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

[slime-devel] Multiple REPLs
by Sebastian Tennant 12 Mar '10

12 Mar '10

Hi all, In buffer *Slime Connections* the marked entry is the 'default' REPL. What's the significance of 'default' status? Is it that in the case of multiple REPLS on a single lisp instance, it determines which REPL receives the lisp output? And if the above thinking is correct, does it have any significance at all when each REPL is on a separate lisp instance? 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] just testing
by Helmut Eller 12 Mar '10

12 Mar '10

nothing to see

1 0

[slime-devel] Daily ChangeLog diff
by Helmut Eller 12 Mar '10

12 Mar '10

Index: slime/ChangeLog diff -u slime/ChangeLog:1.2033 slime/ChangeLog:1.2035 --- slime/ChangeLog:1.2033 Wed Mar 10 06:30:22 2010 +++ slime/ChangeLog Thu Mar 11 04:05:50 2010 @@ -1,5 +1,20 @@ 2010-03-10 Tobias C. Rittweiler <tcr(a)freebits.de> + * swank.lisp (signal-interrupt): Removed. + (interrupt-worker-thread): Slurp in definition of + signal-interrupt. No need for invoke-or-queue-interrupt twice in + case we do not use threads. Thus micro-prettification of + backtraces. + +2010-03-10 Tobias C. Rittweiler <tcr(a)freebits.de> + + * swank-ecl.lisp (source-location): Move call to + TRANSLATE-LOGICAL-PATHNAME from here into MAKE-FILE-LOCATION + because locations-via-annotations may now also involve logical + pathnames. + +2010-03-10 Tobias C. Rittweiler <tcr(a)freebits.de> + * swank-ecl.lisp (*original-sigint-handler*) (install-sigint-handler): Deleted; we directly implement call-with-user-break-handler instead.

1 0

[slime-devel] Bug Report: SB-KERNEL:SIMPLE-PACKAGE-ERROR not invoking the debugger
by Russ Tyndall 11 Mar '10

11 Mar '10

Slime Version: 3/10 from git mirror of CVS SBCL 1.0.36 GNU Emacs 22.1.1 and 23.x.x (checked on two different versions). Using the latest version of slime, executing the following code in the slime repl results in a totally locked lisp that can't be quit or interacted with via slime. There is also nothing of interest in the backing SBCL terminal. (make-package :short-name) (make-package :long-name :nicknames (list :short-name)) Executing these same forms in the sbcl command line repl correctly results in a debugger asking for me to abort being stupid. The last slime event we see is: (:emacs-rex (swank:listener-eval "(make-package :long-name :nicknames (list :short-name))\n") "COMMON-LISP-USER" :repl-thread 6) This is my setup: (slime-setup '(slime-fancy slime-asdf slime-tramp slime-indentation)) Thanks for the great software, SLIME is a pleasure to use, Russ Tyndall

2 1

[slime-devel] Daily ChangeLog diff
by Helmut Eller 11 Mar '10

11 Mar '10

Index: slime/ChangeLog diff -u slime/ChangeLog:1.2032 slime/ChangeLog:1.2033 --- slime/ChangeLog:1.2032 Tue Mar 9 19:02:53 2010 +++ slime/ChangeLog Wed Mar 10 06:30:22 2010 @@ -1,3 +1,12 @@ +2010-03-10 Tobias C. Rittweiler <tcr(a)freebits.de> + + * swank-ecl.lisp (*original-sigint-handler*) + (install-sigint-handler): Deleted; we directly implement + call-with-user-break-handler instead. + (call-with-user-break-handler): New. Correctly interrupt main + thread instead of newly spawned handle-signal thread on SIGINT. + (make-interrupt-handler): New helper. + 2010-03-09 Stas Boukarev <stassats(a)gmail.com> * swank-ccl.lisp (emacs-inspect function): Print closed over variables @@ -62,7 +71,7 @@ it boils down to SIGNAL-INTERRUPT which uses USE-THREADS-P which needs a connection. (install-fd-handler): Adapted accordingly. - (simple-serve-event): Adapted accordingly. Additionally, remove + (simple-serve-requests): Adapted accordingly. Additionally, remove superfluous WITH-SWANK-PROTOCOL-HANDLER as that's established by WITH-CONNECTION already. (simple-repl): Show "abort inferior lisp" restart only if not a

1 0

[slime-devel] sldb-initial-restart-limit
by Robert Goldman 10 Mar '10

10 Mar '10

The current definition of this custom is as follows: (defcustom sldb-initial-restart-limit 6 "Maximum number of restarts to display initially." :group 'slime-debugger :type 'integer) Reading the way this is used, I believe this could be changed to (defcustom sldb-initial-restart-limit 6 "Maximum number of restarts to display initially. Will always display all restarts if NIL." :group 'slime-debugger :type '(choice integer (const :tag "Show all" nil)) I don't venture to offer this as a patch because I'm not an expert with customize. HtH, r

1 0

[slime-devel] slime-compile-file gets error in ACL 8.2
by Takuo Watanabe 10 Mar '10

10 Mar '10

Dear Slime Developers I am using Slime with the following environment: Slime 2010-03-01 (CVS head) Emacs 22.3.1 (i386-apple-darwin9.8.0, Carbon Version 1.6.0) Allegro CL Professional 8.2 (allegro-8.2a-ics-macosx-x86) Mac OS X (10.6.2), MacBook After upgrading Allegro to 8.2, slime-compile-file (also slime-compile-and-load-file) gets errors like this: ------------------------------------------------------------ `(6 6)' is not of the expected type `NUMBER' [Condition of type TYPE-ERROR] ------------------------------------------------------------ The backtrace of the error looks like: ------------------------------------------------------------ Backtrace: 0: (ERROR TYPE-ERROR :DATUM (6 6) :EXPECTED-TYPE NUMBER ...) 1: (SWANK-BACKEND::LOCATION-FOR-WARNING #<SIMPLE-WARNING @ #x20a331c2>) Locals: CONDITION = #<SIMPLE-WARNING @ #x20a331c2> SWANK-BACKEND::POS = (6 6) NIL = (:FILE "/Users/takuo/foo.lisp") NIL = #<STANDARD-CLASS READER-ERROR> SWANK-BACKEND::LOC = (#P"/Users/takuo/foo.lisp" 6 6) SWANK-BACKEND::FILE = #P"/Users/takuo/foo.lisp" 2: (SWANK-BACKEND::HANDLE-COMPILER-WARNING #<SIMPLE-WARNING @ #x20a331c2>) 3: (SIGNAL #<SIMPLE-WARNING @ #x20a331c2>) ------------------------------------------------------------ The source of this error seems to be the change of the format of condition location (i.e., the value of the :loc property provided with the excl::plist slot in a warning object). For example, in the previous ACL (8.1), the value of :loc property looks like this. ------------------------------------------------------------ CL-USER> (handler-case (compile-file "foo.lisp") (warning (w) (getf (slot-value w 'excl::plist) :loc))) (#P"foo.lisp" . 6) ------------------------------------------------------------ The car part of the value is the file name and the cdr is the warning position in the file. After upgrading to 8.2, the value of the above expression becomes (#P"foo.lisp" 6 6). I don't know about the difference among these two numbers. Anyway, the function location-for-warning in swank-allegro.lisp assumes that the cdr part of the value should be a number. The (kludgey) patch I applied is the following. I'm not sure this is really a solution to the problem. ------------------------------------------------------------ --- swank-allegro.lisp 2010-03-02 12:25:46.000000000 +0900 +++ swank-allegro.lisp.orig 2010-03-02 12:25:24.000000000 +0900 @@ -279,7 +279,7 @@ (destructuring-bind (file . pos) loc (make-location (list :file (namestring (truename file))) - (list :position (1+ (if (listp pos) (car pos) pos)))))) + (list :position (1+ pos))))) (t (make-error-location "No error location available."))))) ------------------------------------------------------------ Sorry, I didn't check any other platforms (Linux, Windows, etc.). I only have a license for the OS X edition. Cheers Takuo Watanabe

3 2

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

10 Mar '10

Index: slime/ChangeLog diff -u slime/ChangeLog:1.2029 slime/ChangeLog:1.2032 --- slime/ChangeLog:1.2029 Tue Mar 9 03:15:05 2010 +++ slime/ChangeLog Tue Mar 9 19:02:53 2010 @@ -1,3 +1,12 @@ +2010-03-09 Stas Boukarev <stassats(a)gmail.com> + + * swank-ccl.lisp (emacs-inspect function): Print closed over variables + in case of closure. + +2010-03-09 Stas Boukarev <stassats(a)gmail.com> + + * slime.el (slime-parse-context): Add defclass. + 2010-03-09 Helmut Eller <heller(a)common-lisp.net> Some more fixes for Allegro @@ -10,6 +19,8 @@ vs. buffer issue. (ldb-code-to-src-loc): Don't use *temp-file-map* before it is declared. + (pc-source-location): Be a bit more fuzzy when searching the + code-location for a pc. 2010-03-08 Stas Boukarev <stassats(a)gmail.com> Index: slime/contrib/ChangeLog diff -u slime/contrib/ChangeLog:1.353 slime/contrib/ChangeLog:1.356 --- slime/contrib/ChangeLog:1.353 Mon Mar 8 11:35:06 2010 +++ slime/contrib/ChangeLog Tue Mar 9 18:26:23 2010 @@ -1,3 +1,19 @@ +2010-03-09 Stas Boukarev <stassats(a)gmail.com> + + * swank-arglists.lisp (arglist-ref): Don't error if a &key name isn't + a :keyword symbol. And also handle non-:keyword keyword parameters. + +2010-03-09 Stas Boukarev <stassats(a)gmail.com> + + * swank-presentations.lisp (inspect-presentation): Throw an error when + trying to access unrecorded object. + (lookup-presented-object-or-lose): Rename from `get-repl-result'. + +2010-03-09 Stas Boukarev <stassats(a)gmail.com> + + * slime-repl.el (slime-call-defun): When on defclass insert + (make-instance 'name). + 2010-03-08 Stas Boukarev <stassats(a)gmail.com> * swank-fancy-inspector.lisp (all-slots-for-inspector): Sort class

1 0

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

09 Mar '10

Index: slime/ChangeLog diff -u slime/ChangeLog:1.2021 slime/ChangeLog:1.2029 --- slime/ChangeLog:1.2021 Sun Mar 7 11:22:17 2010 +++ slime/ChangeLog Tue Mar 9 03:15:05 2010 @@ -1,3 +1,86 @@ +2010-03-09 Helmut Eller <heller(a)common-lisp.net> + + Some more fixes for Allegro + + * swank-allegro.lisp (function-source-location): Use + xref::object-to-function-name which seems to work better for some + cases. + (fspec-definition-locations): For :top-level-forms return a list + of ((fspec loc)) not just (fspec loc). Also deal with the file + vs. buffer issue. + (ldb-code-to-src-loc): Don't use *temp-file-map* before it is + declared. + +2010-03-08 Stas Boukarev <stassats(a)gmail.com> + + * slime.el (slime-extract-context): Add defvar and defparameter. + +2010-03-08 Helmut Eller <heller(a)common-lisp.net> + + Fix some of the brokeness in the last change. + + * swank-allegro.lisp (frame-source-location): Deal with frames for + undefined functions better. + (ldb-code-to-src-loc): Handle temp-files properly. + +2010-03-08 Helmut Eller <heller(a)common-lisp.net> + + Try to use source-level debugging features in Allegro 8.2 + + * swank-allegro.lisp (disassemble-frame): Use undocumented + debugger::dyn-fd-analyze to figure out the PC and display it. + (pc-source-location, ldb-code-to-src-loc, longest-common-prefix) + (source-paths-of): New functions. + (frame-source-location): Use pc-source-location. Still far from + optimal since Allegro rarely records source regions and anonymous + functions don't seem to carry source level debug-info at all. + (*temp-file-map*, buffer-or-file-location, find-fspec-location): + Use a table to map temp-file names back to Emacs buffers instead + of putting an eval-when-compile form in the source. The + eval-when-compile form messed up source positions. + (*temp-file-header-end-position*, find-definition-in-buffer): + Deleted. + (compile-from-temp-file): Bind excl:*load-source-debug-info* and + compiler:save-source-level-debug-info-switch so that Allegro + doesn't try to load debug-info from deleted files. Also put + the filename in *temp-file-map*. + +2010-03-08 Tobias C. Rittweiler <tcr(a)freebits.de> + + * swank.lisp (dispatch-interrupt-event): Take a connection because + it boils down to SIGNAL-INTERRUPT which uses USE-THREADS-P which + needs a connection. + (install-fd-handler): Adapted accordingly. + (simple-serve-event): Adapted accordingly. Additionally, remove + superfluous WITH-SWANK-PROTOCOL-HANDLER as that's established by + WITH-CONNECTION already. + (simple-repl): Show "abort inferior lisp" restart only if not a + more appropriate "abort some REX" restart is available. Also make + sure to return in case of END-OF-FILE, otherwise there's an + infinite loop where we end up in the debugger again and again + until the user eventually selects close-connection restart himself. + (make-repl-input-stream): Use WITH-TOP-LEVEL-RESTART so + `sldb-quit' can be used in SLDB. + +2010-03-08 Tobias C. Rittweiler <tcr(a)freebits.de> + + * swank.lisp (close-connection): Include initially passed + coding-system in debugging output. + +2010-03-08 Tobias C. Rittweiler <tcr(a)freebits.de> + + Make swank:connection-info include information about initially + passed coding-system, and the resulting external-format of the + socket. Debugging aid. + + * swank.lisp (connection.external-format): New function. + (start-server, create-server): Pass down coding-system, not + external-format. + (setup-server): Pass down both, coding-system and external-format. + (serve-connection): Ditto. + (create-connection): Set coding-system slot of CONNECTION. + (connection-info): Include coding-system and external-format. + 2010-03-07 Helmut Eller <heller(a)common-lisp.net> * swank-allegro.lisp (count-cr): Deleted. No longer used. Index: slime/contrib/ChangeLog diff -u slime/contrib/ChangeLog:1.351 slime/contrib/ChangeLog:1.353 --- slime/contrib/ChangeLog:1.351 Sun Mar 7 12:04:00 2010 +++ slime/contrib/ChangeLog Mon Mar 8 11:35:06 2010 @@ -1,7 +1,16 @@ +2010-03-08 Stas Boukarev <stassats(a)gmail.com> + + * swank-fancy-inspector.lisp (all-slots-for-inspector): Sort class + names when grouping by inheritance the same way as slots are sorted. + +2010-03-08 Stas Boukarev <stassats(a)gmail.com> + + * slime-repl.el (slime-call-defun): Work also on defvar/defparameter. + 2010-03-07 Stas Boukarev <stassats(a)gmail.com> - * swank-fancy-inspector.lisp (stable-sort-by-inheritance): Remove copy-seq, - unnecessarily put in the previous commit. + * swank-fancy-inspector.lisp (stable-sort-by-inheritance): Remove + copy-seq, unnecessarily put in the previous commit. 2010-03-07 Stas Boukarev <stassats(a)gmail.com>

2 1

[slime-devel] communication-style nil and interrupts
by Tobias C. Rittweiler 08 Mar '10

08 Mar '10

Currently, we use NIL on ECL because on one hand while it has threads, some parts are not yet thread-safe, and on the other hand, I do not fully understand Slime's :fd-handler architecture. :FD-HANDLER seems to be tied to an I/O-implies-serve-event architecture like SBCL (and CMUCL?) have. OTOH, I just took a courtesy look at WAIT-FOR-INPUT in swank-cmucl and see *INTERRUPT-QUEUED-HANDLER*. What's the purpose of that? Why does WAIT-FOR-INPUT write to that pipe? :-) ECL does have a select()-based serve-event, but it's not implicitly called. Where in the backend should it call serve-event? The problem with NIL is that if ECL was built with threads enabled, it executes the SIGINT handler function inside a new thread. But we actually want to interrupt the REPL thread -- however SIGNAL-INTERRUPT just ignores its thread argument in case :SPAWN is not used. So I do not see an obvious way out. That said, the problem won't be solved by implementing :FD-HANDLER for ECL either. Thoughts? -T.

2 1