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

[slime-devel] Daily ChangeLog diff
by Helmut Eller 27 Nov '09

27 Nov '09

Index: slime/ChangeLog diff -u slime/ChangeLog:1.1921 slime/ChangeLog:1.1922 --- slime/ChangeLog:1.1921 Thu Nov 26 02:06:50 2009 +++ slime/ChangeLog Thu Nov 26 05:58:08 2009 @@ -3,7 +3,7 @@ * swank-abcl.lisp (arglist): Fixes for functions with non-nil arglist and for generic functions with empty argument lists. - Contributed by Matthias. + Diagnosed and cured by Matthias Hölzl. 2009-11-23 Stas Boukarev <stassats(a)gmail.com>

1 0

[slime-devel] Daily ChangeLog diff
by Helmut Eller 26 Nov '09

26 Nov '09

Index: slime/ChangeLog diff -u slime/ChangeLog:1.1920 slime/ChangeLog:1.1921 --- slime/ChangeLog:1.1920 Mon Nov 23 14:54:55 2009 +++ slime/ChangeLog Thu Nov 26 02:06:50 2009 @@ -1,3 +1,10 @@ +2009-11-26 Mark Evenson <evenson(a)panix.com>> + + * swank-abcl.lisp (arglist): Fixes for functions with non-nil + arglist and for generic functions with empty argument lists. + + Contributed by Matthias. + 2009-11-23 Stas Boukarev <stassats(a)gmail.com> * slime.el (slime-set-connection-info): Set slime-current-thread to t

1 0

[slime-devel] [patch] ABCL arglist display
by Matthias Hölzl 26 Nov '09

26 Nov '09

The current implementation of swank::arglist for ABCL seems to be not quite correct. The attached patch fixes two problems: 1. In the form (multiple-value-bind (arglist present) (or (sys::arglist ...) ...) the value of "present" is always nil if sys::arglist returns a non-nil arglist (since, for any form but the last, "or" returns the primary value only). Hence swank::arglist always returns :not-available if sys::arglist returns a non-nil primary value, therefore no arglist info is displayed for non-generic functions. The attached patch fixes this. 2. The current implementation returns :not-available for generic functions with empty argument list. This does not seem to be quite correct since generic functions without parameters are allowed according to Section 3.4.2 of the HyperSpec. The attached patch changes this by always setting "present" to true when the function is of type "standard-generic-function"; this seems to be correct according to the definition of generic-function-lambda-list in AMOP, p. 216. Best regards, Matthias

2 1

[slime-devel] Daily ChangeLog diff
by Helmut Eller 25 Nov '09

25 Nov '09

Index: slime/contrib/ChangeLog diff -u slime/contrib/ChangeLog:1.286 slime/contrib/ChangeLog:1.287 --- slime/contrib/ChangeLog:1.286 Mon Nov 23 16:48:52 2009 +++ slime/contrib/ChangeLog Tue Nov 24 08:17:00 2009 @@ -1,3 +1,23 @@ +2009-11-24 Tobias C. Rittweiler <tcr(a)freebits.de> + + Fix a few edge cases in new arglist code. + + * slime-parse.el (slime-parse-form-upto-point): Regard + beginning-of-line as whitespace, and DTRT. + + * swank-arglists.lisp (empty-arg-p): Input may not only be an + arglist-dummy. + (print-decoded-arglist-as-template): Do not print superfluuous + newline before &body. + (arglist-for-echo-area): Catch errors. + (find-subform-with-arglist): Deal properly with NIL as argument. + (find-immediately-containing-arglist): Do not erroneously complete + form with an unsuited arglist of the parent form. + (last-arg): New helper. + (arglist-path-to-nested-arglist): Use it. + + Reported by Ariel Badichi. + 2009-11-23 Tobias C. Rittweiler <tcr(a)freebits.de> * slime-asdf.el (slime-rgrep-system): Conditionalize on whether

1 0

[slime-devel] *default-pathname-defaults* not reset after file compilation?
by Slawek Zak 24 Nov '09

24 Nov '09

Hi, Current slime doesn't reset *default-pathname-defaults* after file has been compiled to it's previous value - is there a reason for it? Thanks, /S

2 2

[slime-devel] swank-arglists fragility
by Ariel Badichi 24 Nov '09

24 Nov '09

Hello, The arglist mechanism is too vocal when presumptions about its input are not met. A case in point, SWANK::EXTRACT-LOCAL-OP-ARGLISTS expects the second element of the supplied "form", given that there is such an element, to be a cons. If it is not a cons, it will signal an error. A simple example of how this results in unsatisfactory behavior: (defmacro foo (bar zot) `(labels ,bar (| Where | is the place following point. Another example: (defmacro foo (bar zot) `(labels (quux (boo)) (| This time an error may or may not be signaled, depending on whether symbols corresponding to the strings in the source text exist. I know that these particular cases can (and should) be fixed, but I think in general the arglist mechanism shouldn't be allowed to scream so loud, demanding the immediate attention of anyone within debugger range when they're not even interested. Ariel

2 2

[slime-devel] Daily ChangeLog diff
by Helmut Eller 24 Nov '09

24 Nov '09

Index: slime/ChangeLog diff -u slime/ChangeLog:1.1919 slime/ChangeLog:1.1920 --- slime/ChangeLog:1.1919 Sat Nov 21 16:32:28 2009 +++ slime/ChangeLog Mon Nov 23 14:54:55 2009 @@ -1,3 +1,8 @@ +2009-11-23 Stas Boukarev <stassats(a)gmail.com> + + * slime.el (slime-set-connection-info): Set slime-current-thread to t + before doing anything. Solves a bug reported by Slawek Zak. + 2009-11-21 Tobias C. Rittweiler <tcr(a)freebits.de> * swank-sbcl.lisp (who-specializes): Implement. Index: slime/contrib/ChangeLog diff -u slime/contrib/ChangeLog:1.282 slime/contrib/ChangeLog:1.286 --- slime/contrib/ChangeLog:1.282 Sun Nov 22 08:03:10 2009 +++ slime/contrib/ChangeLog Mon Nov 23 16:48:52 2009 @@ -1,3 +1,30 @@ +2009-11-23 Tobias C. Rittweiler <tcr(a)freebits.de> + + * slime-asdf.el (slime-rgrep-system): Conditionalize on whether + `rgrep' is available (it isn't on Emacs 21.) Also make sure to + call `grep-compute-defaults' before `grep-read-regexp' because the + former does some necessary setup. (Seems to be of a problem on + Emacs from CVS.) + +2009-11-23 Tobias C. Rittweiler <tcr(a)freebits.de> + + * slime-repl.el (slime-repl-previous-matching-input) + (slime-repl-next-matching-input): Read input by means of + `slime-read-from-minibuffer' so TAB will complete symbols for us. + +2009-11-23 Tobias C. Rittweiler <tcr(a)freebits.de> + + * swank-asdf.lisp (operate-on-system-for-emacs), + (operate-on-system): Muffle ASDF:COMPILE-ERROR because we reuse + Slime's compilation error reportery anyway, and sldb popping up is + just annoying. + +2009-11-23 Tobias C. Rittweiler <tcr(a)freebits.de> + + * swank-asdf.lisp (system-contains-file-p): Previous definition + didn't properly propagate positive return value of recursive + call. Fix that. + 2009-11-22 Stas Boukarev <stassats(a)gmail.com> * slime-repl.el (slime-repl-history-remove-duplicates): New variable,

1 0

[slime-devel] Wire protocol cleanup issue??
by Slawek Zak 23 Nov '09

23 Nov '09

Hi, There is some problem with buffer cleanup when lisp connection is broken. Here's a minimal example. 1. Run your lisp remotely and start swank server 2. Connect w/ Emacs 3. Run remotely: pkill -9 <your favorite lisp> # the lisp dies 'orrible death 4. Restart lisp 5. Reconnect w/ Emacs Following messages get logged when you try to clean buffer (C-c M-o): error in process filter: slime-dispatch-event: Invalid rpc: Thread not found: :REPL-THREAD error in process filter: Invalid rpc: Thread not found: :REPL-THREAD error in process filter: slime-dispatch-event: Invalid rpc: Thread not found: :REPL-THREAD error in process filter: Invalid rpc: Thread not found: :REPL-THREAD Connected. Lemonodor-fame is but a hack away! error in process filter: slime-dispatch-event: Invalid protocol message error in process filter: Invalid protocol message And a slime error too: Invalid protocol message: Symbol "CLEAR-REPL-RESULTS" not found in the SWANK package. [file position = 37] "(:emacs-rex (swank:clear-repl-results) \"USER\" :repl-thread 7) " Thanks!, /S

2 2

[slime-devel] Error/warning location reporting in ACL 8.2 beta vs slime
by Slawek Zak 23 Nov '09

23 Nov '09

Hi, There is an issue with parsing of error/warning location during compilation in the all-new ACL 8.2 beta. Attached patch solves the issue. Thanks!, /S

1 0

[slime-devel] Daily ChangeLog diff
by Helmut Eller 23 Nov '09

23 Nov '09

Index: slime/contrib/ChangeLog diff -u slime/contrib/ChangeLog:1.280 slime/contrib/ChangeLog:1.282 --- slime/contrib/ChangeLog:1.280 Sat Nov 21 11:27:55 2009 +++ slime/contrib/ChangeLog Sun Nov 22 08:03:10 2009 @@ -1,3 +1,28 @@ +2009-11-22 Stas Boukarev <stassats(a)gmail.com> + + * slime-repl.el (slime-repl-history-remove-duplicates): New variable, + if set to T previous matching history entries are removed before + appending a new item. Default value is NIL. + (slime-repl-history-trim-whitespaces): New variable, when T remove + whitespaces at the beginning and end of a new history item. + Default value is NIL. + (slime-repl-add-to-input-history): Implement behaviour of the variables + above. + (slime-string-trim): New function, works like cl:string-trim. + +2009-11-22 Tobias C. Rittweiler <tcr(a)freebits.de> + + * slime-fontifying-fu.el (slime-compute-region-for-font-lock): Set + an upper limit for the starting point of searching for suppressed + forms. + + * slime-parse.el (slime-make-form-spec-from-string): Minor + optimizations. + (slime-parse-form-upto-point): Refactored to not use `reduce' but + bultins. + (slime-make-form-spec-from-string, slime-parse-form-upto-point) + (slime-compare-char-syntax): Byte-compile. + 2009-11-21 Tobias C. Rittweiler <tcr(a)freebits.de> * swank-asdf.lisp (asdf-determine-system): Also try to determine

1 0