in XEmacs, and it seems to interact badly with SLIME. An example: I begin with: CL-USER> () ^point I then type asdf-u<TAB>:u<TAB> and end up with this: CL-USER> (asdf-upgrade:upgrade)) ^point Notice the extra parenthesis that SLIME inserted. My guess is that SLIME shouldn't insert this parenthesis. I can understand where this came from (no extra args, so the user surely wants to close the parenthesis), but for someone mostly working with balanced parentheses this isn't too useful. --J.

2 3

[slime-devel] patch: make slime-echo-arglist show initargs/initforms for make-instance
by Luís Oliveira 06 Feb '06

06 Feb '06

Hello, The attached patch makes slime-echo-arglist detect (make-instance 'quoted-symbol ...) forms and uses the quoted symbol to find the class and present the initargs and initforms in the mini-buffer. For example: CL-USER> (defclass foo () ((a :initarg :initarg-a :initform nil) (b :initarg :initarg-b) c)) #<STANDARD-CLASS FOO> CL-USER> (make-instance<space> ;; minibuffer displays: (make-instance class &rest initargs) CL-USER> (make-instance 'foo<space> ;; minibuffer displays: (make-instance 'foo &key (initarg-a nil) initarg-b) The code is a bit ugly, specially on the elisp side. My elisp-fu being very weak doesn't help. Suggestions on how to improve that are most welcome! BTW, the ChangeLog's encoding seems to be latin-1, wouldn't utf-8 be better? -- Luís Oliveira luismbo (@) gmail (.) com Equipa Portuguesa do Translation Project http://www.iro.umontreal.ca/translation/registry.cgi?team=pt

2 3

[slime-devel] Followup on dangerous bug
by Bradford W. Miller 06 Feb '06

06 Feb '06

Last week I reported on a problem where a buffer in the slime-typeout window would be overwritten if the *SLIME-Typeout* buffer is killed by the user. A suggested fix follows: The code calls the function slime-typeout-active-p in order to assure the window is open. However, this call does NOT insure that the correct buffer is exposed in that window. The saved window is in the variable "slime-typeout-window", but this will reflect any exposed buffer in that window, not "*SLIME-Typeout*". I have modified the slime-typeout-active-p function as follows, to make sure that the window has the correct buffer in the foreground before the caller does the write. (There is still the potential for a collision if the buffer is switched between calling this function and writing to it, however, I'm unfamiliar with mechanisms for locking frames in emacs to prevent that). Instead, perhaps for full safety, the call (window-buffer slime-typeout-window) in slime-typeout- message should be (get-buffer-create "*SLIME-Typeout*") instead. That way the write always occurs to the correct buffer, and not the one that happens to be exposed. At least on my local copy, these functions works normally, and killing *SLIME-Typeout* causes it to be restored before, e.g., showing the arglist for the function at prompt while in the slime REPL. These patches fix two problems: 1) that the use of the typeout window may go to the wrong buffer if the wrong buffer happens to be exposed and 2) that the *SLIME-Typeout* buffer is not exposed in the window created for that express purpose. Allow me to suggest these patchs for inclusion in the CVS. (defun slime-typeout-active-p () ;; modified 2/6/06 BWM (cond ((and slime-typeout-window (window-live-p slime-typeout-window)) (save-selected-window (select-window slime-typeout-window) (switch-to-buffer "*SLIME-Typeout*")) t) ; return true -it's active and ready (t nil))) (defun slime-typeout-message (format-string &rest format-args) (assert (slime-typeout-active-p)) (with-current-buffer (get-buffer-create "*SLIME-Typeout*") ;; modified 2/6/06 BWM (erase-buffer) (insert (apply #'format format-string format-args))))

1 0

[slime-devel] Bug fix for inspecting logical-pathnames
by Lynn Quam 05 Feb '06

05 Feb '06

There is a missing comma for backquote substitution in the following method that causes an error when inpecting logical-pathnames using SLIME. Error in function SWANK::INSPECTOR-CONTENT-FOR-EMACS: destructure-case failed: (PATHNAME-HOST PATHNAME) [Condition of type SIMPLE-ERROR] Restarts: 0: [ABORT] ABORT 1: [ABORT-REQUEST] Abort handling SLIME request. 2: [ABORT] Return to Top-Level. Backtrace: 0: (SWANK::INSPECTOR-CONTENT-FOR-EMACS ("Namestring" ": " (:VALUE "FREEDIUS:LISP;MATH;MATRICES.LISP") (:NEWLINE) "Physical pathname: " ...)) 1: (SWANK::INSPECT-OBJECT #P"FREEDIUS:LISP;MATH;MATRICES.LISP" #<SWANK-BACKEND::CMUCL-INSPECTOR {5BA62CBD}>) ... The literal form (pathname-host pathname) gets passed to SWANK::INSPECTOR-CONTENT-FOR-EMACS instead of its value. I have added the comma in the line indicated by ; <<< COMMA NEEDED HERE (defmethod inspect-for-emacs ((pathname logical-pathname) inspector) (declare (ignore inspector)) (values "A logical pathname." (append (label-value-line* ("Namestring" (namestring pathname)) ("Physical pathname: " (translate-logical-pathname pathname))) `("Host: " ,(pathname-host pathname) ; <<< COMMA NEEDED HERE " (" (:value ,(logical-pathname-translations (pathname-host pathname))) "other translations)" (:newline)) (label-value-line* ("Directory" (pathname-directory pathname)) ("Name" (pathname-name pathname)) ("Type" (pathname-type pathname)) ("Version" (pathname-version pathname)) ("Truename" (if (not (wild-pathname-p pathname)) (probe-file pathname)))))))

1 0