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

[slime-devel] Problems starting SLIME
by Stelian Ionescu 12 Aug '08

12 Aug '08

If I save an image containg latest SWANK and try to start SLIME with it, I get an error. OTOH, if I start SLIME with an image that doesn't contain SWANK(loading it on the spot), all works well. I've traced back the origin of this problem to the commit of Sun Aug 3 18:23:10 entitled "Add some flow-control." with CLISP: [1]> ;; Loading file /usr/share/emacs/site-lisp/slime/swank-loader.lisp ... 0 errors, 0 warnings ;; Loaded file /usr/share/emacs/site-lisp/slime/swank-loader.lisp *** - WRITE-CHAR on #<CLOSED IO TERMINAL-STREAM> is illegal The following restarts are available: ABORT :R1 Abort main loop Break 1 [2]> :bt <1/181> #<SYSTEM-FUNCTION SHOW-STACK> 3 <2/174> #<COMPILED-FUNCTION SYSTEM::PRINT-BACKTRACE> <3/168> #<COMPILED-FUNCTION SYSTEM::DEBUG-BACKTRACE> <4/160> #<SYSTEM-FUNCTION SYSTEM::READ-EVAL-PRINT> 2 <5/157> #<COMPILED-FUNCTION SYSTEM::BREAK-LOOP-2-3> <6/153> #<SYSTEM-FUNCTION SYSTEM::SAME-ENV-AS> 2 <7/139> #<COMPILED-FUNCTION SYSTEM::BREAK-LOOP-2> <8/137> #<SYSTEM-FUNCTION SYSTEM::DRIVER> <9/97> #<COMPILED-FUNCTION SYSTEM::BREAK-LOOP> <10/94> #<SYSTEM-FUNCTION INVOKE-DEBUGGER> 1 <11/84> #<SYSTEM-FUNCTION WRITE-STRING> <12/79> #<COMPILED-FUNCTION SWANK::SIMPLE-ANNOUNCE-FUNCTION-1> <13/71> #<COMPILED-FUNCTION SYSTEM::FORMAT-APPLY> <14/66> #<COMPILED-FUNCTION FORMAT> <15/62> #<COMPILED-FUNCTION SWANK::SIMPLE-ANNOUNCE-FUNCTION> <16/60> #<COMPILED-FUNCTION SWANK::ANNOUNCE-SERVER-PORT> <17/57> #<COMPILED-FUNCTION SWANK::START-SERVER-1> <18/51> #<COMPILED-FUNCTION SWANK::SETUP-SERVER> <19/44> #<COMPILED-FUNCTION SWANK:START-SERVER> <20/43> #<SYSTEM-FUNCTION FUNCALL> 4 [39] EVAL frame for form (FUNCALL (READ-FROM-STRING "swank:start-server") "/tmp/.private/hechee/slime.3239193" :CODING-SYSTEM "utf-8-unix") and with SBCL: S/HU[1]> ; loading #P"/usr/share/emacs/site-lisp/slime/swank-loader.lisp" debugger invoked on a SB-SYS:MEMORY-FAULT-ERROR in thread #<THREAD "initial thread" RUNNING {B6209F9}>: Unhandled memory fault at #x-58A13000. Type HELP for debugger help, or (SB-EXT:QUIT) to exit from SBCL. restarts (invokable by number or by possibly-abbreviated name): 0: [ABORT] Exit debugger, returning to top level. (SB-SYS:MEMORY-FAULT-ERROR) 0] backtrace 0: (SB-SYS:MEMORY-FAULT-ERROR) 1: (SB-SYS:MEMORY-FAULT-ERROR)[:EXTERNAL] 2: ("foreign function: #x8063A1C") 3: ("foreign function: #x8063C70") 4: (SB-IMPL::OUTPUT-BYTES/UTF-8 #<SB-SYS:FD-STREAM for "standard error" {9143371}> ";; Swank started at port: " NIL 0 26) 5: (SB-IMPL::FD-SOUT #<SB-SYS:FD-STREAM for "standard error" {9143371}> ";; Swank started at port: " 0 26) 6: (SB-IMPL::%WRITE-STRING ";; Swank started at port: " #<SB-SYS:FD-STREAM for "standard error" {9143371}> 0 26) 7: (WRITE-STRING ";; Swank started at port: " #<SB-SYS:FD-STREAM for "standard error" {9143371}> :START 0 :END NIL) 8: (SB-FORMAT::INTERPRET-DIRECTIVE-LIST #<SB-SYS:FD-STREAM for "standard error" {9143371}> (";; Swank started at port: " #<~D> "." #<~%>) (39770) (39770)) 9: (SB-FORMAT::%FORMAT #<SB-SYS:FD-STREAM for "standard error" {9143371}> "~&;; Swank started at port: ~D.~%" (39770) (39770)) 10: (FORMAT #<SB-SYS:FD-STREAM for "standard error" {9143371}> "~&;; Swank started at port: ~D.~%")[:EXTERNAL] 11: (SWANK::SIMPLE-ANNOUNCE-FUNCTION 39770) 12: (SWANK::ANNOUNCE-SERVER-PORT "/tmp/.private/hechee/slime.3239193" 39770) 13: ((LAMBDA (SWANK::PORT)) 39770) 14: (SWANK::SETUP-SERVER 0 #<CLOSURE (LAMBDA (SWANK::PORT)) {B6F6B75}> :SPAWN NIL :UTF-8) 15: (SWANK:START-SERVER "/tmp/.private/hechee/slime.3239193")[:EXTERNAL] 16: (SB-INT:SIMPLE-EVAL-IN-LEXENV (FUNCALL (READ-FROM-STRING "swank:start-server") "/tmp/.private/hechee/slime.3239193" :CODING-SYSTEM "utf-8-unix") #<NULL-LEXENV>) -- Stelian Ionescu a.k.a. fe[nl]ix Quidquid latine dictum sit, altum videtur.

3 4

[slime-devel] swank-sbcl.lisp:1.213 breaks on windows
by Bart Botta 12 Aug '08

12 Aug '08

Running slime from CVS on sbcl 1.0.19/win32 ends up in the debugger with The variable SB-UNIX:SIGINT is unbound. [Condition of type UNBOUND-VARIABLE] Disabling the (defimplementation install-sigint-handler ...) added in 1.213 with #-win32 seems to fix the problem, no idea if that is the correct solution though. -b-

2 1

[slime-devel] Daily ChangeLog diff
by mbaringer＠common-lisp.net 12 Aug '08

12 Aug '08

Index: slime/ChangeLog diff -u slime/ChangeLog:1.1428 slime/ChangeLog:1.1442 --- slime/ChangeLog:1.1428 Sat Aug 9 15:57:37 2008 +++ slime/ChangeLog Mon Aug 11 13:41:59 2008 @@ -1,3 +1,69 @@ +2008-08-11 Helmut Eller <heller(a)common-lisp.net> + + Improve interrupt safety for single-threaded lisps. + + * slime.el (slime-interrupt): Send a :emacs-interrupt message + together with SIGINT. SIGINT now means "check for new events" + instead of "invoke the debugger". + + * swank-backend.lisp (install-sigint-handler) + (call-with-user-break-handler): New functions. + + * swank.lisp (simple-serve-requests,install-fd-handler): Use it. + (read-packet): New function. Check for interrupts. + (decode-message): Use it. + (wait-for-event/event-loop): Check for interrupts. + +2008-08-11 Helmut Eller <heller(a)common-lisp.net> + + * swank-abcl.lisp (preferred-communication-style): Return nil + until we implement receive-if. + + * swank-openmcl.lisp (receive-if): Support timeout argument. + * swank-allegro.lisp (receive-if): Ditto. + +2008-08-11 Helmut Eller <heller(a)common-lisp.net> + + * swank.lisp (*global-debugger*): Change default back to t. + +2008-08-10 Helmut Eller <heller(a)common-lisp.net> + + * swank-ecl.lisp (thread-id): Assign an non-nil id to unknown + threads. + + * slime.el (slime-execute-tests): Call slime-test-should-fail-p + before executing the test (which may close the connection). + (def-slime-test): Use slime-sync-to-top-level with a timeout. + (slime-temp-directory): New function. + (slime-run-tests): Save repl-history in /tmp. + (slime-repl-mode): Ignore persistent history if + slime-repl-history-file is nil. + (slime-quit-lisp-internal): New function. + (slime-quit-lisp, slime-restart-inferior-lisp): Use it + (slime-batch-test): Exit, if the Lisp isn't up and running after + 30 secs. + +2008-08-10 Helmut Eller <heller(a)common-lisp.net> + + * swank.lisp (wait-for-event): Add timeout argument. This is used + for :fd-handler and :sigio style where we only process events as + long as we don't block. + (wait-for-event/event-loop, read-event) + (decode-message, receive-if): Ditto. + (process-requests): Renamed from read-from-emacs. + (handle-requests): Renamed from handle-request. Take timeout + argument. Update callers. + (process-available-input): Deleted. + (with-swank-error-handler): Renamed from + with-reader-error-handler. + (with-connection): Use it. + +2008-08-10 Helmut Eller <heller(a)common-lisp.net> + + * swank.lisp (invoke-slime-debugger): New function. + Analagous to cl:invoke-debugger. + (swank-debugger-hook): Use it. + 2008-08-09 Helmut Eller <heller(a)common-lisp.net> * slime.el (slime-quit-lisp): Optionally send kill signal.

1 0

[slime-devel] SLIME warns repeatedly after 1.209
by thom.goodsell＠mac.com 12 Aug '08

12 Aug '08

Something in the changes on the 6th cause slime to freak out on sbcl 1.0.18 w/ threads on OS X. (This is sbcl from macports, if it might make a difference.) cvs update -D '2008-08-06 17:50:37' is the earliest change that reproduces the problem. (The preceding changeset won't start at all; changesets before that work fine.) The symptom is a never-ending chain of these warnings: WARNING: Releasing #S(SB-THREAD:MUTEX :NAME NIL :%OWNER NIL :LUTEX #<unknown pointer object, widetag=#x5E {11D7111F}>), owned by another thread: NIL I haven't tried to track down the problem, yet. Does anything seem obvious to those responsible for the relevant changes? Thanks, Thom

3 2

[slime-devel] new debugger behavior?
by Carlos Konstanski 11 Aug '08

11 Aug '08

Last week I upgraded from slime-2.0_p20070816 to slime-2.0_p20080731 (these are the gentoo ebuild names). We're looking at a pretty big jump here; these builds are a year apart. In that year, something changed in the debugger that is not to my liking. I am hoping that it is a feature that I can turn off. When I hit an error in my code, I no longer enter the debugger. Instead, I just get a printed backtrace in the REPL. If I put a (break) in the code, I do enter the debugger. I would like to enter the debugger on all errors, like the good old days. Does this sound like a feature that was added in the last year? Or do I need to look elsewhere for the cause? Carlos Konstanski

4 13

[slime-devel] [RfC] SPC vs. RET in Xref Buffers
by Tobias C. Rittweiler 10 Aug '08

10 Aug '08

Hi, In XREF buffers, pressing RET will open the definition at point in the buffer where you originally issued the xref command, but let the XREF buffer open and let point still be in there. Pressing, SPACE will bring one to the definition at point, but will also close the XREF buffer. I never get that right, so I'd like to swap the key bindings, as this makes more sense to me (and is what I want more often in my experience.) Anyone's muscle memories against that? -T.

3 7

[slime-devel] Daily ChangeLog diff
by mbaringer＠common-lisp.net 10 Aug '08

10 Aug '08

Index: slime/ChangeLog diff -u slime/ChangeLog:1.1419 slime/ChangeLog:1.1428 --- slime/ChangeLog:1.1419 Fri Aug 8 18:46:54 2008 +++ slime/ChangeLog Sat Aug 9 15:57:37 2008 @@ -1,3 +1,59 @@ +2008-08-09 Helmut Eller <heller(a)common-lisp.net> + + * slime.el (slime-quit-lisp): Optionally send kill signal. + (slime-quit-connection-at-point): Disconnect after some timeout. + +2008-08-09 Helmut Eller <heller(a)common-lisp.net> + + Fixes for heap dumping. + + * swank.lisp (*log-output*): Don't initialize at load-time, + otherwise the stream would end up in a heap image. + +2008-08-09 Helmut Eller <heller(a)common-lisp.net> + + * swank-lispworks.lisp (defimplementation): Record location. + +2008-08-09 Helmut Eller <heller(a)common-lisp.net> + + * swank.lisp (*maximum-pipelined-output-chunks*): New variable + +2008-08-09 Helmut Eller <heller(a)common-lisp.net> + + Display the "Use default debugger" restart more prominently + for testing. + + * swank.lisp (swank-debugger-hook): Bind a user visible restart + to invoke the native debugger. + (*global-debugger*): Make this nil by default. + (sldb-loop): Minor cleanups. + (sldb-break-with-default-debugger): Invoke the native debugger + on top of the slime debugger. + + * slime.el (slime-simple-completions): Bind slime-current-thread + to t so that completion is peformed in a fresh thread. + (sldb-setup): Always pop to the debugger buffer. + (sldb-activate): Optionally select the window. + + * swank-sbcl.lisp (sb-thread::get-foreground): Override this + as the default implementation is unusable for Slime. + + * swank-lispworks.lisp (environment-display-notifier): Just + return t. + +2008-08-09 Helmut Eller <heller(a)common-lisp.net> + + * swank-lispworks.lisp (disassemble-frame): Implemented. + +2008-08-09 Tobias C. Rittweiler <tcr(a)freebits.de> + + M-x slime doesn't destroy the window layout anymore when you + switch windows/frames while the connection is being set up. + + * slime.el (slime-hide-inferior-lisp-buffer): Search for the + inferior-lisp buffer's window in all frames. + (slime-repl-update-banner): Do not pop to the REPL buffer. + 2008-08-08 Tobias C. Rittweiler <tcr(a)freebits.de> * slime.el: Fixing `q' in Xref buffers.

1 0

[slime-devel] Proof-of-concept Name-conflict resolution
by Michael Weber 09 Aug '08

09 Aug '08

Hi, attached is a patch for name conflict resolution within SLIME (Zach Beane wrote about the issue quite some time ago: <http://xach.livejournal.com/96625.html >). The patch is not meant for immediate inclusion into SLIME, more as a request for feedback. There are some rough edges in there: * works only with SBCL at the moment. In case of a name conflict, SBCL signals a condition which gives access to the set of conflicting symbols, and also provides a RESOLVE-CONFLICT restart which lets me choose programmatically which symbol to favor out of the conflicting ones. Other Lisps provide different restarts to resolve conflicts, and after superficial looks they appear to be less amenable to programmatic use. * works correctly only with *recent* SBCL (>= 1.0.19.15) * In case of many conflicts (e.g., due to USEing a package), on top of the RESOLVE-CONFLICT restart one can build some automation: instead of ending up a zillion times in the debugger, I can choose to favor a package, and for the current request conflicts are resolved in favor of symbols from that package. If no symbol of the chosen package is in conflict set, I'll end up in the debugger again. I found this behavior quite pleasant to work with in everyday situations. * this is currently realized as a contrib, however, some hooks are needed. I think they are reasonable overall. I provided them as separate patch (0001). * eval-in-emacs is used to trigger the UI (minibuffer & history list--- works for me), which might not be ideal. Cheers, Michael

2 1

[slime-devel] Daily ChangeLog diff
by mbaringer＠common-lisp.net 09 Aug '08

09 Aug '08

Index: slime/ChangeLog diff -u slime/ChangeLog:1.1406 slime/ChangeLog:1.1419 --- slime/ChangeLog:1.1406 Thu Aug 7 13:07:22 2008 +++ slime/ChangeLog Fri Aug 8 18:46:54 2008 @@ -1,3 +1,91 @@ +2008-08-08 Tobias C. Rittweiler <tcr(a)freebits.de> + + * slime.el: Fixing `q' in Xref buffers. + + (slime-popup-buffer-snapshot-unchanged-p): New. + (slime-popup-buffer-restore-snapshot): New. + (slime-xref-quit): Can't use slime-popup-buffer directly. Instead + implement its own quit behaviour using the above functions. + +2008-08-08 Helmut Eller <heller(a)common-lisp.net> + + Use wait-for-event instead of catch/throw where needed. + + * swank.lisp (read-user-input-from-emacs, y-or-n-p-in-emacs) + (eval-in-emacs): Use wait-for-event. + (make-tag): Replaces intern-catch-tag. + (take-input): Deleted. + (dispatch-event): Remove some redundancy. + +2008-08-08 Tobias C. Rittweiler <tcr(a)freebits.de> + + * slime.el: Make xref buffers use `slime-with-popup-buffer', + nee `slime-with-temp-buffer'. + + (slime-with-xref-buffer): Rewritten using a popup buffer. + (slime-init-xref-buffer): Removed. + (slime-display-xref-buffer): Removed. + +2008-08-08 Tobias C. Rittweiler <tcr(a)freebits.de> + + * slime.el: Rename slime's temp-buffer to popup-buffer. + +2008-08-08 Tobias C. Rittweiler <tcr(a)freebits.de> + + * slime.el (slime-compiler-notes-mode, slime-connection-ist-mode): + Add slime-temp-buffer-mode-map to docstring. + +2008-08-08 Tobias C. Rittweiler <tcr(a)freebits.de> + + * slime.el (slime-connections-buffer-name): New variable. + (slime-list-connections): Use it. + (def-slime-selector ?c): Ditto + +2008-08-08 Tobias C. Rittweiler <tcr(a)freebits.de> + + * slime.el: Make the threads browser use `slime-with-temp-buffer'. + + (slime-temp-buffer-quit-function): New buffer-local + variable. Defaults to `slime-temp-buffer-quit'. + (slime-temp-buffer-mode): `q' invokes above variable now. + + (slime-threads-buffer-name): New variable. + (slime-list-threads): Use slime-with-temp-buffer. + (slime-update-threads-buffer): New. Lifted from + slime-list-threads. + (slime-thread-quit): Renamed to slime-quit-threads-buffer. + +2008-08-08 Helmut Eller <heller(a)common-lisp.net> + + Unify event dispatching for single and multi-threaded cases. + + * swank.lisp (send-to-control-thread,read-from-control-thread) + (send-to-socket-io,read-from-socket-io): Deleted. + (send-event, read-event, send-to-emacs) + (signal-interrupt, use-threads-p): New functions. + And more random changes. + +2008-08-08 Helmut Eller <heller(a)common-lisp.net> + + Spawn the auto-flush thread in the front end. + This removes some copy&paste code in various backends. + + * swank.lisp (auto-flush-loop): New function. + (open-streams): Use it. + * swank-backend.lisp (make-stream-interactive): Deleted. + +2008-08-08 Helmut Eller <heller(a)common-lisp.net> + + * slime.el (test disconnect): Call slime-inferior-process + with explicit connection argument to avoid clashes with + buffer-local connections. + +2008-08-08 Tobias C. Rittweiler <tcr(a)freebits.de> + + * slime.el (slime-create-macroexpansion-buffer): Make fontifying + case insensitive as the result from macroexpansion is most likely + printed all upper-case. + 2008-08-07 Helmut Eller <heller(a)common-lisp.net> * slime.el (slime-with-temp-buffer): By default, no longer inherit

1 0

[slime-devel] Mercurial
by Helmut Eller 08 Aug '08

08 Aug '08

What do people think of switching from CVS to Mercurial[*] ? Helmut. [*] http://www.selenic.com/mercurial/

20 50