June 2004 - slime-devel - mailman3.common-lisp.net

[slime-devel] LispWorks OS X create-swank-server
by Brian Downing 21 Jun '04

21 Jun '04

I need this patch to make create-swank-server work on Mac OS X LispWorks Personal (v.4.3.7). It works fine on the same version on Windows without it. Without it I get a network error condition: CL-USER 2 > (swank:create-swank-server) Error: BIND failed: (Can't assign requested address) (49) 1 (abort) Return to level 0. 2 Return to top loop level 0. Type :b for backtrace, :c <option number> to proceed, or :? for other options Obviously this is not the solution, but it demonstrates that there's a problem with the :address keyword argument on Mac OS X, at least on my machine. diff -ur slime-pristine/swank-lispworks.lisp slime/swank-lispworks.lisp --- slime-pristine/swank-lispworks.lisp 2004-06-17 10:59:46.000000000 -0500 +++ slime/swank-lispworks.lisp 2004-06-18 14:51:37.000000000 -0500 @@ -37,8 +37,7 @@ (defimplementation create-socket (host port) (multiple-value-bind (socket where errno) - #-lispworks4.1(comm::create-tcp-socket-for-service port :address host) - #+lispworks4.1(comm::create-tcp-socket-for-service port) + (comm::create-tcp-socket-for-service port) (cond (socket socket) (t (error 'network-error :format-control "~A failed: ~A (~D)" -bcd

2 1

[slime-devel] SBCL I/O streams broken?
by Brian Downing 21 Jun '04

21 Jun '04

I've been having a lot of trouble with I/O to SBCL through slime recently. This is on PPC, using either :sigio or :fd-handler communication style, and dedicated-output-stream either t or nil. It's easy for me to reproduce (Slime as of fifteen minutes ago, SBCL 0.8.11.14): CL-USER> (inspect nil) The object is a SYMBOL. 0. Name: "NIL" 1. Package: #<PACKAGE "COMMON-LISP"> 2. Value: NIL 3. Function: "unbound" 4. Plist: NIL > 0 0 0 00 0 0 90 <C-c C-c, select the second to bottom-most reset> ; Evaluation aborted The second to bottom-most one (usually an abort, but not TOPLEVEL) seems the most reliable to clear it up. What's worse is when (as happens on SBCL/PPC sometimes) you get a recursive debug when the backtrace fails, it drops you into a TTY debugger in the Slime REPL. Since you can't interact with it, and it's difficult to get another SLDB debugger to come up, this is kind of bad. I'm not sure when this started. Also, is there any way to get the slime version anymore? The date no longer appears at the top of a repl, which is kind of annoying since it was easy to identify what era of slime you were working with. -bcd -- *** Brian Downing <bdowning at lavos dot net>

1 0

[slime-devel] Daily ChangeLog diff
by Luke Gorrie 21 Jun '04

21 Jun '04

Index: slime/ChangeLog diff -u slime/ChangeLog:1.411 slime/ChangeLog:1.417 --- slime/ChangeLog:1.411 Sat Jun 19 22:49:08 2004 +++ slime/ChangeLog Sun Jun 20 23:27:24 2004 @@ -1,3 +1,57 @@ +2004-06-21 Luke Gorrie <luke(a)bluetail.com> + + * swank-loader.lisp (*lisp-name*): Add version number to + Lispwork's fasl directory. We should do this for ACL and OpenMCL + too, but for some reason my ACL 5.0 gets an error when trying to + create a directory with a version number in its name, and I don't + have OpenMCL to test with. + + * swank-backend.lisp, swank.lisp (add-hook, run-hook): Moved the + hook mechanism and all hooks to swank.lisp (from + swank-backend.lisp). There is no compelling use for the hooks in + backends yet and I want to pass swank.lisp-internal data + structures in the existing hooks. + (notify-backend-of-connection): Call `emacs-connected' with the + user-io stream for its argument. Should fix previous breakage + where the connection structure was passed instead. + (*globally-redirect-io*): New configurable: when true the standard + streams are globally redirected to Emacs. That way even + e.g. SERVE-EVENT handlers will print to Emacs. Currently does not + handle standard input -- that is trickier since the Lisp's native + REPL can be trying to read from that. + + * slime.el (slime-complete-maybe-restore-window-configuration): + Only restore the window configuration if the completions buffer is + currently visible in the window that we popped it up in. + (slime-complete-maybe-save-window-configuration): Don't save the + window configuration if the completions buffer is already visible. + (slime-repl-return): Make sure the newline goes at the end of the + input, not at point. + (slime-complete-restore-window-configuration): Wrap the + `set-window-configuration' call in (run-at-time 0 ..). XEmacs does + not allow us to set the window configuration from inside + pre-command-hook. + +2004-06-20 Helmut Eller <e9626484(a)stud3.tuwien.ac.at> + + * swank-sbcl.lisp (emacs-connected): Set *invoke-debugger-hook* to + our debugger hook. Not optimal, but at least BREAK will then + invoke our debugger. + (*trap-load-time-warnings*): New variable. If it is true, + conditions, most notably redefinition warnings, signalled at load + time are not trapped. + (swank-compile-file, swank-compile-string): Use it. + + * swank.lisp (guess-buffer-package): Don't signal a continuable + error if the package doesn't exists; that's too annoying. + + * slime.el: Fix outline structure. + (slime-maybe-list-compiler-notes): Fix thinko. + (break): New test. Reorganize the test-suite a bit to support + "expected failures". + (slime-eval-feature-conditional, slime-to-feature-keyword): Add a + ?: to the symbol-name if needed. + 2004-06-20 Luke Gorrie <luke(a)bluetail.com> * swank.lisp (changelog-date): Removed unneeded function.

1 0

[slime-devel] *new-connection-hook*
by Marco Baringer 21 Jun '04

21 Jun '04

*new-connection-hook* passes the connection struct to emacs-connected which expects a stream argument. something has to be changed. should emacs-connected be changed to take a connection object or should the elements in *new-connection-hook* be passed the stream? -- -Marco Ring the bells that still can ring. Forget your perfect offering. There is a crack in everything. That's how the light gets in. -Leonard Cohen

3 5

[slime-devel] swank:*globally-redirect-io*
by Luke Gorrie 21 Jun '04

21 Jun '04

Ahoy, I added a new configurable called swank:*globally-redirect-io*. If you set this to true in your ~/.swank.lisp then we hijack the standard streams globally for the duration of the SLIME connection. That way e.g. serve-event handler output will go to the REPL instead of *inferior-lisp*. Currently it doesn't redirect *standard-input*. That's trickier since we can get the Lisp's native REPL trying to read its input from our SLIME REPL :-). You can also globally install SLDB as the debugger like this: (setq *debugger-hook* #'swank:swank-debugger-hook) But that's old news.. Cheers, Luke (who just found his first use for synonym streams)

1 0

[slime-devel] Error when typing `(in-package' in Lisp mode buffer
by Paolo Amoroso 20 Jun '04

20 Jun '04

As soon as I type the "e" of "(in-package" in a Lisp mode buffer, SLIME stops with the error message included below, and reissues the error until I delete enough characters to make the string no longer recognizable as a package setting form. I use SLIME CVS sources updated a few hours ago with GNU Emacs 20.7.2 and CMUCL 18e under Debian Woody. Paolo ----------------------------------------------------------------- Error in function SWANK::GUESS-BUFFER-PACKAGE: Package nil not found. [Condition of type SIMPLE-ERROR] Restarts: 0: [CONTINUE] Use current package. [COMMON-LISP-USER] 1: [ABORT] Abort handling SLIME request. 2: [ABORT] Return to Top-Level. Backtrace: 0: (SWANK::GUESS-BUFFER-PACKAGE "nil") 1: (SWANK::EVAL-FOR-EMACS (SWANK:ARGLIST-FOR-ECHO-AREA '("in-package")) "nil" 11) 2: ("DEFUN HANDLE-REQUEST") 3: (SWANK::CALL-WITH-REDIRECTED-IO #S(SWANK::CONNECTION :SOCKET-IO #<Stream for descriptor 6> :DEDICATED-OUTPUT #1=#<Stream for descriptor 7> :USER-INPUT #2=#<SWANK-BACKEND::SLIME-INPUT-STREAM {48607ECD}> :USER-OUTPUT #1# :USER-IO #<Two-Way Stream, Input = #2#, Output = #1#> ...) #<Function "DEFUN HANDLE-REQUEST" {4845C931}>) 4: (SWANK::HANDLE-REQUEST #S(SWANK::CONNECTION :SOCKET-IO #<Stream for descriptor 6> :DEDICATED-OUTPUT #1=#<Stream for descriptor 7> :USER-INPUT #2=#<SWANK-BACKEND::SLIME-INPUT-STREAM {48607ECD}> :USER-OUTPUT #1# :USER-IO #<Two-Way Stream, Input = #2#, Output = #1#> ...)) 5: (SWANK::PROCESS-AVAILABLE-INPUT #<Stream for descriptor 6> #<Closure Over Function "DEFUN INSTALL-SIGIO-HANDLER" {48682EB1}>) 6: (SWANK::HANDLER) 7: (SWANK-BACKEND::SIGIO-HANDLER #<#1=unused-arg> #<#1#> #<#1#>) 8: ("Foreign function call land") 9: ("Foreign function call land") 10: ("Foreign function call land") 11: ("Foreign function call land") 12: ("Foreign function call land") 13: (SYSTEM:WAIT-UNTIL-FD-USABLE 0 :INPUT NIL) 14: (COMMON-LISP::DO-INPUT #<Stream for Standard Input>) 15: (COMMON-LISP::INPUT-CHARACTER #<Stream for Standard Input> NIL (COMMON-LISP::*EOF*)) 16: (COMMON-LISP::SYNONYM-IN #<Synonym Stream to SYSTEM:*STDIN*> NIL (COMMON-LISP::*EOF*)) 17: (COMMON-LISP::TWO-WAY-IN #<Two-Way Stream, Input = #<Synonym Stream to SYSTEM:*STDIN*>, Output = #<Synonym Stream to SYSTEM:*STDOUT*>> NIL (COMMON-LISP::*EOF*)) 18: (READ-CHAR #<Two-Way Stream, Input = #<Synonym Stream to SYSTEM:*STDIN*>, Output = #<Synonym Stream to SYSTEM:*STDOUT*>> NIL (COMMON-LISP::*EOF*) #<unused-arg>) 19: (READ-PRESERVING-WHITESPACE #<Two-Way Stream, Input = #<Synonym Stream to SYSTEM:*STDIN*>, Output = #<Synonym Stream to SYSTEM:*STDOUT*>> NIL (:EOF) T) 20: (READ-PRESERVING-WHITESPACE #<Two-Way Stream, Input = #<Synonym Stream to SYSTEM:*STDIN*>, Output = #<Synonym Stream to SYSTEM:*STDOUT*>> NIL (:EOF) NIL) 21: (READ-PRESERVING-WHITESPACE 4 #<Two-Way Stream, Input = #<Synonym Stream to SYSTEM:*STDIN*>, Output = #<Synonym Stream to SYSTEM:*STDOUT*>> NIL (:EOF) ...)[:EXTERNAL] 22: (READ #<Two-Way Stream, Input = #<Synonym Stream to SYSTEM:*STDIN*>, Output = #<Synonym Stream to SYSTEM:*STDOUT*>> NIL (:EOF) NIL) 23: (COMMON-LISP::%TOP-LEVEL) 24: (COMMON-LISP::RESTART-LISP) ----------------------------------------------------------------- -- Why Lisp? http://alu.cliki.net/RtL%20Highlight%20Film

1 1

[slime-devel] Compilation policy in and out of Slime
by Brian Downing 20 Jun '04

20 Jun '04

[Slime 2004-06-18, SBCL Darwin/PPC 0.8.11.14+many hacks, an oldish Emacs CVS] Once thing I noticed when working on some SBCL internals tonight is that Slime is compiling things for SBCL with very little debug. This can be seen by compiling: (defun bar (x y) (zot x y)) from the Slime REPL and the *inferior-lisp* REPL. The Slime-compiled one has a very short disassembly, and tail-calls the FDEFINITION. The REPL-compiled one is quite long, and doesn't. This is strange, though, since sb-c::*policy* evaluates to the same in either environment. While this was sort of nice for me (Slime's low debug made my find and fix a nasty stack-destroying bug), it seems bad that Slime-compiled code is defaulting to low debug levels, and seems to not be respecting the *policy*. Is this happening on anyone else's installation, and does anyone have an idea why this is happening? (I've not CCed sbcl-devel on this - I figure enough of the SBCL developers who know anything about Slime are on this list.) -bcd

1 1

[slime-devel] Daily ChangeLog diff
by Luke Gorrie 20 Jun '04

20 Jun '04

Index: slime/ChangeLog diff -u slime/ChangeLog:1.409 slime/ChangeLog:1.411 --- slime/ChangeLog:1.409 Fri Jun 18 14:54:34 2004 +++ slime/ChangeLog Sat Jun 19 22:49:08 2004 @@ -1,3 +1,46 @@ +2004-06-20 Luke Gorrie <luke(a)bluetail.com> + + * swank.lisp (changelog-date): Removed unneeded function. + (connection-info): No more version field in result. + + * slime.el: Audited to remove namespace slipups. Tracking a really + horrible clashing-with-some-user-configuration bug and want to + eliminate potential symbol conflicts. + (sldb-get-buffer): Renamed from `get-sldb-buffer'. + (slime-emacs-20-p): Renamed from `emacs-20-p'. + (slime-defun-if-undefined): Renamed from `defun-if-undefined'. + (slime-isearch): Small bugfix that could cause M-. to go to the + wrong place in CMUCL. + (slime-changelog-date, slime-check-protocol-version): Removed + unneeded functions. + + * swank-backend.lisp (add-hook, run-hook): Added an Emacs-like + hook mechanism. The hope is that this will make some sections of + the code more self-describing by showing where they hook in. + (*new-connection-hook*): Hook run when a new connection is + established. Initialized to '(swank-backend:emacs-connected). + (*pre-reply-hook*): Hook run before sending a reply to Emacs. + + * swank.lisp: Added some comments and docstrings. + (package-external-symbols): Removed unused function. + (serve-connection): Call *new-connection-hook*. + (eval-for-emacs): Call *pre-reply-hook*. + (sync-features-to-emacs, sync-indentation-to-emacs): Added to + *pre-reply-hook*. + (cl-package, keyword-package): Now defconstant instead of + defvar. Removed the *'s accordingly. + + * slime.el (slime-abort-connection): Renamed from + `slime-connection-abort'. The new name is easier to find with + completion. + + * swank-sbcl.lisp: Change sb-posix:: to sb-posix: + +2004-06-19 Luke Gorrie <luke(a)bluetail.com> + + * swank.lisp (known-to-emacs-p): Bugfix. Indentation-updates was + broken. + 2004-06-18 Luke Gorrie <luke(a)bluetail.com> * slime.el (slime-buffer-package): If DONT-CACHE is true and no

1 0

[slime-devel] M-x slime to go back to the REPL
by Pascal J.Bourguignon 20 Jun '04

20 Jun '04

Both ilisp and run-lisp when invoked a second time switch back to their buffer. slime did not. Now it does too: (defun slime () "Start an inferior^_superior Lisp and connect to its Swank server, or switch to the slime REPL buffer if it's already running." (interactive) (if (get-buffer-process (get-buffer "*inferior-lisp*")) (switch-to-buffer (slime-output-buffer)) (when (or (not (slime-bytecode-stale-p)) (slime-urge-bytecode-recompile)) (cond ((and current-prefix-arg (slime-connected-p) (get-buffer "*inferior-lisp*")) (unless (slime-maybe-rearrange-inferior-lisp) (slime-disconnect))) (t (slime-disconnect))) (slime-maybe-start-lisp) (slime-read-port-and-connect)))) -- __Pascal Bourguignon__ http://www.informatimago.com/ There is no worse tyranny than to force a man to pay for what he does not want merely because you think it would be good for him. -- Robert Heinlein

1 0

[slime-devel] splash patch version 3
by Pascal J.Bourguignon 20 Jun '04

20 Jun '04

Thanks to user feedback, here is the third release of the splash screen for slime. Added: - selection of splash-dark.png or splash-light.png depending on the dark/light status of the frame. - search the splash-*.png files first in ~/.slime/, then in load-paths. That should give each user an opportunity to have his own splash screen. splash-3.patch, to be applied after the previous patch. -- __Pascal Bourguignon__ http://www.informatimago.com/ There is no worse tyranny than to force a man to pay for what he does not want merely because you think it would be good for him. -- Robert Heinlein

1 0