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

[slime-devel] Closing a SLIME connection without stopping the SWANK server
by Sebastian Tennant 10 Oct '09

10 Oct '09

Hi all, Having now got a SWANK server that doesn't kill its host Lisp when it is stopped, I'm wondering if it's possible to close a SLIME connection without stopping the SWANK server? At the moment, when I do ',disconnect' in my SLIME REPL it stops the SWANK server and I have to re-attach to the SBCL image (using attachtty) and recreate the SWANK server before I can re-connect again with SLIME. I've tried setting swank:*dont-close* to true, but it makes no difference. Perhaps ',disconnect' is supposed to stop the SWANK server... I don't know. My swank:*comunication-style* is currently 'spawn'. I wonder if changing this might affect things? (Strangely, swank-backend:preferred-communication-style is unbound). Kind regards, Seb -- Emacs' AlsaPlayer - Music Without Jolts Lightweight, full-featured and mindful of your idyllic happiness. http://home.gna.org/eap

3 3

[slime-devel] Stopping swank server kills sbcl
by Sebastian Tennant 10 Oct '09

10 Oct '09

Hi all, In bash: # sbcl --core sbcl.core-with-swank This is SBCL 1.0.30, an implementation of ANSI Common Lisp. More information about SBCL is available at <http://www.sbcl.org/>. SBCL is free software, provided as is, with absolutely no warranty. It is mostly in the public domain; some portions are provided under BSD-style licenses. See the CREDITS and COPYING files in the distribution for more information. * (swank:create-server :port 49153) ;; Swank started at port: 49153. 49153 * (swank:stop-server 49153) # sbcl dies, which I obviously don't want it to. My core includes the latest slime/swank from CVS, on a Debian (Lenny) box. I need to be able to stop swank the server cleanly, whilst leaving sbcl running. Is this a (serious) bug or am I missing something? Regards, Seb -- Emacs' AlsaPlayer - Music Without Jolts Lightweight, full-featured and mindful of your idyllic happiness. http://home.gna.org/eap

2 4

[slime-devel] dwim.hu slime branch
by Attila Lendvai 10 Oct '09

10 Oct '09

dear list, as some of you may remember, we've forked slime some time ago. short after that slime proper was refactored which made our fork a dead-end, and required manual rebasing of all our changes. recently i've started using the refactored slime HEAD and started to gradually rebase the most missed modifications from us to the new HEAD. the work is not finished, but we are using this new branch on a daily basis now. the result of this is two public darcs2 repos: 1) a tailor converted darcs2 mirror of the official slime repo without any changes by us: http://dwim.hu/darcsweb/darcsweb.cgi?r=HEAD%20slime;a=summary 2) a darcs repo based on 1) containing our extra patches: http://dwim.hu/darcsweb/darcsweb.cgi?r=HEAD%20hu.dwim.slime;a=summary currently the most notable differences are in the inspector. some of the extras: - i've re-added the short-lived dwim inspecting feature: when C-i is pressed on a name, tries to look up things with that name, then puts the name on the inspector stack, and instead of the name, it inspects the thing that was looked up by that name. IOW, if you press C-i on the name of a class, then the inspector will come up with the class, and if it's not what you want, then you can press 'l' to go back to the symbol. - extended the fancy inspector by an inspectable type, and the object itself is also an inspector link, so that it's copyable to the repl as needed. - several type of objects are displayed nicer in the inspector - inspecting stick to the same thread, so when e.g. inspecting an object in the debugger then the svuc customizations run in the same thread (think of *transaction*...) besides the inspector changes, there are numerous small convenience changes, some bugfixes, and a bit of fuzzy completion speedup. the detailed list of extra patches is available at the end of the mail. i'm not welcom on slime-devel, so i cut it short here, and i'm not planning to write much more about this fork here. but i don't mind if someone with the commit bit commits some of these changes to the official branch either. it's a matter of darcs diff -u -p "some regexp in the patch description" >/tmp/x.diff happy hacking, -- attila currently these extra paches are in 2) compared to 1): Wed Sep 9 18:29:13 CEST 2009 attila.lendvai(a)gmail.com * Pad restart descriptions to start at the same column in sldb Wed Sep 9 20:13:52 CEST 2009 attila.lendvai(a)gmail.com * In xref buffer, switch the enter and space key bindings to follow the common "enter for sideffects" behaviour Wed Sep 9 20:45:24 CEST 2009 attila.lendvai(a)gmail.com * Smarter sldb fame details when there's nothing to display Thu Sep 10 02:47:51 CEST 2009 attila.lendvai(a)gmail.com * support an &key length-limit in to-string Fri Sep 11 10:53:48 CEST 2009 attila.lendvai(a)gmail.com * reintroduce dwim inspecting to the fancy inspector contrib Fri Sep 11 10:58:05 CEST 2009 attila.lendvai(a)gmail.com * fix: presentations.el depends on repl.el Fri Sep 11 11:10:35 CEST 2009 attila.lendvai(a)gmail.com * some inspector features: labels, displaying the type (further inspectable) Fri Sep 11 11:35:17 CEST 2009 attila.lendvai(a)gmail.com * follow sb-debug:backtrace-as-list changes (WARNING: as of this writing, the sbcl side of this patch is not in sbcl head, yet) Fri Sep 11 12:06:15 CEST 2009 attila.lendvai(a)gmail.com * Fuzzy completion: speed up by 2-4 times (on sbcl) Fri Sep 11 12:12:10 CEST 2009 attila.lendvai(a)gmail.com * Fix safe-source-location-for-emacs on sbcl to be safe actually Fri Sep 11 12:16:46 CEST 2009 attila.lendvai(a)gmail.com * disable undo in fuzzy buffer and in the repl while printing non-user input Fri Sep 11 12:21:25 CEST 2009 attila.lendvai(a)gmail.com * drop default key binding for slime-quit-connection-at-point, add disconnect on key 'k' no, you don't want to press that button on a production system accidentally Fri Sep 11 15:05:29 CEST 2009 attila.lendvai(a)gmail.com * patch slime-current-package in the repl contrib to return (slime-lisp-package) as a last try Fri Sep 11 15:06:37 CEST 2009 attila.lendvai(a)gmail.com * fix: printing of special variables in the minibuffer is now safe from print-object errors (and popping up debugger windows) Fri Sep 11 15:59:09 CEST 2009 attila.lendvai(a)gmail.com * Line up frame locals in the same column in sldb Fri Sep 11 16:20:49 CEST 2009 attila.lendvai(a)gmail.com * fix: presentations are properly recognized at the very beginning of a buffer Fri Sep 11 16:53:35 CEST 2009 attila.lendvai(a)gmail.com * move sbcl specific type inspectors into the swank-sbcl-exts contrib, use fancy inspector features Fri Sep 11 17:10:33 CEST 2009 attila.lendvai(a)gmail.com * fix: inspector keeps using the same thread and package throughout inspecting (e.g. for svuc customizations depending on a transaction) Fri Sep 11 17:33:34 CEST 2009 attila.lendvai(a)gmail.com * added some default values for the sldb faces Fri Sep 11 18:12:55 CEST 2009 attila.lendvai(a)gmail.com * make the inspector work more like web browsers: bind left click to follow link, and mouse 6/7 for forward and back Fri Sep 11 18:29:38 CEST 2009 attila.lendvai(a)gmail.com * fix: slime-repl-delete-from-input-history saved back the old history entries before the in-memory ones Fri Sep 11 18:31:26 CEST 2009 attila.lendvai(a)gmail.com * Always jump to the prompt when navigating the repl history Fri Sep 11 18:39:00 CEST 2009 attila.lendvai(a)gmail.com * Lower the score of fuzzy matches towards the end except for exact matches at the very end Sat Sep 12 21:29:57 CEST 2009 attila.lendvai(a)gmail.com * try not to (goto-char (point-max)) when copying presentations to repl for more flexible editing doesn't work from sldb yet, due to slime-repl-send-string doing the same Sat Sep 12 21:33:03 CEST 2009 attila.lendvai(a)gmail.com * added separate host and port history for slime-connect Sat Sep 12 21:33:12 CEST 2009 attila.lendvai(a)gmail.com * bind M-RET to copy-down-to-repl in the inspector, too (as in sldb) Sat Sep 12 21:40:34 CEST 2009 attila.lendvai(a)gmail.com * sldb is smarter in pruning swank frames Sat Sep 12 22:36:23 CEST 2009 attila.lendvai(a)gmail.com * slime-selector can receive the input key as an argument for using from key bindings Sun Sep 13 00:47:28 CEST 2009 attila.lendvai(a)gmail.com * fix: don't just define, actually use slime-inspector-insert-ispec-function among other things, it fixes sexp navigation in the inspector on unprintable objects Sun Sep 13 01:44:56 CEST 2009 attila.lendvai(a)gmail.com * Assign id's to restarts in sldb in such a way that older restarts are assigned a lower id As a result of this the restarts registered earlier (things like "Kill thread") will always get the same id and key shortcut. Tue Sep 15 17:56:05 CEST 2009 attila.lendvai(a)gmail.com * new features for label-value-line macros Thu Sep 17 16:02:18 CEST 2009 attila.lendvai(a)gmail.com * FIX: source-path-parser honours the swank:*readtable-alist* This fixes jumping to source forms in files with a custom readtable Tue Sep 22 10:21:32 CEST 2009 attila.lendvai(a)gmail.com * sensible error when a lisp implementation entry is not found Tue Sep 22 10:53:59 CEST 2009 attila.lendvai(a)gmail.com * Print out the current package at various minibuffer prompts Tue Sep 22 11:02:17 CEST 2009 attila.lendvai(a)gmail.com * widen a bit the sb-sprof output to make it more readable Wed Sep 23 19:16:03 CEST 2009 attila.lendvai(a)gmail.com * smarter inspector for several types Wed Sep 23 19:16:47 CEST 2009 attila.lendvai(a)gmail.com * added *skip-slot-value-errors-while-inspecting* and some useful restarts for mis-behaving svuc's in the inspector Fri Sep 25 09:44:10 CEST 2009 attila.lendvai(a)gmail.com * merge with 2009-09-25 06:43:33 by trittweiler Mon Sep 28 21:44:27 CEST 2009 attila.lendvai(a)gmail.com * push :hu.dwim.slime in *features* Thu Oct 8 21:53:31 CEST 2009 attila.lendvai(a)gmail.com * make slime-parent-map bindings available when reading from the minibuffer

1 0

[slime-devel] Daily ChangeLog diff
by Helmut Eller 10 Oct '09

10 Oct '09

Index: slime/ChangeLog diff -u slime/ChangeLog:1.1869 slime/ChangeLog:1.1870 --- slime/ChangeLog:1.1869 Tue Oct 6 16:12:03 2009 +++ slime/ChangeLog Fri Oct 9 19:05:12 2009 @@ -1,3 +1,8 @@ +2009-10-09 Stas Boukarev <stassats(a)gmail.com> + + * swank.lisp (stop-server): (list-threads) returns threads offset by 1, + don't kill the wrong thread. + 2009-10-06 Stas Boukarev <stassats(a)gmail.com> * swank-ccl.lisp (map-backtrace): Handle null end-frame-number argument. Index: slime/contrib/ChangeLog diff -u slime/contrib/ChangeLog:1.250 slime/contrib/ChangeLog:1.252 --- slime/contrib/ChangeLog:1.250 Mon Sep 28 23:21:30 2009 +++ slime/contrib/ChangeLog Fri Oct 9 10:57:45 2009 @@ -1,3 +1,14 @@ +2009-10-09 Stas Boukarev <stassats(a)gmail.com> + + * swank-sprof.lisp: Add ability to exclude functions which symbols + are from swank package. + * slime-sprof.el (slime-sprof-toggle-swank-exclusion): New function. + Bound to s in the slime-sprof buffer. + * slime-sprof.el (slime-sprof-exclude-swank): New variable for + controlling exclusion of swank functions. + + * slime-sprof.el: Slightly factor code, add menu entries. + 2009-09-29 Stas Boukarev <stassats(a)gmail.com> * slime-repl.el (slime-sync-package-and-default-directory):

1 0

[slime-devel] Daily ChangeLog diff
by Helmut Eller 07 Oct '09

07 Oct '09

Index: slime/ChangeLog diff -u slime/ChangeLog:1.1868 slime/ChangeLog:1.1869 --- slime/ChangeLog:1.1868 Mon Sep 28 10:56:19 2009 +++ slime/ChangeLog Tue Oct 6 16:12:03 2009 @@ -1,3 +1,7 @@ +2009-10-06 Stas Boukarev <stassats(a)gmail.com> + + * swank-ccl.lisp (map-backtrace): Handle null end-frame-number argument. + 2009-09-28 Stas Boukarev <stassats(a)gmail.com> * swank.lisp (set-package): Provide a more meaningful error

1 0

Re: [slime-devel] SLIME+SBCL+WIN32
by Nikodemus Siivola 05 Oct '09

05 Oct '09

2009/10/5 Luke J Crook <luke(a)balooga.com>: >> What is the value of SBCL_HOME in the Emacs process? > > I have this set as a win32 environmental variable; > > C:\home\bin\sbcl\1.0.29\ Does this directory contain subdirectories such as asdf sb-grovel sb-simple-streams asdf-install sb-introspect sb-sprof and does each of them contain a .fasl of the same name (asdf/asdf.fasl, etc)? > There is a way to set this for the emacs process itself? That should not be necessary, as long as M-x getenv SBCL_HOME gives you the right result. >> What is value of inferior-lisp-program in elisp? > > "c:/home/bin/sbcl/1.0.29/sbcl.exe --noinitform" ITYM --noinform ...not that it should matter. How did you install SBCL? If you install SBCL using the MSI installer, it should set up SBCL_HOME and PATH for you, so that (setq inferior-lisp-program "sbcl") should just work. Cheers, -- Nikodemus

1 0

[slime-devel] SLIME+SBCL+WIN32
by Luke J Crook 05 Oct '09

05 Oct '09

I've been getting the following slime error from CVS using SBCL 1.0.29 for win32 in Emacs; See also: The ANSI Standard, Macro DEFPACKAGE ; compiling file "c:\\home\\bin\\central-registry\\slime\\swank-source-path- parser.lisp" (written 25 JUN 2009 08:23:18 PM): ; c:\home\.slime\fasl\2009-09-24\sbcl-1.0.29-win32-x86\swank-source-path- parser.fasl written ; compilation finished in 0:00:00.125 ; loading #P"c:\\home\\.slime\\fasl\\2009-09-24\\sbcl-1.0.29-win32-x86\\swank- source-path-parser.fasl" ; compiling file "c:\\home\\bin\\central-registry\\slime\\swank-source-file- cache.lisp" (written 15 MAY 2009 04:33:37 PM): ; c:\home\.slime\fasl\2009-09-24\sbcl-1.0.29-win32-x86\swank-source-file- cache.fasl written ; compilation finished in 0:00:00.094 ; loading #P"c:\\home\\.slime\\fasl\\2009-09-24\\sbcl-1.0.29-win32-x86\\swank- source-file-cache.fasl" ; compiling file "c:\\home\\bin\\central-registry\\slime\\swank- sbcl.lisp" (written 25 SEP 2009 11:32:52 PM): ; ; compilation unit aborted ; caught 1 fatal ERROR condition ; compilation aborted after 0:00:00.094 ;; ;; Error while loading: c:\home\.slime\fasl\2009-09-24\sbcl-1.0.29-win32- x86\swank-sbcl.fasl ;; Condition: Don't know how to REQUIRE SB-BSD-SOCKETS. ;; See also: ;; The SBCL Manual, Variable *MODULE-PROVIDER-FUNCTIONS* ;; The SBCL Manual, Function REQUIRE ;; Aborting.

3 3

[slime-devel] slime-indentation breaks indentation
by Stanislaw Halik 30 Sep '09

30 Sep '09

I have a macro with such lambda-list: (defmacro fluid-let (&environment env (&rest bindings) &body body) ...) With „slime-indentation”, it gets indented pretty badly: (fluid-let () 42) Without it, it's (correctly) indented as with LET bodies. -- polscy Lispnicy, łączmy się -- #lisp-pl na freenode Lisp stuff: <http://tehran.lain.pl>

1 0

[slime-devel] copious output makes point jump around
by Stanislaw Halik 29 Sep '09

29 Sep '09

Sometimes, not very deterministically, after (require 'foo), point jumps from the prompt way up, into a seemingly random place in copious output. It's not a compilation note, not a warning, but something like "compiling (defun foo ...". It happens even if no notes are emitted. Someone on #lisp confirmed that as well. I don't have any idea on how to debug or isolate it. -- polscy Lispnicy, łączmy się -- #lisp-pl na freenode Lisp stuff: <http://tehran.lain.pl>

1 0

[slime-devel] Daily ChangeLog diff
by Helmut Eller 29 Sep '09

29 Sep '09

Index: slime/ChangeLog diff -u slime/ChangeLog:1.1866 slime/ChangeLog:1.1868 --- slime/ChangeLog:1.1866 Sat Sep 26 19:24:50 2009 +++ slime/ChangeLog Mon Sep 28 10:56:19 2009 @@ -1,3 +1,11 @@ +2009-09-28 Stas Boukarev <stassats(a)gmail.com> + + * swank.lisp (set-package): Provide a more meaningful error + message when package doesn't exist. + + * swank-lispworks.lisp (replace-strings-with-symbols): Didn't work on + non-proper lists. Reported by Madhu. + 2009-09-27 Tobias C. Rittweiler <tcr(a)freebits.de> * swank-sbcl.lisp (call-with-debugger-hook): Correctly deal with Index: slime/contrib/ChangeLog diff -u slime/contrib/ChangeLog:1.248 slime/contrib/ChangeLog:1.250 --- slime/contrib/ChangeLog:1.248 Fri Sep 25 02:43:33 2009 +++ slime/contrib/ChangeLog Mon Sep 28 23:21:30 2009 @@ -1,3 +1,13 @@ +2009-09-29 Stas Boukarev <stassats(a)gmail.com> + + * slime-repl.el (slime-sync-package-and-default-directory): + Better checking for packages. + +2009-09-28 Stas Boukarev <stassats(a)gmail.com> + + * slime-repl.el (slime-sync-package-and-default-directory): + Do not try to set package if it doesn't exists. + 2009-09-25 Tobias C. Rittweiler <tcr(a)freebits.de> * swank-sbcl-exts.lisp: Add EVAL-WHEN to prevent subtle dependency

1 0