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

[slime-devel] New trace fixes for CMUCL
by Dan Pierson 28 Apr '05

28 Apr '05

Here's a patch to more carefully try both (METHOD ...) and (PCL:FAST-METHOD ...) when tracing methods on CMUCL. Dan Pierson Index: swank-cmucl.lisp =================================================================== RCS file: /project/slime/cvsroot/slime/swank-cmucl.lisp,v retrieving revision 1.143 diff -u -r1.143 swank-cmucl.lisp --- swank-cmucl.lisp 3 Apr 2005 23:27:41 -0000 1.143 +++ swank-cmucl.lisp 28 Apr 2005 15:11:12 -0000 @@ -2136,9 +2136,13 @@ (let ((name (second spec))) (toggle-trace-aux name :methods name))) ((:defmethod) - (toggle-trace-aux `(method ,(cdr spec))) - ;; Man, is this ugly - (toggle-trace-aux `(pcl::fast-method ,(cdr spec)))) + (cond ((fboundp `(method ,@(cdr spec))) + (toggle-trace-aux `(method ,(cdr spec)))) + ;; Man, is this ugly + ((fboundp `(pcl::fast-method ,@(cdr spec))) + (toggle-trace-aux `(pcl::fast-method ,@(cdr spec)))) + (t + (error 'undefined-function :name (cdr spec))))) ((:call) (destructuring-bind (caller callee) (cdr spec) (toggle-trace-aux (process-fspec callee)

2 1

[slime-devel] Fixes for slime-browse-xrefs on cmucl
by Dan Pierson 28 Apr '05

28 Apr '05

I've made a couple of fixes to slime-browse-xrefs: 1. Add support for the common 'q' keystroke to quit out of the xref. 2. Remove the (FLET ...) entries which appear on at least CMUCL. I don't believe you can actually expand them on any current implementation and they just mess up the browse tree. 3. Hack swank:xref to look up (METHOD ...) entries on CMUCL. This really shouldn't be here, but I can't see how to put it in swank-cmucl.lisp and avoid the error thrown by swank:xref. 4. Fix slime-parse-context so that pressing, say, C-cC-t when point is on a '-' in a symbol name won't break. (This is part of a CMUCL patch for the new trace which will appear in my next post.) Dan Pierson Index: swank.lisp =================================================================== RCS file: /project/slime/cvsroot/slime/swank.lisp,v retrieving revision 1.295 diff -u -r1.295 swank.lisp --- swank.lisp 21 Apr 2005 07:39:12 -0000 1.295 +++ swank.lisp 28 Apr 2005 15:10:45 -0000 @@ -3038,18 +3038,29 @@ errors)))))))) (defslimefun xref (type symbol-name) - (let ((symbol (parse-symbol-or-lose symbol-name *buffer-package*))) - (group-xrefs - (ecase type - (:calls (who-calls symbol)) - (:calls-who (calls-who symbol)) - (:references (who-references symbol)) - (:binds (who-binds symbol)) - (:sets (who-sets symbol)) - (:macroexpands (who-macroexpands symbol)) - (:specializes (who-specializes symbol)) - (:callers (list-callers symbol)) - (:callees (list-callees symbol)))))) + ;; Minimal attempts to retrieve something when passed "(METHOD FOO ...)". + ;; Return nil when passed something incomprehensible like "(FLET FOO ...)". + ;;---*** This is way too CMUCL-specific for the front end, but I have to + ;;---*** do something before parse-symbol-or-lose looses. + (let* ((spec (from-string symbol-name)) + (symbol-string nil)) + (cond ((and (listp spec) (string-equal (string (first spec)) "METHOD")) + (setq symbol-string (string (second spec)))) + ((not (listp spec)) + (setq symbol-string symbol-name))) + (when symbol-string + (let ((symbol (parse-symbol-or-lose symbol-string *buffer-package*))) + (group-xrefs + (ecase type + (:calls (who-calls symbol)) + (:calls-who (calls-who symbol)) + (:references (who-references symbol)) + (:binds (who-binds symbol)) + (:sets (who-sets symbol)) + (:macroexpands (who-macroexpands symbol)) + (:specializes (who-specializes symbol)) + (:callers (list-callers symbol)) + (:callees (list-callees symbol)))))))) ;;;; Inspecting Index: slime.el =================================================================== RCS file: /project/slime/cvsroot/slime/slime.el,v retrieving revision 1.484 diff -u -r1.484 slime.el --- slime.el 18 Apr 2005 19:23:40 -0000 1.484 +++ slime.el 28 Apr 2005 15:10:48 -0000 @@ -5587,7 +5587,7 @@ (backward-up-list 1) (slime-parse-context `(setf ,name))) ((slime-in-expression-p '(defmethod *)) - (unless (looking-at "\\>\\|\\s ") + (unless (looking-at "\\s ") (forward-sexp 1)) ; skip over the methodname (let (qualifiers arglist) (loop for e = (read (current-buffer)) @@ -7485,6 +7485,14 @@ :dynargs 'slime-expand-class-node :has-echildren t)))) +(defvar slime-browser-map nil + "Keymap for tree widget browsers") + +(unless slime-browser-map + (setq slime-browser-map (make-sparse-keymap)) + (set-keymap-parent slime-browser-map widget-keymap) + (define-key slime-browser-map "q" 'bury-buffer)) + (defun slime-call-with-browser-setup (buffer package title fn) (require 'tree-widget) (switch-to-buffer buffer) @@ -7495,9 +7503,9 @@ (funcall fn) (lisp-mode) (slime-mode t) - (use-local-map widget-keymap) + (use-local-map slime-browser-map) (widget-setup)) - + ;;;; Xref browser @@ -7510,6 +7518,9 @@ append specs))) (loop for (dspec . _) in specs + ;; We can't expand FLET references so they're useless + unless (string= "(FLET" (substring dspec 0 5)) + ;;---*** Need to find and remove duplicates collect `(tree-widget :tag ,dspec :xref-type ,type :dynargs slime-expand-xrefs

2 2

[slime-devel] Daily ChangeLog diff
by lgorrie＠common-lisp.net 27 Apr '05

27 Apr '05

Index: slime/ChangeLog diff -u slime/ChangeLog:1.672 slime/ChangeLog:1.674 --- slime/ChangeLog:1.672 Thu Apr 21 14:07:42 2005 +++ slime/ChangeLog Wed Apr 27 16:54:41 2005 @@ -1,3 +1,8 @@ +2005-04-27 Helmut Eller <heller(a)common-lisp.net> + + * swank-cmucl.lisp (+header-type-symbols+): Drop the third arg to + apropos-list; it's no longer supported in recent CMUCLs. + 2005-04-21 Luke Gorrie <luke(a)synap.se> * swank.lisp (arglist-to-string): Rolled back the previous change

1 0

[slime-devel] Daily ChangeLog diff
by lgorrie＠common-lisp.net 21 Apr '05

21 Apr '05

Index: slime/ChangeLog diff -u slime/ChangeLog:1.669 slime/ChangeLog:1.672 --- slime/ChangeLog:1.669 Wed Apr 20 19:27:52 2005 +++ slime/ChangeLog Thu Apr 21 14:07:42 2005 @@ -1,3 +1,8 @@ +2005-04-21 Luke Gorrie <luke(a)synap.se> + + * swank.lisp (arglist-to-string): Rolled back the previous change + because it interferred with values appearing in parameter lists. + 2005-04-20 Luke Gorrie <luke(a)synap.se> * swank.lisp (arglist-to-string): Bind *PRINT-ESCAPE* to NIL. This @@ -8,9 +13,6 @@ `linux_no_threads_p' alien variable to decide whether to use :SPAWN. From dan_b for compatibility with new SBCLs. -2005-04-19 heller <heller@xaital> - - * /project/slime/cvsroot/slime/PROBLEMS: Warn about old kernels. 2005-04-19 Helmut Eller <heller(a)common-lisp.net> * PROBLEMS: Warn about old kernels. @@ -1417,7 +1419,7 @@ * swank-cmucl.lisp (return-from-frame): Add it. -2004-10-11 Thomas Burdick <tfb(a)OCF.Berkeley.EDU> +2004-10-11 Thomas F. Burdick <tfb(a)OCF.Berkeley.EDU> * swank-sbcl.lisp (function-definitions): Find compiler macros, too. (find-defintions, compiler-definitions)

1 0

[slime-devel] Any word on a 1.1 release?
by Peter Seibel 21 Apr '05

21 Apr '05

So I'm lagging behind my original plan vis a vis the Practical Common Lisp Lispbox distribution. (I've got preliminary versions of OS X and GNU/Linux Lispboxes up.) So now that it's mid-April is there any chance of an official 1.1 release of SLIME. To date I've been making my own "releases" out of CVS HEAD but it'd be nice to have an official one. -Peter -- Peter Seibel peter(a)gigamonkeys.com Lisp is the red pill. -- John Fraser, comp.lang.lisp

5 12

[slime-devel] Stepping in SBCL
by Svein Ove Aas 21 Apr '05

21 Apr '05

As you may have noticed, single-stepping using the step macro doesn't work with SBCL. [1] After some poking around, I figured out that the cause is a missing handler-bind, and is appaarently easily fixable - just insert the following code. (handler-bind ((sb-impl::step-condition #'sb-impl::invoke-stepper)) <code leading to a (step) form>) My problem is that I can't decide where to put it. The call stack for code run in the REPL seems to only include functions defined in swank.lisp, which is presumably the wrong place to put implementation-dependent code. What should I do?

2 1

[slime-devel] Daily ChangeLog diff
by lgorrie＠common-lisp.net 20 Apr '05

20 Apr '05

Index: slime/ChangeLog diff -u slime/ChangeLog:1.667 slime/ChangeLog:1.669 --- slime/ChangeLog:1.667 Tue Apr 19 22:24:49 2005 +++ slime/ChangeLog Wed Apr 20 19:27:52 2005 @@ -1,3 +1,13 @@ +2005-04-20 Luke Gorrie <luke(a)synap.se> + + * swank.lisp (arglist-to-string): Bind *PRINT-ESCAPE* to NIL. This + way symbols in arglists are printed as with PRINC, i.e. without + package qualifier. + + * swank-sbcl.lisp (preferred-communication-style): Use + `linux_no_threads_p' alien variable to decide whether to use + :SPAWN. From dan_b for compatibility with new SBCLs. + 2005-04-19 heller <heller@xaital> * /project/slime/cvsroot/slime/PROBLEMS: Warn about old kernels.

1 0

[slime-devel] Re: Slime and :sb-futex on recent CVS SBCLs
by Luke Gorrie 20 Apr '05

20 Apr '05

Vincent Arkesteijn <vincent(a)arkesteijn.net> writes: > Doesn't work for me. It looks like you've added a new function > communication-style, which never gets called. Changing the existing > function preferred-communication-style does work for me (on 2.4 both > with and without NPTL). See attached patch. Wow :-) thanks for catching that!

1 0

[slime-devel] Re: [Sbcl-devel] Slime and :sb-futex on recent CVS SBCLs
by dan＠telent.net 20 Apr '05

20 Apr '05

This isn't just about CVS SBCL; SB-FUTEX was removed last month before the 0.8.21 release On Tue, Apr 19, 2005 at 01:56:21PM +0200, Vincent Arkesteijn wrote: > A somewhat related problem, which will be uncovered by fixing the > one that Andreas describes, surfaces when running SLIME+SBCL on > a non-NPTL 2.4 kernel. :sb-thread is still on *features*, so > swank:*communication-style* will be set to :spawn, but calling > make-thread will then fail. Not nice. > > Possible solutions: > 1) Remove :sb-thread from *features* when running on a non-NPTL > kernel. I think this (and the consequently necessary changes to the fasl compatibility checker) is probably TRT, but not this close to a release. > 2) Add some function or variable to SBCL that can be used to check > whether threads actually work, and change SLIME to use that. * (defun communication-style () (if (and (member :sb-thread *features*) (not (sb-alien:extern-alien "linux_no_threads_p" sb-alien:boolean))) :spawn :fd-handler)) COMMUNICATION-STYLE * (communication-style) :SPAWN ... but note that this uses implementation-internal stuff and is not guaranteed to stay working. As the useful working life of any SLIME version on SBCL rarely exceeds three months anyway, that might not be such a big deal -dan

3 2

[slime-devel] Difficulty connecting to external process
by Dan Knapp 20 Apr '05

20 Apr '05

Much like in the recent thread on ssh, I'm trying to use Slime by connecting to a running Lisp. In theory, this ought to be simpler, because I'm not using ssh for it. I am not having much luck. I'm using the latest CVS version of Slime (just ran an update and repeated my test); OpenMCL 1.14.3; and the Emacs version: GNU Emacs 21.3.50.1 (powerpc-apple-darwin7.3.0) of 2004-05-08 on iBook.local In the host process, I load Slime through asdf, then I start it by invoking (setq swank:*use-dedicated-output-stream* nil) (swank:create-server :port 9090 :dont-close t :style nil) The *use-dedicated-output-stream* is there because an old thread on this list advised it. I have also tried without the :style parameter, not that that should have any relevance. Then in Emacs, I do M-x slime-connect RET 127.0.0.1 RET 9090 RET. Emacs says this: Connecting to Swank on port 9090.. net-read error: (end-of-file) error in process-filter: byte-code: PANIC! error in process-filter: PANIC! Lisp says this (I have it set up to dump into syslog): ;; Connection to Emacs lost. ;; [Short read: length=3158064 pos=58] If I try to connect again, I get the same error messages on both ends, so at least I know the loop isn't falling through... Both processes happen to be on the same computer, but are using different copies of the same version of Slime (one of them is inside a .app bundle...), in case that's relevant. Please advise? -- Dan Knapp

2 2