August 2007 - slime-devel - mailman3.common-lisp.net

[slime-devel] Daily ChangeLog diff
by mbaringer＠common-lisp.net 24 Aug '07

24 Aug '07

Index: slime/ChangeLog diff -u slime/ChangeLog:1.1145 slime/ChangeLog:1.1153 --- slime/ChangeLog:1.1145 Wed Aug 22 17:41:00 2007 +++ slime/ChangeLog Thu Aug 23 18:50:57 2007 @@ -1,9 +1,179 @@ +2007-08-23 Matthias Koeppe <mkoeppe(a)mail.math.uni-magdeburg.de> + + Repair inspection of presentations. + + * swank.lisp (inspect-presentation): New slimefun. + * slime.el (slime-inspect-presentation-at-mouse): Use it here. + +2007-08-23 Helmut Eller <heller(a)common-lisp.net> + + Move Marco Baringer's inspector to contrib. + + * swank.lisp (*default-inspector*): New variable. Set this + variable dispatch to different inspectors. + (inspect-object): Use it. + + * swank-loader.lisp (*contribs*): Add 'swank-fancy-inspector. + + * swank-backend.lisp (backend-inspector): New class. Introduce a + named class to give as another way to dispatch to backend methods. + + * swank-cmucl.lisp: Use backend-inspector class. + * swank-sbcl.lisp: Use backend-inspector class. + * swank-clisp.lisp: Use backend-inspector class. + * swank-lispworks.lisp: Use backend-inspector class. + * swank-allegro.lisp: Use backend-inspector class. + * swank-openmcl.lisp: Use backend-inspector class. + * swank-abcl.lisp: Use backend-inspector class. + * swank-corman.lisp: Use backend-inspector class. + * swank-scl.lisp: Use backend-inspector class. + +2007-08-23 Tobias C. Rittweiler <tcr(a)freebits.de> + + Added arglist display for declaration specifiers and type + specifiers. + + Examples: + + `(declare (type' will display + + (declare (type type-specifier &rest vars)) + + `(declare (type (float' will display + + [Typespec] (float &optional lower-limit upper-limit) + + `(declare (optimize' will display + + (declare (optimize &any (safety 1) (space 1) (speed 1) ...)) + + &ANY is a new lambda keyword that is introduced for arglist + description purpose, and is very similiar to &KEY, but isn't based + upon plists; they're more based upon *FEATURES* lists. (See the + comment near the ARGLIST defstruct in `swank.lisp'.) + + * slime.el: + (slime-to-feature-keyword): Renamed to `slime-keywordify'. + (slime-eval-feature-conditional): Adapted to use `slime-keywordify'. + (slime-ensure-list): New utility. + (slime-sexp-at-point): Now takes an argument that specify how many + sexps at point should be returned. + (slime-enclosing-operator-names): Renamed to + `slime-enclosing-form-specs'. + (slime-enclosing-form-specs): Returns a list of ``raw form specs'' + instead of what was called ``extended operator names'' before, see + `swank::parse-form-spec' for more information. This is a + simplified superset. Additionally as tertiary return value return + a list of points to let the caller see where each form spec is + located. Adapted callers accordingly. Extended docstring. + (slime-parse-extended-operator-name): Adapted to changes in + `slime-enclosing-form-specs'. Now gets more context, and is such + more powerful. This was needed to allow parsing DECLARE forms. + (slime-make-extended-operator-parser/look-ahead): Because the + protocol for arglist display was simplified, it was possible to + replace the plethora of parsing function just by this one. + (slime-extended-operator-name-parser-alist): Use it. Also add + parser for DECLARE forms. + (slime-parse-extended-operator/declare): Responsible for parsing + DECLARE forms. + (%slime-in-mid-of-typespec-p): Helper function for + `slime-parse-extended-operator/declare'. + (slime-incomplete-form-at-point): New. Return the ``raw form + spec'' near point. + (slime-complete-form): Use `slime-incomplete-form-at-point'. + + * swank.lisp: New Helper functions. + (length=, ensure-list, recursively-empty-p): New. + (maybecall, exactly-one-p): New. + + * swank.lisp (arglist-for-echo-area): Adapted to take ``raw form + specs'' from Slime. + (parse-form-spec): New. Takes a ``raw form spec'' and returns a + ``form spec'' for further processing in Swank. Docstring documents + these two terms. + (split-form-spec): New. Return relevant information from a form spec. + (parse-first-valid-form-spec): Replaces `find-valid-operator-name'. + (find-valid-operator-name): Removed. + (operator-designator-to-form): Removed. Obsoleted by `parse-form-spec'. + + (defstruct arglist): Add `any-p' and `any-args' slots to contain + arguments belonging to the &ANY lambda keyword. + (print-arglist): Adapted to also print &ANY args. + (print-decoded-arglist-as-template): Likewise. + (decode-arglist): Adapted to also decode &ANY args. + (remove-actual-args): Adapted to also remove &ANY args. + (remove-&key-args): Split out from `remove-actual-args'. + (remove-&any-args): New. Removes already provided &ANY args. + (arglist-from-form-spec): New. Added detailed docstring. + (arglist-dispatch): Dispatching generic function for + `arglist-from-form-spec' that does all the work. Renamed from + prior `form-completion'. + (arglist-dispatch) Added methods for dealing with declaration and + type-specifiers. + (complete-form): Adapted to take ``raw form specs'' from Slime. + (completions-for-keyword): Likewise. + (format-arglist-for-echo-area): Removed. Not needed anymore. + + * swank-backend.lisp (declaration-arglist): New generic + function. Returns the arglist for a given declaration + identifier. (Backends are supposed to specialize it if they can + provide additional information.) + (type-specifier-arglist): New generic function. Returns the + arglist for a given type-specifier operator. (Backends are + supposed to specialize it if they can provide additional + information.) + (*type-specifier-arglists*): New variable. Contains the arglists + for the type specifiers in Common Lisp. + + * swank-sbcl.lisp: Now depends upon sb-cltl2. + (declaration-arglist 'optimize): Specialize the `optimize' + declaration identifier to pass it to + sb-cltl2:declaration-information. + + +2007-08-23 Helmut Eller <heller(a)common-lisp.net> + + Some inspector cleanups. + + * slime.el (slime-inspect): Remove dwim stuff and drop keyword + args. + (slime-read-object): Killed. + (slime-open-inspector): Drop keyword args. Update callers + accodordingly, expect presentation related code. Presentations no + longer work in the inspector. + + * swank.lisp (*inspector-dwim-lookup-hooks*) + (default-dwim-inspector-lookup-hook): Deleted. + (init-inspector): Sanitize arglist. + (inspect-object): Don't return an :id for *inspectee-parts*. + + * swank-backend (type-for-emacs): Removed. No backend implemented + it. + +2007-08-23 Helmut Eller <heller(a)common-lisp.net> + + * slime.el (slime-fuzzy-upgrade-notice): New function. Bound to + the key where slime-fuzzy-complete-symbol used to be. + 2007-08-22 Tobias C. Rittweiler <tcr(a)freebits.de> * slime.el (slime-close-all-parens-in-sexp): Fix interplay with `slime-close-parens-limit'. This should also affect `slime-complete-form' (C-c C-s) in a positive way. +2007-08-19 Helmut Eller <heller(a)common-lisp.net> + + * contrib: New directory. Move fuzzy completion code to that + directory. + + * swank.lisp (swank-require): New function to load contrib code. + (*find-module*, module-filename, *load-path*, merged-directory) + (find-module, module-canditates): New. Pathname acrobatics for + swank-require. + + * swank-loader.lisp: Compile (but don't load) contribs. + (*contribs*, contrib-source-files): New. + 2007-08-16 Tobias C. Rittweiler <tcr(a)freebits.de> * slime.el (slime-process-available-input): Correct yesterday's @@ -15,7 +185,7 @@ to send a `:debug-return' message before the acknowledgement message for `sldb-quit' is sent. So the acknowledgement is received in a context where the sldb-buffer is closed already. - + 2007-08-15 Tobias C. Rittweiler <tcr(a)freebits.de> * slime.el (slime-process-available-input): Make sure that the @@ -24,7 +194,7 @@ Previously, the clauses of `slime-rex' were processed in the internal *cl-connection* buffer. And as a result the continuations passed to `slime-eval' and `slime-eval-async' ditto. - + 2007-08-15 Tobias C. Rittweiler <tcr(a)freebits.de> Make `M-.' work on definitions outside the current restriction.

1 0

[slime-devel] Inspector in contrib
by Helmut Eller 23 Aug '07

23 Aug '07

Hi, I moved most of the inspector code to contrib and called the module "fancy-inspector". Slime still has a (non-fancy) inspector which shows whatever the backend returns. Except, for lists, arrays, and hashtables. On the implementation side, this is done by dispatching on the second argument of inspect-for-emacs. The code in the contrib is almost unchanged. Only the specializers of the second argument are different. To use the fancy inspector add the usual: (add-to-list 'load-path "<contrib-directory>") (add-hook 'slime-load-hook (lambda () (require 'slime-fancy-inspector))) to your .emacs. Helmut.

1 0

[slime-devel] How would you like unlimited hits to your website 15 minutes from now ?
by Hit-Booster＠common-lisp.net 23 Aug '07

23 Aug '07

1 0

[slime-devel] Daily ChangeLog diff
by mbaringer＠common-lisp.net 23 Aug '07

23 Aug '07

Index: slime/ChangeLog diff -u slime/ChangeLog:1.1144 slime/ChangeLog:1.1145 --- slime/ChangeLog:1.1144 Thu Aug 16 04:27:31 2007 +++ slime/ChangeLog Wed Aug 22 17:41:00 2007 @@ -1,3 +1,9 @@ +2007-08-22 Tobias C. Rittweiler <tcr(a)freebits.de> + + * slime.el (slime-close-all-parens-in-sexp): Fix interplay with + `slime-close-parens-limit'. This should also affect + `slime-complete-form' (C-c C-s) in a positive way. + 2007-08-16 Tobias C. Rittweiler <tcr(a)freebits.de> * slime.el (slime-process-available-input): Correct yesterday's

1 0

[slime-devel] Contrib branch: File error: Cannot open load file, slime-fuzzy
by Frank Goenninger 21 Aug '07

21 Aug '07

Hi all: I just updated Slime from CVS and ... it did not install the CONTRIB branch. Oh, yes, the contrib directory is there but it just contains a CVS directory and no files. Using CVS commands manually I was not able to download any slime-fuzzy.* files ... Adding the two lines to .emacs file leads to the error indicated in the subject. Could you please restore the old structure that made it simple and easy to use/update/install SLIME ?! Thx! Best regards Frank

2 2

[slime-devel] indent functions broken under Emacs-22.1?
by Jeff Cunningham 21 Aug '07

21 Aug '07

A short time ago both slime-indent-sexp and slime-reindent-defun stopped working, the former giving the error message: slime-repl-at-prompt-start-p: Wrong type argument: number-or-marker-p, nil and the latter doing nothing whatever. Since I had just installed emacs-22.1, cedet and a couple other modes, I figured it was an interaction problem and would look at it later. Now is later and in the course of trying to debug this, I've reduced my .emacs to nothing but the following: (setq inferior-lisp-program "/usr/local/bin/sbcl") (add-to-list 'load-path "/home/jcunningham/site/slime") (require 'slime) (slime-setup) (global-set-key "\C-cs" 'slime-selector) And there is nothing in the load path (I moved everything elsewhere). Slime is straight out of CVS this morning. The breakage persists. So its starting to look like a problem in slime or between slime and emacs-22.1. Has anyone else run into this? --Jeff

2 3

[slime-devel] Re: [slime-cvs] CVS slime
by Helmut Eller 19 Aug '07

19 Aug '07

* trittweiler [2007-08-15 15:45+0200] writes: > Make `M-.' work on definitions outside the current restriction. > `M-,' will also properly restore the narrowing as of before the > jump. Similiarly for quiting from the compilation notes buffer and > the Xref buffers. > > * slime.el (slime-narrowing-configuration, slime-emacs-snapshot), > (current-slime-narrowing-configuration), > (set-slime-narrowing-configuration), > (current-slime-emacs-snapshot), > (set-slime-emacs-snapshot), > (current-slime-emacs-snapshot-fingerprint): New. Emacs' window > configurations do not restore narrowing, so introduce a > snapshot facility that contains the necessary information. This is quite a heavy abstraction. Can we at least agree to use "slime" as prefix in ELisp code please? May I also point out that "slime-emacs-snapshot" is a rather unspecific term? Perhaps something like "slime-window-config" would be more suggestive. Or at least drop the "emacs" part. Helmut.

3 4

[slime-devel] Daily ChangeLog diff
by mbaringer＠common-lisp.net 17 Aug '07

17 Aug '07

Index: slime/ChangeLog diff -u slime/ChangeLog:1.1143 slime/ChangeLog:1.1144 --- slime/ChangeLog:1.1143 Wed Aug 15 11:11:24 2007 +++ slime/ChangeLog Thu Aug 16 04:27:31 2007 @@ -1,3 +1,15 @@ +2007-08-16 Tobias C. Rittweiler <tcr(a)freebits.de> + + * slime.el (slime-process-available-input): Correct yesterday's + change: the buffer a request was originally performed in doesn't + necessarily exist at this time anymore, so we check for buffer + liveness now. + + The problem arised when quitting in SLDB which would cause Swank + to send a `:debug-return' message before the acknowledgement + message for `sldb-quit' is sent. So the acknowledgement is + received in a context where the sldb-buffer is closed already. + 2007-08-15 Tobias C. Rittweiler <tcr(a)freebits.de> * slime.el (slime-process-available-input): Make sure that the

1 0

[slime-devel] (string #\U65B0)
by Chris Laux 16 Aug '07

16 Aug '07

Hi all, I'm working with data that contains foreign characters using SLIME and get the errors below when these become visible in the REPL. > error in process filter: byte-code: net-read error: (end-of-file) > error in process filter: net-read error: (end-of-file) This can be provoked by evaluating e.g.: (string #\U65B0) I tried the remedies discussed previously for similar problems, without solving the problem. Interestingly something like "(string (code-char #x2222))" will actually work without complaining. For reference, I am using SLIME CVS, Emacs 21.4.1 and SBCL 1.0.5.34 and I ensured the following settings: swank::*coding-system* --> "utf-8-unix" (setf slime-lisp-implementations '((sbcl ("sbcl") :coding-system utf-8-unix))) This topic has come up a few times, this thread seems to be most similar: http://thread.gmane.org/gmane.lisp.slime.devel/5895/focus=5901 Unfortunately it doesn't conclude with a solution. Has there been progress in the meantime? Thanks, Chris

2 2

[slime-devel] Daily ChangeLog diff
by mbaringer＠common-lisp.net 16 Aug '07

16 Aug '07

Index: slime/ChangeLog diff -u slime/ChangeLog:1.1141 slime/ChangeLog:1.1143 --- slime/ChangeLog:1.1141 Sun Aug 5 23:20:28 2007 +++ slime/ChangeLog Wed Aug 15 11:11:24 2007 @@ -1,3 +1,57 @@ +2007-08-15 Tobias C. Rittweiler <tcr(a)freebits.de> + + * slime.el (slime-process-available-input): Make sure that the + event received from SWANK is processed in the context of the + original buffer the request of the response was performed in. + Previously, the clauses of `slime-rex' were processed in the + internal *cl-connection* buffer. And as a result the continuations + passed to `slime-eval' and `slime-eval-async' ditto. + +2007-08-15 Tobias C. Rittweiler <tcr(a)freebits.de> + + Make `M-.' work on definitions outside the current restriction. + `M-,' will also properly restore the narrowing as of before the + jump. Similiarly for quiting from the compilation notes buffer and + the Xref buffers. + + * slime.el (slime-narrowing-configuration, slime-emacs-snapshot), + (current-slime-narrowing-configuration), + (set-slime-narrowing-configuration), + (current-slime-emacs-snapshot), + (set-slime-emacs-snapshot), + (current-slime-emacs-snapshot-fingerprint): New. Emacs' window + configurations do not restore narrowing, so introduce a + snapshot facility that contains the necessary information. + + * slime.el: Various renaming and adaptions in the Slime temp + buffer, xref, goto-definition and compilation notes section to use + the newly introduced snapshots instead of mere window + configurations. + + * slime.el: (slime-highlight-notes, slime-remove-old-overlays): + Still operate on whole buffer, but restore previous restriction if + there was any. + (slime-goto-location-position): Now widens the buffer to properly + jump to definitions outside of the current restriction. + + * slime.el (slime-push-definition-stack), + (slime-pop-find-definition-stack): Now also stores information + about narrowing on the definition stack, in order to properly + restore narrowing on `M-,'. + + * slime.el (def-slime-test narrowing): Test case for properly + dealing with narrowing. + + * slime.el (slime-buffer-narrowed-p): New function, tests whether + the current buffer is narrowed or not. + (save-restriction-if-possibly): Like `save-restriction', but not + as strict---see doc string. + + * slime.el (slime-length=): New function; semantically the same + as (= (length seq) n), but more efficiently implemented for lists. + Changed the above pattern into a call to SLIME-LENGTH= where + appropriate. + 2007-08-05 Matthias Koeppe <mkoeppe(a)mail.math.uni-magdeburg.de> * swank.lisp (backtrace): Handle printer errors while printing a

1 0