September 2005 - slime-devel - mailman3.common-lisp.net

[slime-devel] SLIME/Allegro communication-style
by Lynn Quam 15 Sep '05

15 Sep '05

Is it possible to run SLIME with Allegro using communication-style = :sigio rather than :spawn? I have a custom REPL loop which handles tcl/tk events in addition to input from *standard-input*. This work fine with SLIME and CMUCL with communication-style = :sigio. I have encountered difficulties implementing event-handling for tcl/tk with SLIME using communication-style = :spawn. One problem is that (method stream-listen slime-input-stream) on Allegro just checks if the buffer contains characters. It does not attempt to ask the socket is there are characters available. This does not appear to be a problem in CMUCL.

2 1

[slime-devel] Problem with EUC-JP encoding on Clisp-2.35 Win32 native (+fix)
by VAUCHER Laurent 15 Sep '05

15 Sep '05

Hi. My current configuration is Windoze/Emacs-21.3/Clisp-2.35 native (not the cygwin port, which is much slower) and it seems that a recent fix in swank-clisp.lisp in function find-encoding is not "portable". In fact, it is compatible with Clisp-2.35-Unix-flavor (MacOSX and Cygwin included) but not with the native W32 version which lacks some encodings (only 85 vs 121 for the unix version, to be accurate). I include a patch that solves the problem for me (by not using the inexistant symbol directly) ant should also work for those using the japanese encoding on Clisp versions supporting it. Laurent. --- swank-clisp.lisp.bak 2005-09-14 09:16:12.523260100 +0200 +++ swank-clisp.lisp 2005-09-14 09:19:14.141146000 +0200 @@ -123,7 +123,7 @@ :line-terminator :unix)) (:utf-8-unix (ext:make-encoding :charset 'charset:utf-8 :line-terminator :unix)) - (:euc-jp-unix (ext:make-encoding :charset 'charset:euc-jp + (:euc-jp-unix (ext:make-encoding :charset (find-symbol "EUC-JP" :charset) :line-terminator :unix)))) (defimplementation accept-connection (socket

2 1

[slime-devel] Daily ChangeLog diff
by lgorrie＠common-lisp.net 14 Sep '05

14 Sep '05

Index: slime/ChangeLog diff -u slime/ChangeLog:1.777 slime/ChangeLog:1.778 --- slime/ChangeLog:1.777 Tue Sep 13 07:37:16 2005 +++ slime/ChangeLog Wed Sep 14 22:18:39 2005 @@ -1,3 +1,8 @@ +2005-09-14 Matthias Koeppe <mkoeppe(a)mail.math.uni-magdeburg.de> + + * slime.el (slime-presentation-expression): Remove handling of + cons presentation-ids. + 2005-09-13 Alan Ruttenberg <alanr-l(a)mumble.net> * slime.el (defcustom slime-ed-use-dedicated-frame ... vs defvar

1 0

[slime-devel] disconnect after errors with openmcl
by Otto Diesenbacher 14 Sep '05

14 Sep '05

Hi, I have an error, when trying to evaluate any non-existent or wrong expression in the slime-repl running openmcl 14.3. After quiting the error-message I lose the connection and have to restart slime. The last lines in the *Message* buffer: error in process filter: slime-process-available-input: Wrong type argument: listp, S error in process filter: Wrong type argument: listp, S net-read error: (cl-assertion-failed (plusp length)) Lisp connection closed unexpectedly: deleted my emacs: (version) "GNU Emacs 22.0.50.1 (powerpc-apple-darwin8.2.0) of 2005-08-06 on quartet.local" slime is the last CVS-version btw.: with sbcl and clisp I haven't any problem. As I just startet learning lisp - I hope my descriptions are specific enough :) Thanks for any help in advance! otto

1 0

[slime-devel] Lisp connection closed unexpectedly: connection broken by remote peer
by martin 14 Sep '05

14 Sep '05

Dear all, Sorry for bothering you, but I have no idea how to fix a problem that I have with slime. I'm working on MacOsX 10.3 and I'm trying to switch from MCL to Emacs (22.0.50.1 powerpc-apple-darwin7.9.0) + Slime (1.2.1) + CMUCL (cmucl-darwin-exp-2004-02-03-010-ppc-darwin). It works and I'm really enthusiastic about it. But there is one problem. If I try to evaluate a s-expression that has more than 4040 characters, the lisp connection breaks with this message: Lisp connection closed unexpectedly: connection broken by remote peer In the *inferior-lisp* buffer, I get this: * ;; Connection to Emacs lost. ;; [ Error in function LISP::ASSERT-ERROR: Short read: length=4091 pos=4090] I attached a lisp file with two defuns. I can evaluate "test1", but not "test2" (test1 has 4040 characters, test2 has 4041 characters). Does anybody have an idea how to fix this? Thanks a lot, Martin

2 2

[slime-devel] clisp problems
by Rahul Garg 14 Sep '05

14 Sep '05

Hi. I am using CLISP 2.34 + slime 1.2.1 stable release (not cvs) version. Whenever i type out a function name slime is supposed to display the name and parameters in the mini-buffer. For example typing (gethash <space> will start displaying (gethash arg0 arg1) etc. Can i do something so that it displays the name of the arguments instead of arg0 and arg1. the names are displayed if i use cmucl but not if i use clisp. Any solutions? I also tried with Lisp In a Box using clisp 2.33 on winxp but it seems its a problem with clisp everywhere. sincerely, rahul

3 2

[slime-devel] Daily ChangeLog diff
by lgorrie＠common-lisp.net 13 Sep '05

13 Sep '05

Index: slime/ChangeLog diff -u slime/ChangeLog:1.774 slime/ChangeLog:1.777 --- slime/ChangeLog:1.774 Sat Sep 10 20:28:32 2005 +++ slime/ChangeLog Tue Sep 13 07:37:16 2005 @@ -1,3 +1,80 @@ +2005-09-13 Alan Ruttenberg <alanr-l(a)mumble.net> + + * slime.el (defcustom slime-ed-use-dedicated-frame ... vs defvar + + (defcustom slime-when-complete-filename-expand: Use + comint-replace-by-expanded-filename instead of + comint-dynamic-complete-as-filename to complete file names + + * swank.lisp (run-repl-eval-hooks .. finally (return vs no return + + inspector-call-nth-action Allow second value :replace for inspector actions + + (defvar *slime-inspect-contents-limit* default nil. How many elements of + a hash table or array to show by default. If table has more than + this then offer actions to view more. Set to nil for no limit. Probably should + set default to reasonable value - I like 200. + + (inspect-for-emacs ((ht hash-table) inspector)) - banner line is hash table object. + Respect *slime-inspect-contents-limit* + + (defmethod inspect-for-emacs ((array array) inspector) + Respect *slime-inspect-contents-limit* + + * swank-openmcl.lisp inspector for closures shows closed-over + values. To be fixed: inspector-princ needs to be loaded earlier + since swank package not available when compiling + +2005-09-13 Helmut Eller <heller(a)common-lisp.net> + + * present.lisp (menu-choices-for-presentation-id): Use + lookup-presented-object secondary return value instead of + *not-present*. + (execute-menu-choice-for-presentation-id, presenting-object-1): + Remove references to *can-print-presentation*. + + * slime.el (slime-current-output-id): Remove this ugly klugde. + (slime-repl-insert-result): New function. Handle the presentations + and other special cases cleaner. + (slime-repl-insert-prompt): Use it. The `result' arg is now a + structured list; update callers accordingly. + (slime-repl-return): Make the prefix arg work again. + (package-updating): The result of swank::listener-eval changed a + bit. Update the test. + + Remove some unnecessary uses of `defun*' and reindent it to 80 + columns. + + * swank.lisp: Simplify the object <-> presentation-id mapping. + (save-presented-object): Remove the optional `id' arg. + (lookup-presented-object): Id's should be fixnums not some cons + with fuzzy/non-documented meaning. Use the secondary return value + to test for absence of the id. Update callers accordingly. + (*not-present*): Deleted. + + Remove the repl result special cases, let the general presentation + machinery handle it. + (*last-repl-result-id*, add-repl-result, *current-id*) + (clear-last-repl-result): Deleted. + (listener-eval): Don't *current-id* to tag result values. + + (*can-print-presentation*): Deleted. Nobody quite knows whether + it's still needed so let just try without it. Updated referrers + accordingly. + + (eval-region, run-repl-eval-hooks): Move the eval hook stuff to + a separate function. + + * swank-loader.lisp (lisp-version-string)[cmu]: Replace spaces + with underscores. + +2005-09-12 NIIMI Satoshi <sa2c(a)sa2c.net> + + * swank.lisp, slime.el, swank-clisp.lisp, swank-sbcl.lisp: add + EUC-JP as coding system. This patch eliminates the requirement of + Mule-UCS to use Japanese characters. (Nice for pre-22 Emacs + users.) + 2005-09-10 Matthias Koeppe <mkoeppe(a)mail.math.uni-magdeburg.de> * slime.el (slime-enable-evaluate-in-emacs): Resurrected. @@ -69,7 +146,7 @@ (slime-choose-overlay-for-sexp): Ignore errors when stepping over forms. (slime-search-method-location, slime-goto-location-position): Move - all this regexpery to it's own function. + all this regexpery to its own function. (slime-recenter-if-needed, slime-repl-return): Factor some duplicated code into its own function. (slime-presentation-bounds, slime-presentation-around-point) @@ -121,6 +198,7 @@ (slime-y-or-n-p): New. 2005-08-29 Alan Ruttenberg <alanr-l(a)mumble.net> + * slime.el (sldb-insert-condition) - Add tooltip for long condition string which otherwise falls off the right of the screen * swank.lisp (list-threads) - thread name might be a symbol - pass

1 0

[slime-devel] [PATCH] add optimization settings menu
by Klaus Weidner 13 Sep '05

13 Sep '05

Hello, this patch adds menu options for changing the optimization settings in the lisp global environment. This way, you can recompile things for improved debuggability or speed without having to put declaims or declares in your source file. At least with sbcl, the proclaim call affects following compilations, but the compiled code can still override the global settings with decaim/declare. I know that this is just a quick hack, and I'm sure there's a more elegant way to configure global properties for the lisp environment, so please consider this as a proof of concept and not necessarily something for applying... -Klaus =================================================================== RCS file: /project/slime/cvsroot/slime/slime.el,v retrieving revision 1.544 diff -u -r1.544 slime.el --- slime.el 13 Sep 2005 05:37:16 -0000 1.544 +++ slime.el 13 Sep 2005 20:46:52 -0000 @@ -780,6 +780,10 @@ [ "Check Parens" check-parens t] [ "Update Indentation" slime-update-indentation ,C] [ "Select Buffer" slime-selector t]) + ("Optimization" + [ "Max Debug" slime-optimize-max-debug ,C ] + [ "High Speed" slime-optimize-high-speed ,C ] + [ "Unsafe Speed" slime-optimize-unsafe-speed ,C ]) ("Profiling" [ "Toggle Profiling..." slime-toggle-profile-fdefinition ,C ] [ "Profile Package" slime-profile-package ,C] @@ -6304,6 +6308,21 @@ (buffer-file-name))))))) (let ((lisp-filename (slime-to-lisp-filename (expand-file-name filename)))) (slime-eval-with-transcript `(swank:load-file ,lisp-filename)))) + + +;;;; Optimization + +(defun slime-optimize-max-debug () + (interactive) + (slime-eval-async '(cl:proclaim '(cl:optimize (cl:debug 3) (cl:speed 0) (cl:safety 3))))) + +(defun slime-optimize-high-speed () + (interactive) + (slime-eval-async '(cl:proclaim '(cl:optimize (cl:debug 0) (cl:speed 3) (cl:safety 3))))) + +(defun slime-optimize-unsafe-speed () + (interactive) + (slime-eval-async '(cl:proclaim '(cl:optimize (cl:debug 0) (cl:speed 3) (cl:safety 0))))) ;;;; Profiling

1 0

[slime-devel] bug in timezone formatting
by Vincent Arkesteijn 13 Sep '05

13 Sep '05

Hi, When inspecting an integer, one of its interpretations is as a universal time, which is decoded and printed in iso 8601 format. For example, C-c I (get-universal-time) showed: As time: 2005-09-13T00:43:42+01:00 Unfortunately, the indicated number of minutes in the timezone is always 00 or 01, which is problematic when the timezone offset from UTC is not an integer number of hours. Also, the dst return value from decode-universal-time is ignored, leading to a one-hour error when DST is in effect. Fix attached. Vincent.

1 0

[slime-devel] function names that start with "@"
by Alan Ruttenberg 13 Sep '05

13 Sep '05

I can't meta-point or get the arglist on space for a function that starts with an "@". Could someone suggest a way to fix this? Thanks, Alan

2 3