In current CVS HEAD using Carbon Emacs from CVS (2004-10-29) and OpenMCL I'm thrown immediately into the debugger when I do M-x slime. a does not get me out:
In the *sldb* buffer:
No applicable method for args: (#<SWANK-BACKEND::SLIME-INPUT-STREAM #x649EA86>) to #<STANDARD-GENERIC-FUNCTION STREAM-FORCE-OUTPUT #x60C1146> [Condition of type SIMPLE-ERROR]
Restarts: 0: [ABORT] #<RESTART ABORT #xC8F76> 1: [ABORT-BREAK] #<RESTART ABORT-BREAK #xC8F9E>
Backtrace: 0: (#<CCL::STANDARD-KERNEL-METHOD NO-APPLICABLE-METHOD (T)> #<STANDARD-GENERIC-FUNCTION STREAM-FORCE-OUTPUT #x60C1146>) 1: (CCL::EVENT-POLL) 2: (CCL::HOUSEKEEPING) 3: (#<Anonymous Function #x60F911E>)
In the *slime-events* buffer:
(:emacs-rex (swank:connection-info) nil t 1) (:open-dedicated-output-stream 52683) (:indentation-update (("with-lock-grabbed-maybe" . 1) ("with-ioblock-input-locked" . 1) ("with-package-write-lock" . 1) ("with-slot-values" . 2) ("with-eagain" . 2) ("without-gcing" . 0) ("with-restart" . 1) ("with-auxiliary-foreign-types" . 0) ("with-ppc-local-vinsn-macros" . 1) ("with-package-read-lock" . 1) ("while" . 1) ("%ilogior" . 1) ("with-call-method-context" . 1) ("with-dll-node-freelist" . 1) ("with-open-dir" . 1) ("ignoring-without-interrupts" . 0) ("with-area-macptr" . 1) ("with-standard-abort-handling" . 1) ("once-only" . 1) ("def-kernel-restart" . 3) ...)) (:return (:ok (25784 "OpenMCL" "openmcl" (:cltl2 :mcl2.0 :primary-classes :ccl :ccl-2 :ccl-3 :ccl-4 :coral :common-lisp :mcl :openmcl :ansi-cl :processes :unix :openmcl-native-threads :openmcl-partial-mop :mcl-common-mop-subset :openmcl-mop-2 :interfaces-2 :powerpc ...) :spawn)) 1) (:debug 0 1 ("No applicable method for args:\n (#<SWANK-BACKEND::SLIME-INPUT-STREAM #x649EA86>)\n to #<STANDARD-GENERIC-FUNCTION STREAM-FORCE-OUTPUT #x60C1146>" " [Condition of type SIMPLE-ERROR]" nil nil) (("ABORT" "#<RESTART ABORT #xC8F76>") ("ABORT-BREAK" "#<RESTART ABORT-BREAK #xC8F9E>")) ((0 "(#<CCL::STANDARD-KERNEL-METHOD NO-APPLICABLE-METHOD (T)> #<STANDARD-GENERIC-FUNCTION STREAM-FORCE-OUTPUT #x60C1146>)") (1 "(CCL::EVENT-POLL)") (2 "(CCL::HOUSEKEEPING)") (3 "(#<Anonymous Function #x60F911E>)"))) (:debug-activate 0 1) (:emacs-rex (swank:sldb-abort) nil 0 2) (:return (:abort) 2) (:debug-return 0 1 nil) (:debug 0 1 ("No applicable method for args:\n (#<SWANK-BACKEND::SLIME-INPUT-STREAM #x649EA86>)\n to #<STANDARD-GENERIC-FUNCTION STREAM-FORCE-OUTPUT #x60C1146>" " [Condition of type SIMPLE-ERROR]" nil nil) (("ABORT" "#<RESTART ABORT #xC8F76>") ("ABORT-BREAK" "#<RESTART ABORT-BREAK #xC8F9E>")) ((0 "(#<CCL::STANDARD-KERNEL-METHOD NO-APPLICABLE-METHOD (T)> #<STANDARD-GENERIC-FUNCTION STREAM-FORCE-OUTPUT #x60C1146>)") (1 "(CCL::EVENT-POLL)") (2 "(CCL::HOUSEKEEPING)") (3 "(#<Anonymous Function #x60F911E>)"))) (:debug-activate 0 1) (:emacs-rex (swank:sldb-abort) nil 0 3) (:return (:abort) 3) (:debug-return 0 1 nil) (:debug 0 1 ("No applicable method for args:\n (#<SWANK-BACKEND::SLIME-INPUT-STREAM #x649EA86>)\n to #<STANDARD-GENERIC-FUNCTION STREAM-FORCE-OUTPUT #x60C1146>" " [Condition of type SIMPLE-ERROR]" nil nil) (("ABORT" "#<RESTART ABORT #xC8F76>") ("ABORT-BREAK" "#<RESTART ABORT-BREAK #xC8F9E>")) ((0 "(#<CCL::STANDARD-KERNEL-METHOD NO-APPLICABLE-METHOD (T)> #<STANDARD-GENERIC-FUNCTION STREAM-FORCE-OUTPUT #x60C1146>)") (1 "(CCL::EVENT-POLL)") (2 "(CCL::HOUSEKEEPING)") (3 "(#<Anonymous Function #x60F911E>)"))) (:debug-activate 0 1) (:emacs-rex (swank:sldb-abort) nil 0 4) (:return (:abort) 4) (:debug-return 0 1 nil) (:debug 0 1 ("No applicable method for args:\n (#<SWANK-BACKEND::SLIME-INPUT-STREAM #x649EA86>)\n to #<STANDARD-GENERIC-FUNCTION STREAM-FORCE-OUTPUT #x60C1146>" " [Condition of type SIMPLE-ERROR]" nil nil) (("ABORT" "#<RESTART ABORT #xC8F76>") ("ABORT-BREAK" "#<RESTART ABORT-BREAK #xC8F9E>")) ((0 "(#<CCL::STANDARD-KERNEL-METHOD NO-APPLICABLE-METHOD (T)> #<STANDARD-GENERIC-FUNCTION STREAM-FORCE-OUTPUT #x60C1146>)") (1 "(CCL::EVENT-POLL)") (2 "(CCL::HOUSEKEEPING)") (3 "(#<Anonymous Function #x60F911E>)"))) (:debug-activate 0 1) (:emacs-rex (swank:sldb-abort) nil 0 5) (:return (:abort) 5) (:debug-return 0 1 nil) (:debug 0 1 ("No applicable method for args:\n (#<SWANK-BACKEND::SLIME-INPUT-STREAM #x649EA86>)\n to #<STANDARD-GENERIC-FUNCTION STREAM-FORCE-OUTPUT #x60C1146>" " [Condition of type SIMPLE-ERROR]" nil nil) (("ABORT" "#<RESTART ABORT #xC8F76>") ("ABORT-BREAK" "#<RESTART ABORT-BREAK #xC8F9E>")) ((0 "(#<CCL::STANDARD-KERNEL-METHOD NO-APPLICABLE-METHOD (T)> #<STANDARD-GENERIC-FUNCTION STREAM-FORCE-OUTPUT #x60C1146>)") (1 "(CCL::EVENT-POLL)") (2 "(CCL::HOUSEKEEPING)") (3 "(#<Anonymous Function #x60F911E>)"))) (:debug-activate 0 1) (:emacs-rex (swank:sldb-abort) nil 0 6) (:return (:abort) 6) (:debug-return 0 1 nil) (:debug 0 1 ("No applicable method for args:\n (#<SWANK-BACKEND::SLIME-INPUT-STREAM #x649EA86>)\n to #<STANDARD-GENERIC-FUNCTION STREAM-FORCE-OUTPUT #x60C1146>" " [Condition of type SIMPLE-ERROR]" nil nil) (("ABORT" "#<RESTART ABORT #xC8F76>") ("ABORT-BREAK" "#<RESTART ABORT-BREAK #xC8F9E>")) ((0 "(#<CCL::STANDARD-KERNEL-METHOD NO-APPLICABLE-METHOD (T)> #<STANDARD-GENERIC-FUNCTION STREAM-FORCE-OUTPUT #x60C1146>)") (1 "(CCL::EVENT-POLL)") (2 "(CCL::HOUSEKEEPING)") (3 "(#<Anonymous Function #x60F911E>)"))) (:debug-activate 0 1) (:emacs-rex (swank:backtrace 4 43) nil 0 7) (:return (:ok nil) 7) (:debug-activate 0 1)
David Steuber david@david-steuber.com writes:
In current CVS HEAD using Carbon Emacs from CVS (2004-10-29) and OpenMCL I'm thrown immediately into the debugger when I do M-x slime. a does not get me out:
try this patch:
--- swank-openmcl.lisp 17 Nov 2004 15:17:26 -0000 1.89 +++ swank-openmcl.lisp 19 Nov 2004 15:49:26 -0000 @@ -143,7 +143,7 @@ (setq ccl::*interactive-abort-process* ccl::*current-process*))
(defimplementation make-stream-interactive (stream) - (push stream ccl::*auto-flush-streams*)) + nil)
;;; Unix signals
the problem is that stream is a gray-stream, and openmcl doesn't take well to having a gray-stream on ccl::*auto-flush-streams*. i don't know what _should_ be done though (there must be something better than doing nothing).
"Marco Baringer" mb@bese.it writes:
the problem is that stream is a gray-stream, and openmcl doesn't take well to having a gray-stream on ccl::*auto-flush-streams*. i don't know what _should_ be done though (there must be something better than doing nothing).
I think the problem is that force-output is called on an input-stream. Perhaps we could try to only push output-streams onto ccl::*auto-flush-streams*.
Helmut.
On Nov 19, 2004, at 11:19 AM, Helmut Eller wrote:
"Marco Baringer" mb@bese.it writes:
the problem is that stream is a gray-stream, and openmcl doesn't take well to having a gray-stream on ccl::*auto-flush-streams*. i don't know what _should_ be done though (there must be something better than doing nothing).
I think the problem is that force-output is called on an input-stream. Perhaps we could try to only push output-streams onto ccl::*auto-flush-streams*.
FWIW, I had done a 'cvs up -D 2004-11-12' to get a working version. I just did a 'cvs up -A' to try Marco Baringer's patch against the current CVS version. I wanted to see if it was still the same and did M-x slime in a fresh Emacs session and got this in the *inferior-lisp* buffer:
Welcome to OpenMCL Version (Beta: Darwin) 0.14.2-p1! ? ;Loading #P"/Users/david/usr/src/slime/swank-loader.lisp"... ;; Compiling /Users/david/usr/src/slime/swank-backend.lisp... ;Loading #P"/Users/david/.slime/fasl/openmcl/swank-backend.dfsl"... ;; Compiling /Users/david/usr/src/slime/nregex.lisp... ;Loading #P"/Users/david/.slime/fasl/openmcl/nregex.dfsl"... ;; Compiling /Users/david/usr/src/slime/metering.lisp... ;Loading #P"/Users/david/.slime/fasl/openmcl/metering.dfsl"... ;; Compiling /Users/david/usr/src/slime/swank-openmcl.lisp... ;Loading #P"/Users/david/.slime/fasl/openmcl/swank-openmcl.dfsl"... ;; Compiling /Users/david/usr/src/slime/swank-gray.lisp... ;Loading #P"/Users/david/.slime/fasl/openmcl/swank-gray.dfsl"... ;; Compiling /Users/david/usr/src/slime/swank.lisp... ;Compiler warnings for "/Users/david/usr/src/slime/swank.lisp" : ; Unused lexical variable SOCKET-IO, in SIMPLE-SERVE-REQUESTS. ;Loading #P"/Users/david/.slime/fasl/openmcl/swank.dfsl"... ; Warning: These Swank interfaces are unimplemented: ; (ACTIVATE-STEPPING ADD-FD-HANDLER ADD-SIGIO-HANDLER MACROEXPAND-ALL REMOVE-FD-HANDLERS REMOVE-SIGIO-HANDLERS SLDB-BREAK-AT-START SLDB-BREAK-ON-RETURN) ; While executing: SWANK-BACKEND::WARN-UNIMPLEMENTED-INTERFACES ;Compiler warnings for "/Users/david/usr/src/slime/swank-openmcl.lisp" : ; Undefined function #:EQL-SPECIALIZER-OBJECT, in SWANK-BACKEND::SPECIALIZER-NAME. ; Undefined function #:CLASS-NAME, in SWANK-BACKEND::SPECIALIZER-NAME. #P"/Users/david/usr/src/slime/swank-loader.lisp" ? ;; Swank started at port: 52924.
Error in process listener(1): value :EXTERNAL-FORMAT is not of the
expected type (MEMBER :SPAWN :FD-HANDLER :SIGIO NIL).
While executing: SWANK::SETUP-SERVER Type :POP to abort.
Type :? for other options. 1 >
David Steuber david@david-steuber.com writes:
FWIW, I had done a 'cvs up -D 2004-11-12' to get a working version. I just did a 'cvs up -A' to try Marco Baringer's patch against the current CVS version. I wanted to see if it was still the same and did M-x slime in a fresh Emacs session and got this in the *inferior-lisp* buffer:
It's possible that you upgraded your version just in a moment when the cvs version was in a inconsistent state. In the newest version, START-SERVER in swank.lisp has only keyword arguments and no optional arguments.
Please remove all fasl files in ~/.slime, cvs up, and retry.
Helmut.
On Nov 19, 2004, at 2:54 PM, Helmut Eller wrote:
David Steuber david@david-steuber.com writes:
FWIW, I had done a 'cvs up -D 2004-11-12' to get a working version. I just did a 'cvs up -A' to try Marco Baringer's patch against the current CVS version. I wanted to see if it was still the same and did M-x slime in a fresh Emacs session and got this in the *inferior-lisp* buffer:
It's possible that you upgraded your version just in a moment when the cvs version was in a inconsistent state. In the newest version, START-SERVER in swank.lisp has only keyword arguments and no optional arguments.
Please remove all fasl files in ~/.slime, cvs up, and retry.
That fixed it. Rather, that got me back to the original state with my previous bug report. I was successful applying Marco Baringer's patch. That seems to have fixed the problem with the debugger popping up.
Thanks, Marco & Helmut.
David Steuber david@david-steuber.com writes:
That fixed it. Rather, that got me back to the original state with my previous bug report. I was successful applying Marco Baringer's patch. That seems to have fixed the problem with the debugger popping up.
i just commited a patch whcih puts only output-streams on ccl::*auto-flush-streams*, this is probably the right thing to do.
any openmcl users (that's you david :)) who applied the patch i psoted to the list will have conflicts when updating.
On Nov 20, 2004, at 7:10 AM, Marco Baringer wrote:
David Steuber david@david-steuber.com writes:
That fixed it. Rather, that got me back to the original state with my previous bug report. I was successful applying Marco Baringer's patch. That seems to have fixed the problem with the debugger popping up.
i just commited a patch whcih puts only output-streams on ccl::*auto-flush-streams*, this is probably the right thing to do.
any openmcl users (that's you david :)) who applied the patch i psoted to the list will have conflicts when updating.
Am I really the only one? Talk about niche!
Blowing away swank-openmcl.lisp before the update avoided the conflict. It looks like it works.
-
David Steuber -
Helmut Eller -
Marco Baringer