September 2006 - slime-devel - mailman3.common-lisp.net

[slime-devel] REPL-to-regression-test made easy.
by Vladimir Sedach 25 Sep '06

25 Sep '06

Hello everybody, At Friday's inaugural Calgary Lisp user group meeting, Michael Beauregard suggested a feature to SLIME that would let you take an expression-result pair from REPL interaction and easily add it as a new regression test to a test suite. Here is some code that does something like that: (defvar *slime-last-repl-input*) (defvar *slime-last-repl-result*) (defvar *slime-last-repl-input-matches-last-result* nil) ;; set to T if last input expression returned normally and it is safe to save the input-result pair as a unit test, nil otherwise (defvar *slime-regression-test-file* nil "Pathname of file containing test cases to which the next test case will be appended to.") (defvar *slime-regression-test-framework* nil "The unit test framework should be a function that takes as arguments the input and expected result of the unit test as strings and produces an object that is suitable to be serialized and appended to a file containing unit tests for that particular framework.") (defun slime-MIT-regression-testing-framework (input result) (princ "(deftest " (current-buffer)) (princ (gensym) (current-buffer)) (newline) (princ "(multiple-value-list " (current-buffer)) (princ input (current-buffer)) (princ ")" (current-buffer)) (newline) (princ result (current-buffer)) (princ ")" (current-buffer))) (defun slime-save-last-input-as-regression-test () "When called, takes the last REPL expression sent to Lisp by SLIME and the output it returns, and adds them as a unit test to the unit test file specified by *slime-regression-test-file*. If the last REPL command resulted in an error, does nothing. The specific brand of regression testing framework you're using is controlled by *slime-regression-test-framework*." (interactive) (unless *slime-regression-test-file* (call-interactively #'slime-set-regression-test-file)) (when *slime-last-repl-input-matches-last-result* (with-temp-buffer (lisp-mode) (newline) (let* ((res *slime-last-repl-result*) (results (cons 'list (cond ((equal :values (car res)) (second res)) ((equal :present (car res)) (mapcar #'car (second res))) (t "nil"))))) (funcall *slime-regression-test-framework* *slime-last-repl-input* results)) (newline) (indent-region (point-min) (point-max) nil) (append-to-file (point-min) (point-max) *slime-regression-test-file*)))) (defun slime-set-regression-test-file (file) (interactive "FRegression test file: ") (setf *slime-regression-test-file* file)) (defadvice slime-repl-eval-string (before regression-testing-string-capturer (string)) "Get the string passed to slime-repl-eval-string before it gets mangled." (setf *slime-last-repl-input* (copy-seq string))) ;;; need to redefine slime-repl-eval-string (defun slime-repl-eval-string (string) (slime-rex () ((list 'swank:listener-eval string) (slime-lisp-package)) ((:ok result) ;;; modified code be here (setf *slime-last-repl-result* result *slime-last-repl-input-matches-last-result* t) ;;; end modified code (with-current-buffer (slime-output-buffer) (slime-repl-insert-prompt result))) ((:abort) ;;; modified code be here (setf *slime-last-repl-input-matches-last-result* nil) ;;; end modified code (slime-repl-show-abort)))) Comments? I don't use regression testing tools very much myself, so the first thing I want to ask, is this approach going to work for the commonly used CL testing frameworks, or is something more flexible necessary? If this were to become part of SLIME, the big question is, what key should slime-save-last-input-as-regression-test be bound to? Having just the last repl-expression-result pair is probably good enough, but is there anyone wanting a history of expression-result pair interactions and a nice colorful-buffer way of choosing which ones to save? Vladimir

1 0

[slime-devel] Daily ChangeLog diff
by mbaringer＠common-lisp.net 22 Sep '06

22 Sep '06

Index: slime/ChangeLog diff -u slime/ChangeLog:1.955 slime/ChangeLog:1.957 --- slime/ChangeLog:1.955 Wed Sep 20 15:28:57 2006 +++ slime/ChangeLog Thu Sep 21 13:36:18 2006 @@ -1,9 +1,33 @@ +2006-09-21 Marco Baringer <mb(a)bese.it> + + * swank.lisp (find-definitions-for-emacs): Don't return locations + whose CAR is :error. + (xref): Process whatever is returned by the various xref functions + with the new sanitize-xrefs functions. + (sanitize-xrefs): Clean up the list of xrefs to remove duplicates. + Patch by Dan Weinreb <dlw(a)itasoftware.com> + + * slime.el (slime-goto-first-note-after-compilation): New + variable. This controls the behaviour of (next|prev)-note + immediatly after a slime-compile-and-load-file. + (slime-compilation-just-finished): New variable. + (slime-compilation-finished): Update slime-compilation-finished. + (slime-next-note, slime-previous-note): Respect + slime-compilation-just-finished. + (slime-autodoc-use-multiline-p): Specify the type. + (slime-repl-grab-old-input): Typo in docstring. + (slime-cheat-sheet): Deal with multiple-bindings + (slime-cheat-sheet-table): Update as per #lisp's suggestions. + 2006-09-20 Marco Baringer <mb(a)bese.it> * slime.el (slime-cheat-sheet): New function. (slime-cheat-sheet-table): New variable which specifies what the cheat sheet should list. - + (slime-read-package-name): Set require to T in the call to + completing read, it doesn't make any sense to switch to an + inexistent package. + * doc/slime.texi: Added "Tips and Tricks" chapter (need a better name for this). @@ -12,10 +36,6 @@ (locate-compiler-note): Say, in the error message, what data caused the error. - * slime.el (slime-read-package-name): Set require to T in the call - to completing read, it doesn't make any sense to switch to an - inexistent package. - 2006-09-20 Juho Snellman <jsnell(a)iki.fi> * swank-sbcl.lisp (call-with-debugger-hook): use INVOKE-STEPPER

1 0

[slime-devel] Daily ChangeLog diff
by mbaringer＠common-lisp.net 21 Sep '06

21 Sep '06

Index: slime/ChangeLog diff -u slime/ChangeLog:1.948 slime/ChangeLog:1.955 --- slime/ChangeLog:1.948 Tue Sep 19 07:56:38 2006 +++ slime/ChangeLog Wed Sep 20 15:28:57 2006 @@ -1,6 +1,29 @@ +2006-09-20 Marco Baringer <mb(a)bese.it> + + * slime.el (slime-cheat-sheet): New function. + (slime-cheat-sheet-table): New variable which specifies what the + cheat sheet should list. + + * doc/slime.texi: Added "Tips and Tricks" chapter (need a better + name for this). + + * swank-sbcl.lisp (fallback-source-location): Use abort-request + instead of error. + (locate-compiler-note): Say, in the error message, what data + caused the error. + + * slime.el (slime-read-package-name): Set require to T in the call + to completing read, it doesn't make any sense to switch to an + inexistent package. + +2006-09-20 Juho Snellman <jsnell(a)iki.fi> + + * swank-sbcl.lisp (call-with-debugger-hook): use INVOKE-STEPPER + instead of calling the stepper hook manually + 2006-09-19 Juho Snellman <jsnell(a)iki.fi> - * swank-backend.lisp (call-with-debugger-hook): make the stepper + * swank-sbcl.lisp (call-with-debugger-hook): make the stepper also work with a threaded SBCL, by binding a handler for sb-ext:stepper-condition instead of relying on the one that SBCL establishes on the toplevel

1 0

[slime-devel] Daily ChangeLog diff
by mbaringer＠common-lisp.net 20 Sep '06

20 Sep '06

Index: slime/ChangeLog diff -u slime/ChangeLog:1.947 slime/ChangeLog:1.948 --- slime/ChangeLog:1.947 Mon Sep 18 17:56:13 2006 +++ slime/ChangeLog Tue Sep 19 07:56:38 2006 @@ -1,3 +1,10 @@ +2006-09-19 Juho Snellman <jsnell(a)iki.fi> + + * swank-backend.lisp (call-with-debugger-hook): make the stepper + also work with a threaded SBCL, by binding a handler for + sb-ext:stepper-condition instead of relying on the one that SBCL + establishes on the toplevel + 2006-09-19 Juho Snellman <jsnell(a)iki.fi> Extend the stepper protocol to work nicely with the SBCL stepper.

1 0

[slime-devel] who-calls vs list-callers and calls-who vs list-callees
by Marco Baringer 19 Sep '06

19 Sep '06

here's the docstring for list-callers: -------------------------------------------------- List the callers of FUNCTION-NAME. This function is like WHO-CALLS except that it is expected to use lower-level means. Whereas WHO-CALLS is usually implemented with special compiler support, LIST-CALLERS is usually implemented by groveling for constants in function objects throughout the heap. The return value is as for WHO-CALLS. -------------------------------------------------- anybody remember why we make this distinction? -- -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

2 3

[slime-devel] Daily ChangeLog diff
by mbaringer＠common-lisp.net 19 Sep '06

19 Sep '06

Index: slime/ChangeLog diff -u slime/ChangeLog:1.945 slime/ChangeLog:1.947 --- slime/ChangeLog:1.945 Wed Sep 13 18:56:14 2006 +++ slime/ChangeLog Mon Sep 18 17:56:13 2006 @@ -1,3 +1,55 @@ +2006-09-19 Juho Snellman <jsnell(a)iki.fi> + + Extend the stepper protocol to work nicely with the SBCL stepper. + + If sldb is invoked on a condition that's sldb-stepper-condition-p, + the sldb functions sldb-step, sldb-next and sldb-out will invoke + the matching backend functions for stepping into the stepped form, + to the next form, or out of the current function. Otherwise the + functions will behave like sldb-step used to (call active-stepping and + select the continue restart). + + * swank-backend.lisp (sldb-stepper-condition-p, sldb-step-into, + sldb-step-next, sldb-step-out): New interface functions + + * swank-sbcl.lisp (activate-stepper, condition-extras, + sldb-stepper-condition-p, sldb-step-into, sldb-step-next, + sldb-step-out): Implemented (conditional on CVS SBCL) + (call-with-debugger-hook): bind sb-ext:*stepper-hook* to + a function that binds *stack-top-hint* and invokes the debugger + (conditional on CVS SBCL) + + * swank.lisp (define-stepper-function): new macro for defining + stepper-related functions, since they all follow the same form + (sldb-step): redefine with define-stepper-function + (sldb-next, sldb-out): new functions + (*sldb-stepping-p*): typo in docstring + + * slime.el (sldb-next, sldb-out): New commands + (sldb-mode-map): bind sldb-next to "x" and sldb-out to "o" + +2006-09-18 Dan Weinreb <dlw(a)itasoftware.com> + + For those cases where SLIME can't complete a user request (like + loading an asdf system without asdf or describing an inexistent + symbol) instead of signaling an error SWANK should politely inform + the user and return normally. + + * swank.lisp (eval-for-emacs): Handle request-abort conditions. + (decode-keyword-arg, get-repl-result, parse-symbol-or-lose): Use + abort-request instead of error. + + * swank-backend.lisp (request-abort): New condition. + (abort-request): Convenience function for signaling request-abort + conditions. + (operate-on-system): Use abort-request instead of error + (:swank-backend): Export the symbols abort-request and + request-abort. + + * slime.el (slime-rex): Update docstring. + (slime-eval, slime-eval-async): Added new REASON parameter sent + along with :abort message. + 2006-09-14 Douglas Crosher <dcrosher(a)common-lisp.net> * swank-scl (arglist, function-arglist, spawn): update for the SCL.

1 0

[slime-devel] Trouble loading new slime
by Matthew D Swank 18 Sep '06

18 Sep '06

Seen in the *inferior-lisp* buffer while trying to load the latest cvs version of slime using sbcl-0.9.15 on Linux x86: ... WARNING: These Swank interfaces are unimplemented: (ACTIVATE-STEPPING CALLS-WHO DISASSEMBLE-FRAME SLDB-BREAK-AT-START SLDB-BREAK-ON-RETURN WHO-BINDS WHO-CALLS WHO-MACROEXPANDS WHO-REFERENCES WHO-SETS WHO-SPECIALIZES) T * Help! 11 nested errors. SB-KERNEL:*MAXIMUM-ERROR-DEPTH* exceeded. 0: (BACKTRACE 536870911 #<SYNONYM-STREAM :SYMBOL SB-SYS:*TTY* {C545449}>) 1: ((LAMBDA NIL)) 2: ((LAMBDA NIL)) 3: (SB-IMPL::%WITH-STANDARD-IO-SYNTAX #<CLOSURE (LAMBDA NIL) {C545565}>) 4: (SB-IMPL::ERROR-ERROR) 5: (SB-IMPL::INFINITE-ERROR-PROTECTOR) 6: (ERROR SB-INT:SIMPLE-STREAM-ERROR) 7: (SB-IMPL::SIMPLE-STREAM-PERROR "couldn't write to ~S" #<SB-SYS:FD-STREAM for "the terminal" {A7C4821}> 32) 8: (SB-IMPL::FLUSH-OUTPUT-BUFFER #<SB-SYS:FD-STREAM for "the terminal" {A7C4821}>) 9: (SB-IMPL::OUTPUT-CHAR-UTF-8-LINE-BUFFERED #<SB-SYS:FD-STREAM for "the terminal" {A7C4821}> #\Newline) 10: (WRITE-CHAR #\Newline #<SB-SYS:FD-STREAM for "the terminal" {A7C4821}>) 11: (TERPRI #<SYNONYM-STREAM :SYMBOL SB-SYS:*TTY* {910E8F1}>) 12: ((LAMBDA NIL)) 13: ((LAMBDA NIL)) 14: (SB-IMPL::%WITH-STANDARD-IO-SYNTAX #<CLOSURE (LAMBDA NIL) {C544FA5}>) 15: (SB-IMPL::ERROR-ERROR) 16: (SB-IMPL::INFINITE-ERROR-PROTECTOR) 17: (ERROR SB-INT:SIMPLE-STREAM-ERROR) 18: (SB-IMPL::SIMPLE-STREAM-PERROR "couldn't write to ~S" #<SB-SYS:FD-STREAM for "the terminal" {A7C4821}> 32) 19: (SB-IMPL::FLUSH-OUTPUT-BUFFER #<SB-SYS:FD-STREAM for "the terminal" {A7C4821}>) 20: (FORCE-OUTPUT #<SYNONYM-STREAM :SYMBOL *TERMINAL-IO* {915C2C1}>) 21: (SB-INT:FLUSH-STANDARD-OUTPUT-STREAMS) 22: (INVOKE-DEBUGGER #<SB-INT:SIMPLE-STREAM-ERROR {C544E91}>) 23: (ERROR SB-INT:SIMPLE-STREAM-ERROR) 24: (SB-IMPL::SIMPLE-STREAM-PERROR "couldn't write to ~S" #<SB-SYS:FD-STREAM for "the terminal" {A7C4821}> 32) 25: (SB-IMPL::FLUSH-OUTPUT-BUFFER #<SB-SYS:FD-STREAM for "the terminal" {A7C4821}>) 26: (FORCE-OUTPUT #<SYNONYM-STREAM :SYMBOL *TERMINAL-IO* {915C2C1}>) 27: (SB-INT:FLUSH-STANDARD-OUTPUT-STREAMS) 28: (INVOKE-DEBUGGER #<SB-INT:SIMPLE-STREAM-ERROR {C544D71}>) 29: (ERROR SB-INT:SIMPLE-STREAM-ERROR) 30: (SB-IMPL::SIMPLE-STREAM-PERROR "couldn't write to ~S" #<SB-SYS:FD-STREAM for "the terminal" {A7C4821}> 32) 31: (SB-IMPL::FLUSH-OUTPUT-BUFFER #<SB-SYS:FD-STREAM for "the terminal" {A7C4821}>) 32: (FORCE-OUTPUT #<SYNONYM-STREAM :SYMBOL *TERMINAL-IO* {915C2C1}>) 33: (SB-INT:FLUSH-STANDARD-OUTPUT-STREAMS) 34: (INVOKE-DEBUGGER #<SB-INT:SIMPLE-STREAM-ERROR {C544C51}>) 35: (ERROR SB-INT:SIMPLE-STREAM-ERROR) 36: (SB-IMPL::SIMPLE-STREAM-PERROR "couldn't write to ~S" #<SB-SYS:FD-STREAM for "the terminal" {A7C4821}> 32) 37: (SB-IMPL::FLUSH-OUTPUT-BUFFER #<SB-SYS:FD-STREAM for "the terminal" {A7C4821}>) 38: (FORCE-OUTPUT #<SYNONYM-STREAM :SYMBOL *TERMINAL-IO* {915C2C1}>) 39: (SB-INT:FLUSH-STANDARD-OUTPUT-STREAMS) 40: (INVOKE-DEBUGGER #<SB-INT:SIMPLE-STREAM-ERROR {C544B31}>) 41: (ERROR SB-INT:SIMPLE-STREAM-ERROR) 42: (SB-IMPL::SIMPLE-STREAM-PERROR "couldn't write to ~S" #<SB-SYS:FD-STREAM for "the terminal" {A7C4821}> 32) 43: (SB-IMPL::FLUSH-OUTPUT-BUFFER #<SB-SYS:FD-STREAM for "the terminal" {A7C4821}>) 44: (FORCE-OUTPUT #<SYNONYM-STREAM :SYMBOL *TERMINAL-IO* {915C2C1}>) 45: (SB-INT:FLUSH-STANDARD-OUTPUT-STREAMS) 46: (INVOKE-DEBUGGER #<SB-INT:SIMPLE-STREAM-ERROR {C544A11}>) 47: (ERROR SB-INT:SIMPLE-STREAM-ERROR) 48: (SB-IMPL::SIMPLE-STREAM-PERROR "couldn't write to ~S" #<SB-SYS:FD-STREAM for "the terminal" {A7C4821}> 32) 49: (SB-IMPL::FLUSH-OUTPUT-BUFFER #<SB-SYS:FD-STREAM for "the terminal" {A7C4821}>) 50: (FORCE-OUTPUT #<SYNONYM-STREAM :SYMBOL *TERMINAL-IO* {915C2C1}>) 51: (SB-INT:FLUSH-STANDARD-OUTPUT-STREAMS) 52: (INVOKE-DEBUGGER #<SB-INT:SIMPLE-STREAM-ERROR {C5448F1}>) 53: (ERROR SB-INT:SIMPLE-STREAM-ERROR) 54: (SB-IMPL::SIMPLE-STREAM-PERROR "couldn't write to ~S" #<SB-SYS:FD-STREAM for "the terminal" {A7C4821}> 32) 55: (SB-IMPL::FLUSH-OUTPUT-BUFFER #<SB-SYS:FD-STREAM for "the terminal" {A7C4821}>) 56: (FORCE-OUTPUT #<SYNONYM-STREAM :SYMBOL *TERMINAL-IO* {915C2C1}>) 57: (SB-INT:FLUSH-STANDARD-OUTPUT-STREAMS) 58: (INVOKE-DEBUGGER #<SB-INT:SIMPLE-STREAM-ERROR {C5447D1}>) 59: (ERROR SB-INT:SIMPLE-STREAM-ERROR) 60: (SB-IMPL::SIMPLE-STREAM-PERROR "couldn't write to ~S" #<SB-SYS:FD-STREAM for "the terminal" {A7C4821}> 32) 61: (SB-IMPL::FLUSH-OUTPUT-BUFFER #<SB-SYS:FD-STREAM for "the terminal" {A7C4821}>) 62: (FORCE-OUTPUT #<SYNONYM-STREAM :SYMBOL *TERMINAL-IO* {915C2C1}>) 63: (SB-INT:FLUSH-STANDARD-OUTPUT-STREAMS) 64: (INVOKE-DEBUGGER #<SB-INT:SIMPLE-STREAM-ERROR {C5446B1}>) 65: (ERROR SB-INT:SIMPLE-STREAM-ERROR) 66: (SB-IMPL::SIMPLE-STREAM-PERROR "couldn't write to ~S" #<SB-SYS:FD-STREAM for "the terminal" {A7C4821}> 32) 67: (SB-IMPL::FLUSH-OUTPUT-BUFFER #<SB-SYS:FD-STREAM for "the terminal" {A7C4821}>) 68: (FORCE-OUTPUT #<SYNONYM-STREAM :SYMBOL *TERMINAL-IO* {915C2C1}>) 69: (SB-INT:FLUSH-STANDARD-OUTPUT-STREAMS) 70: (INVOKE-DEBUGGER #<SB-INT:SIMPLE-STREAM-ERROR {C544591}>) 71: (ERROR SB-INT:SIMPLE-STREAM-ERROR) 72: (SB-IMPL::SIMPLE-STREAM-PERROR "couldn't write to ~S" #<SB-SYS:FD-STREAM for "the terminal" {A7C4821}> 32) 73: (SB-IMPL::FLUSH-OUTPUT-BUFFER #<SB-SYS:FD-STREAM for "the terminal" {A7C4821}>) 74: (FORCE-OUTPUT #<SYNONYM-STREAM :SYMBOL *TERMINAL-IO* {915C2C1}>) 75: (SB-INT:FLUSH-STANDARD-OUTPUT-STREAMS) 76: (INVOKE-DEBUGGER #<SB-INT:SIMPLE-STREAM-ERROR {C544471}>) 77: (ERROR SB-INT:SIMPLE-STREAM-ERROR) 78: (SB-IMPL::SIMPLE-STREAM-PERROR "couldn't write to ~S" #<SB-SYS:FD-STREAM for "the terminal" {A7C4821}> 32) 79: (SB-IMPL::FLUSH-OUTPUT-BUFFER #<SB-SYS:FD-STREAM for "the terminal" {A7C4821}>) 80: (FORCE-OUTPUT #<SYNONYM-STREAM :SYMBOL *TERMINAL-IO* {915C2C1}>) 81: (SB-INT:FLUSH-STANDARD-OUTPUT-STREAMS) 82: (INVOKE-DEBUGGER #<SB-INT:SIMPLE-STREAM-ERROR {C544171}>) 83: (ERROR SB-INT:SIMPLE-STREAM-ERROR) 84: (SB-IMPL::SIMPLE-STREAM-PERROR "couldn't write to ~S" #<SB-SYS:FD-STREAM for "the terminal" {A7C4821}> 32) 85: (SB-IMPL::FLUSH-OUTPUT-BUFFER #<SB-SYS:FD-STREAM for "the terminal" {A7C4821}>) 86: (SB-IMPL::OUTPUT-CHAR-UTF-8-LINE-BUFFERED #<SB-SYS:FD-STREAM for "the terminal" {A7C4821}> #\Newline) 87: (WRITE-CHAR #\Newline #<SB-SYS:FD-STREAM for "the terminal" {A7C4821}>) 88: (WRITE-CHAR #\Newline #<SYNONYM-STREAM :SYMBOL SB-SYS:*TTY* {910E8F1}>) 89: (TERPRI #<SYNONYM-STREAM :SYMBOL *TERMINAL-IO* {915C2C1}>) 90: (SB-FORMAT::%-FORMAT-DIRECTIVE-INTERPRETER #<SYNONYM-STREAM :SYMBOL *TERMINAL-IO* {915C2C1}> #<~%> NIL #<unavailable argument> NIL) 91: (SB-FORMAT::INTERPRET-DIRECTIVE-LIST #<SYNONYM-STREAM :SYMBOL *TERMINAL-IO* {915C2C1}> (#<~%>) (32933) NIL) 92: (SB-FORMAT::%FORMAT #<SYNONYM-STREAM :SYMBOL *TERMINAL-IO* {915C2C1}> "~&;; Swank started at port: ~D.~%" (32933) (32933)) 93: (FORMAT #<SYNONYM-STREAM :SYMBOL *TERMINAL-IO* {915C2C1}> "~&;; Swank started at port: ~D.~%") 94: (SWANK::SIMPLE-ANNOUNCE-FUNCTION 32933) 95: (SWANK::SETUP-SERVER 0 #<CLOSURE (LAMBDA (SWANK::PORT)) {C481565}> :SPAWN NIL :UTF-8-UNIX) 96: (SWANK:START-SERVER "/tmp/slime.18179" :STYLE NIL :DONT-CLOSE NIL :EXTERNAL-FORMAT :UTF-8-UNIX) 97: (SB-INT:EVAL-IN-LEXENV (SWANK:START-SERVER "/tmp/slime.18179" :EXTERNAL-FORMAT :UTF-8-UNIX) #<NULL-LEXENV>) 98: (INTERACTIVE-EVAL (SWANK:START-SERVER "/tmp/slime.18179" :EXTERNAL-FORMAT :UTF-8-UNIX)) 99: (SB-IMPL::REPL-FUN NIL) 100: ((LAMBDA NIL)) 101: ((LAMBDA NIL)) 102: (SB-IMPL::%WITH-REBOUND-IO-SYNTAX #<CLOSURE (LAMBDA NIL) {A7C3C6D}>) 103: (SB-IMPL::TOPLEVEL-REPL NIL) 104: (SB-IMPL::TOPLEVEL-INIT) 105: ((LABELS SB-IMPL::RESTART-LISP)) debugger invoked on a SIMPLE-ERROR in thread #<THREAD "initial thread" {A7C4489}>: Maximum error nesting depth exceeded 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-IMPL::SIMPLE-STREAM-PERROR "couldn't write to ~S" #<SB-SYS:FD-STREAM for "the terminal" {A7C4821}> 32) 0] -- "You do not really understand something unless you can explain it to your grandmother." — Albert Einstein.

3 2

[slime-devel] RE: arglist for macros in clisp
by Anton Vodonosov 15 Sep '06

15 Sep '06

>...but it seems really fragile to me. is >there not somethnig else we can do? I saw someone trying to shadow standard defmacro with its own version: http://groups.google.com/group/comp.lang.lisp/browse_thread/thread/dff827eeb 21412d4/cc2a69182c2734a#cc2a69182c2734a Another way is to implement needed functionality in clisp. Maybe something else... -----Original Message----- From: Marco Baringer [mailto:mb@bese.it] Sent: Friday, September 15, 2006 7:43 AM To: Vodonosov Anton Subject: Re: arglist for macros in clisp "Vodonosov Anton" <vodonosov(a)mail.ru> writes: > Of course, this solution is implementation-tied, treats &rest and &body > parameters equally and works only for interpreted code. But it works (seems > to be). the fact that it's implementation tied is not a problem (it is swank-CLISP.lisp after all), but it seems really fragile to me. is there not somethnig else we can do? -- -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

1 0

[slime-devel] arglist for macros in clisp
by Vodonosov Anton 15 Sep '06

15 Sep '06

In clisp swank is unable to get arglist for macros. As a result emacs is unable to properly indent macro call, when macro has a &body parameter. I've just developed a solution. It works for interpreted macros, but not for compiled. Lets define two macros: (defmacro m-with-body (a &body b) nil) (defmacro m-without-body (a b) nil) And then look to the macro expansion function associated with the symbols m-with-body and m-without body: (function-lambda-expression (macro-function 'm-with-body)) (LAMBDA (SYSTEM::<MACRO-FORM> SYSTEM::<ENV-ARG>) (DECLARE (CONS SYSTEM::<MACRO-FORM>)) (DECLARE (IGNORE SYSTEM::<ENV-ARG>)) (IF (< (LIST-LENGTH-DOTTED SYSTEM::<MACRO-FORM>) 2) (SYSTEM::MACRO-CALL-ERROR SYSTEM::<MACRO-FORM>) (LET* ((A (CADR SYSTEM::<MACRO-FORM>)) (B (CDDR SYSTEM::<MACRO-FORM>))) (BLOCK M-WITH-BODY NIL)))) #(NIL NIL NIL NIL ((DECLARATION OPTIMIZE DECLARATION))) M-WITH-BODY (function-lambda-expression (macro-function 'm-without-body)) (LAMBDA (SYSTEM::<MACRO-FORM> SYSTEM::<ENV-ARG>) (DECLARE (CONS SYSTEM::<MACRO-FORM>)) (DECLARE (IGNORE SYSTEM::<ENV-ARG>)) (IF (/= (LIST-LENGTH-DOTTED SYSTEM::<MACRO-FORM>) 3) (SYSTEM::MACRO-CALL-ERROR SYSTEM::<MACRO-FORM>) (LET* ((A (CADR SYSTEM::<MACRO-FORM>)) (B (CADDR SYSTEM::<MACRO-FORM>))) (BLOCK M-WITHOUT-BODY NIL)))) #(NIL NIL NIL NIL ((DECLARATION OPTIMIZE DECLARATION))) M-WITHOUT-BODY Note, that variable B is bound to (CDDR SYSTEM::<MACRO-FORM>) for macro with &body argument and to (CADDR SYSTEM::<MACRO-FORM>) for macro without &body argument. Based on this difference I've changed the arglist function in the swank-clisp.lisp file: (defimplementation arglist (fname) (block nil (or (ignore-errors (let ((exp (function-lambda-expression fname))) (and exp (return (second exp))))) (ignore-errors (return (ext:arglist fname))) (ignore-errors (return (macro-arglist fname))) :not-available))) (defun macro-arglist (macro-symbol) (let* ((expansion-fn (function-lambda-expression (macro-function macro-symbol))) (if-form (cadddr (cdr expansion-fn))) (let-form (cadddr if-form)) (let-vars (cadr let-form)) (arg-list nil)) (dolist (var-def let-vars) (if (char-equal (char (symbol-name (caadr var-def)) 1) #\A) (setq arg-list (append arg-list (list (car var-def)))) (setq arg-list (append arg-list (list '&body (car var-def)))))) arg-list)) Of course, this solution is implementation-tied, treats &rest and &body parameters equally and works only for interpreted code. But it works (seems to be). I've tried this with clisp-2.34 under Windows. Regards, -Anton

1 0

[slime-devel] Daily ChangeLog diff
by mbaringer＠common-lisp.net 14 Sep '06

14 Sep '06

Index: slime/ChangeLog diff -u slime/ChangeLog:1.937 slime/ChangeLog:1.945 --- slime/ChangeLog:1.937 Mon Sep 11 04:01:58 2006 +++ slime/ChangeLog Wed Sep 13 18:56:14 2006 @@ -1,4 +1,48 @@ -2006-09-1 Nikodemus Siivola <nikodemus(a)random-state.net> +2006-09-14 Douglas Crosher <dcrosher(a)common-lisp.net> + + * swank-scl (arglist, function-arglist, spawn): update for the SCL. + +2006-09-13 Brandon Bergren <bdragon(a)mailsnare.net> + + * slime.el (slime-filename-translations): Fix docstring + +2006-09-13 Bob Halley <halley(a)play-bow.us> + + * swank.lisp (format-iso8601-time): Properly handle non integer + time zones. + +2006-09-13 Taylor R Campbell <campbell(a)mumble.net> + + * slime.el (slime-init-output-buffer): Initial directory and + package stacks should be empty. + (slime-repl-push-package): Push the current package, as opposed to + the new package, and set the new package to whatever the user + specified. + (slime-repl-pop-package): Set the current package to the top of + the package stack, unless it's empty. + +2006-09-13 Daniel Koning <dkoning(a)seas.smu.edu> + + * slime.el (slime-repl-disconnect): New repl shortcut. + +2006-09-13 Marco Baringer <mb(a)bese.it> + + * slime.el (slime-open-inspector): Added a slime-part-number + property to the topline so that you can slime-inspector-copy-down + the object being inspected. There are some cases where we have an + object in the inspector and we'd like to dump it to the repl but + we can't get at it through other means (like in back-traces). + (slime-insert-xrefs): Specify which file the item is in (when that + information is available). + + * swank.lisp (format-arglist-for-echo-area): Instead of using + let+first+rest to destructure a form use destructuring-bind. + (lookup-presented-object): Added (declare (special + *inspectee-parts*)) to silence openmcl's compiler. + (inspect-object): Generate, and send to emacs, an ID for the + object being inspected. + +2006-09-01 Nikodemus Siivola <nikodemus(a)random-state.net> * slime.el (slime-repl-matching-input-regexp): Use the portion between slime-repl-input-mark and point for history search, not

1 0