November 2009 - slime-devel - mailman3.common-lisp.net

[slime-devel] Why are column 1 `(' highlighted in #| ... |# blocks?
by Mirko Vukovic 20 Nov '09

20 Nov '09

Hello, In the #| ... |# comment blocks, opening parenthesis in column 1 (first column) are highlighted in red. What is the reason for that? Thanks, Mirko

2 2

[slime-devel] Daily ChangeLog diff
by Helmut Eller 20 Nov '09

20 Nov '09

Index: slime/ChangeLog diff -u slime/ChangeLog:1.1917 slime/ChangeLog:1.1918 --- slime/ChangeLog:1.1917 Fri Nov 13 15:23:57 2009 +++ slime/ChangeLog Thu Nov 19 08:37:44 2009 @@ -1,3 +1,10 @@ +2009-11-19 Tobias C. Rittweiler <tcr(a)freebits.de> + + * slime.el (slime-minibuffer-map): Nee `slime-read-expression-map' + (slime-minibuffer-history): Nee `slime-read-expression-history' + (slime-minibuffer-setup-hook): Extracted. + (slime-read-from-minibuffer): Adapted accordingly. + 2009-11-13 Stas Boukarev <stassats(a)gmail.com> * swank.lisp (classify-symbol): Check fbound before calling fdefinition, Index: slime/contrib/ChangeLog diff -u slime/contrib/ChangeLog:1.277 slime/contrib/ChangeLog:1.278 --- slime/contrib/ChangeLog:1.277 Tue Nov 17 15:38:22 2009 +++ slime/contrib/ChangeLog Thu Nov 19 08:37:45 2009 @@ -1,3 +1,12 @@ +2009-11-19 Tobias C. Rittweiler <tcr(a)freebits.de> + + * slime-asdf.el (slime-query-replace-system): Enable TAB + completion of symbol names. + + * slime-fuzzy.el (slime-fuzzy-completions): Do not use + `slime-find-buffer-package' it's an internal operation of + `slime-current-buffer'. + 2009-11-17 Tobias C. Rittweiler <tcr(a)freebits.de> M-x slime-isearch-system will run `isearch-forward' on all files

1 0

[slime-devel] slime-contrib idea: *slime-locations* buffer
by Nikodemus Siivola 19 Nov '09

19 Nov '09

First, have a *slime-locations* buffer or similar, which contains references to positions in files/other buffers: (defun process-toplevel-form ...) [src/compiler/main.lisp] (macrolet ...) [src/compiler/ir1tran.lisp] (defun %compile ...) [src/compiler/main.lisp] one position per line. One should be able to: * kill/yank these lines to delete and reorder them. * hit Enter/Space as in an XREF buffer to jump to the location. * save/load these buffers into files. * add current location in a file into the buffer. For extra points things like C-c C-c in the buffer would be interpreted as the relavant command in the location referred to be the current line. One could use the buffer to maintain a shortlist of relevant locations in the source, etc. For example: I often work on SBCL by editing the live source and recompiling stuff on the fly. Every once and a while I mess up my session, however, and need to restart -- which means hunting through buffers for the definitions to recompile. This is often order-sensitive: add an optional argument to one function, then recompile the callers to pass that thing in, etc. When working on certain bits I may be restarting every few minutes! If I could record those locations in a buffer like this, the workflow would be much nicer! Now that I've written this down I'm guessing that aside from slime-compile-all-locations something like this probably exists as a generic Emacs feature. What is it called? Cheers, -- Nikodemus

1 0

[slime-devel] slime repl AND slime-fancy vs. XEmacs
by Mark H. David 18 Nov '09

18 Nov '09

Hi. Sorry if this these are old issues already. I just recently became a subscriber to this list. I've been using SLIME a few years on Linux, and it was fine, and actually set up by others. Recently, I tried setting it up on Windows by myself. Not so smooth. I followed the basic documentation, which was basically (require 'slime) (slime-setup) But that did not start a REPL, but just the *inferior-lisp* buffer. After a while, I looked around, and I found some "chatter" about having to do (require 'slime-fancy) (slime-setup). That almost worked in XEmacs, and did work in GNU Emacs. Is that really what the average user is supposed to do? The main doc on installation pointed to from here: http://common-lisp.net/project/slime/ which doc is here: http://common-lisp.net/project/slime/doc/html/Installation.html#Installation says this is what you should do: (require 'slime) (slime-setup) I.e., the first thing I tried. So, is the basic doc wrong? What's the deal? Also, regarding XEmacs, it did not work off the bat when I required slime-fancy, but did work after I wrapped (unless (featurep 'xemacs) ...) around (require 'slime-autodoc) in a few slime*.el files. Can we get very simple changes such as this, which happen to be needed to make SLIME work with XEmacs, into the main line? Thanks, Mark

6 12

[slime-devel] Daily ChangeLog diff
by Helmut Eller 18 Nov '09

18 Nov '09

Index: slime/contrib/ChangeLog diff -u slime/contrib/ChangeLog:1.275 slime/contrib/ChangeLog:1.277 --- slime/contrib/ChangeLog:1.275 Mon Nov 16 10:47:54 2009 +++ slime/contrib/ChangeLog Tue Nov 17 15:38:22 2009 @@ -1,3 +1,24 @@ +2009-11-17 Tobias C. Rittweiler <tcr(a)freebits.de> + + M-x slime-isearch-system will run `isearch-forward' on all files + pertaining to an ASDF system. + + M-x slime-query-replace-system will run `query-replace' on all + files pertaining to an ASDF system. + + * slime-asdf.el (slime-read-system-name): Refactored so callers + have choice over how the default value is computed. + (slime-find-asd-file): Renamed from `slime-find-asd'. + (slime-determine-asdf-system): New helper. + (slime-isearch-system): New function. Depends on functionality + only available on GNU Emacs 23.1.x. + (slime-query-replace-system): New function. + +2009-11-17 Stas Boukarev <stassats(a)gmail.com> + + * swank-asdf.lisp (asdf-determine-system): Rewritten to be much + faster and to cons less (and look ugly). + 2009-11-16 Stas Boukarev <stassats(a)gmail.com> * swank-asdf.lisp (asdf-determine-system): New function for

1 0

[slime-devel] Daily ChangeLog diff
by Helmut Eller 17 Nov '09

17 Nov '09

Index: slime/contrib/ChangeLog diff -u slime/contrib/ChangeLog:1.274 slime/contrib/ChangeLog:1.275 --- slime/contrib/ChangeLog:1.274 Fri Nov 13 16:04:26 2009 +++ slime/contrib/ChangeLog Mon Nov 16 10:47:54 2009 @@ -1,3 +1,10 @@ +2009-11-16 Stas Boukarev <stassats(a)gmail.com> + + * swank-asdf.lisp (asdf-determine-system): New function for + determining to what system a file belongs. + + * slime-asdf.el (slime-rgrep-system): New function. + 2009-11-13 Tobias C. Rittweiler <tcr(a)freebits.de> * swank-arglists.lisp (extra-keywords/make-instance)

1 0

[slime-devel] Daily ChangeLog diff
by Helmut Eller 14 Nov '09

14 Nov '09

Index: slime/ChangeLog diff -u slime/ChangeLog:1.1914 slime/ChangeLog:1.1917 --- slime/ChangeLog:1.1914 Fri Nov 6 12:59:22 2009 +++ slime/ChangeLog Fri Nov 13 15:23:57 2009 @@ -1,3 +1,20 @@ +2009-11-13 Stas Boukarev <stassats(a)gmail.com> + + * swank.lisp (classify-symbol): Check fbound before calling fdefinition, + ECL doesn't like (fdefinition nil). + + * swank-ecl.lisp (swank-mop:compute-applicable-methods-using-classes): + Add a dummy function. ECL doesn't have it, but some contribs are using it. + Patch by Andy Hefner. + +2009-11-13 Stas Boukarev <stassats(a)gmail.com> + + * swank-ecl.lisp (function-name): Use clos:generic-function-name + for generic functions. + (arglist): Check fro symbol before calling special-operator-p and + macro-function. + Patch by Andy Hefner. + 2009-11-06 Stas Boukarev <stassats(a)gmail.com> * swank-ecl.lisp (grovel-docstring-for-arglist): ECL's arglists Index: slime/contrib/ChangeLog diff -u slime/contrib/ChangeLog:1.273 slime/contrib/ChangeLog:1.274 --- slime/contrib/ChangeLog:1.273 Fri Nov 6 21:04:56 2009 +++ slime/contrib/ChangeLog Fri Nov 13 16:04:26 2009 @@ -1,3 +1,10 @@ +2009-11-13 Tobias C. Rittweiler <tcr(a)freebits.de> + + * swank-arglists.lisp (extra-keywords/make-instance) + (extra-keywords/change-class): Wrap call to CLASS-PROTOTYPE in an + IGNORE-ERRORS because computing a class-prototype involves + evaluating initforms which may be calls to ERROR. + 2009-11-07 Stas Boukarev <stassats(a)gmail.com> * slime-package-fu.el (slime-find-package-definition-regexp): Go

1 0

[slime-devel] swank-mit-scheme and slime-eval-region
by Derrell Piper 14 Nov '09

14 Nov '09

Hi, If I select the following: (define one 1) (define two 2) (define three 3) ...and do a 'slime-eval-region', the result that's printed in the minibuf is "=> one" and in fact only the first expression was actually evaluated: ; SLIME 2009-11-13 (user)> one 1 (user)> two ; Evaluation aborted. (user)> Trying to evaluation 'two' or 'three' signals 'unbound-variable'. Also, the 'use-value' and 'store-value' restarts are not working, FWIW. Derrell

1 0

[slime-devel] swank-ecl method/generic function lambda list printing bug
by Matthew Mondor 13 Nov '09

13 Nov '09

Hello, I noticed lately (using CVS HEAD slime) with CVS HEAD ECL the following error when attempting to define a PRINT-OBJECT method at the REPL: (defmethod print-object <interrupted:> In function SYMBOL-NAME, the value of symbol is #<STANDARD-GENERIC-FUNCTION PRINT-OBJECT> which is not of expected type SYMBOL [Condition of type SIMPLE-TYPE-ERROR] Backtrace: 0: #<bytecompiled-function 097c4d70> Locals: SWANK::OPERATOR = DEFMETHOD SWANK::ARGUMENTS#1 = (PRINT-OBJECT #S(SWANK::ARGLIST-DUMMY :SWANK::STRING-REPRESENTATION "")) #:G4793#2 = (DEFMETHOD PRINT-OBJECT ..) SWANK::GF-NAME#3 = PRINT-OBJECT SWANK::GF#4 = #<STANDARD-GENERIC-FUNCTION PRINT-OBJECT> 1: #<bytecompiled-function ARGLIST-FROM-FORM> Locals: SWANK::FORM = (DEFMETHOD PRINT-OBJECT ..) 2: #<bytecompiled-function YIELD> 3: #<bytecompiled-function GROVEL-FORM> 4: #<bytecompiled-function FIND-SUBFORM-WITH-ARGLIST> 5: #<bytecompiled-function 097c4a78> 6: #<bytecompiled-function ARGLIST-FOR-ECHO-AREA> 7: #<bytecompiled-function BYTECODES> [...] The same actually happens whenever a generic function exists for the method being created. I have the impression a value which shouldn't be evaluated is, although I didn't look into the related swank-ecl code yet. Possibly someone else already knows exactly what's happening? Thanks, -- Matt

2 1

[slime-devel] [patch] ECL compute-applicable-methods-using-classes
by Andy Hefner 13 Nov '09

13 Nov '09

Currently, attempting to produce the arglist for make-instance given some class name causes an error because the compute-applicable-methods-using-classes function isn't defined by ECL. Attached is a patch that adds a stub version of the function which at least makes it possible to type make-instance calls without interruption by an error. In fact, you even get an arglist, although it isn't necessarily correct - for instance, given a defclass such as (defclass foo () (myslot :initarg myinitarg)), Slime presents the arglist as (make-instance 'foo &rest initargs &key (myslot clos::+initform-unsupplied+)), using the slot name rather than the initarg name. I haven't investigated where that comes from, but it's better than nothing.

2 1