January 2010 - slime-devel - mailman3.common-lisp.net

[slime-devel] Standard output no longer appears in REPL
by Steven E. Harris 25 Jun '11

25 Jun '11

It's been a couple of months since I last used SLIME. Today I noticed that when I evaluate expressions in the REPL that write to *standard-output*, the output appears in the *inferior-lisp* buffer, but not in the REPL as it used to. I'm using GNU Emacs on Windows atop the CLISP Cygwin build. SLIME is up to date with CVS. Is there some new configuration required to restore the old behavior with output streams feeding into the REPL output? -- Steven E. Harris

5 12

[slime-devel] broken clojure REPL
by Cyrus Harmon 21 May '10

21 May '10

FWIW, the following patch broke the clojure REPL (*inferior-lisp* still works, but not the fancy REPL): commit addc5be70a489cdc8ed555a2fed7273040da196e Author: Tobias C. Rittweiler <trittweiler(a)common-lisp.net> Date: Sat Oct 31 22:13:55 2009 +0000 * slime.el (slime-inside-string-p, slime-inside-comment-p) (slime-inside-string-or-comment-p): New. * swank-match.lisp: New file. Contains very simple pattern matcher from the CMU AI archive. * swank-loader.lisp: Compile swank-match.lisp. * swank.lisp: Make SWANK package use new SWANK-MATCH package. * slime-autodoc.el, swank-arglists.lisp: Large parts were rewritten. Autodoc is now able to highlight &key parameters, and parameters in nested arglists. * slime-parse.el, slime-c-p-c.el, slime-highlighting-edits.el: Adapted to changes. the symptom is the following: ; SLIME 2009-10-31 user> 3 3 user> (+ 3 4) at which point it just hangs... nothing (helpful) in *messages* or *inferior-lisp*. Reverting to the previous commit fixes the problem. It's really nice to be able to use git to quickly bisect the history to find the offending commit. Not so nice when that commit is >2k lines :( TCR, any ideas? thanks, cyrus

9 33

[slime-devel] Error in swank-fancy-inspector when inspecting network streams [with patch]
by Russ Tyndall 14 May '10

14 May '10

When attempting to inspect a network-stream of underlying type SB-SYS:FD-STREAM I always get the following error: Error: The value NIL is not of type PATHNAME. I wrote a small patch to swank-fancy-inspector to ignore errors coming from accessing the pathname slot and not render the action link (visit file and show current position) when we do not have a pathname. Cheers, Russ Tyndall Acceleration.net Software Developer commit a26d435542936b2065b9d30e94df7bbed156a0e2 Author: Russ Tyndall <russ(a)acceleration.net> Date: Mon Jan 11 12:04:16 2010 -0500 Made swank-fancy-inspector NOT error when attempting to inspect network streams (due to their pathname being unbound) diff --git a/contrib/swank-fancy-inspector.lisp b/contrib/swank-fancy-inspector.lisp index f68c6b0..f30e0c0 100644 --- a/contrib/swank-fancy-inspector.lisp +++ b/contrib/swank-fancy-inspector.lisp @@ -805,27 +805,29 @@ SPECIAL-OPERATOR groups." (defmethod emacs-inspect ((stream file-stream)) (multiple-value-bind (content) (call-next-method) - (append - `("Pathname: " - (:value ,(pathname stream)) - (:newline) " " - ,@(when (open-stream-p stream) - `((:action "[visit file and show current position]" - ,(make-visit-file-thunk stream) - :refreshp nil) - (:newline)))) - content))) + (let ((pathname (ignore-errors (pathname stream)))) + (append + `("Pathname: " + (:value ,pathname) + (:newline) " " + ,@(when (and (open-stream-p stream) pathname) + `((:action "[visit file and show current position]" + ,(make-visit-file-thunk stream) + :refreshp nil) + (:newline)))) + content)))) (defmethod emacs-inspect ((condition stream-error)) (multiple-value-bind (content) (call-next-method) - (let ((stream (stream-error-stream condition))) + (let* ((stream (stream-error-stream condition)) + (pathname (ignore-errors (pathname stream)))) (if (typep stream 'file-stream) (append `("Pathname: " - (:value ,(pathname stream)) + (:value ,pathname) (:newline) " " - ,@(when (open-stream-p stream) + ,@(when (and (open-stream-p stream) pathname) `((:action "[visit file and show current position]" ,(make-visit-file-thunk stream) :refreshp nil)

2 5

[slime-devel] Daily ChangeLog diff
by Helmut Eller 01 Feb '10

01 Feb '10

Index: slime/ChangeLog diff -u slime/ChangeLog:1.1973 slime/ChangeLog:1.1975 --- slime/ChangeLog:1.1973 Thu Jan 28 04:52:18 2010 +++ slime/ChangeLog Sat Jan 30 10:59:52 2010 @@ -1,8 +1,19 @@ +2010-01-30 Stas Boukarev <stassats(a)gmail.com> + + * slime.el (slime-cycle-connections): Change docstring, it + doesn't make connection buffer-local. + * doc/slime.texi (Multiple connections): document + slime-cycle-connections. + + * swank.lisp (pathname-as-directory): Treat "/foo/bar" and "/foo/bar/" + the same. + (compile-file-output): Use the above function. + 2010-01-28 Mark Evenson <evenson(a)dada.local> * swank-abcl.lisp (emacs-inspect): Make inspection of Java objects toString() results dependent on explicit user request to avoid - potentially computationally expensive opreations. + potentially computationally expensive operations. 2010-01-27 Stas Boukarev <stassats(a)gmail.com>

3 2

[slime-devel] swank breakage in Allegro CL
by mikel evins 31 Jan '10

31 Jan '10

CVS head currently has a single upper-case NIL in swank-rpc.lisp that prevents it from working with Allegro CL's mlisp. Downcasing that single token fixes the breakage. It's at line 103 of last night's swank-rc.lisp: (defparameter *validate-input* NIL "Set to true to require input that strictly conforms to the protocol") Changing this form thusly makes mlisp happy again: (defparameter *validate-input* nil "Set to true to require input that strictly conforms to the protocol")

2 1

[slime-devel] slime-repl-find-buffer-package in unconnected lisp
by Madhu 31 Jan '10

31 Jan '10

This commit was a bad idea and has unforseen, undesirable consequences: * 2009-10-15 Helmut Eller <heller(a)common-lisp.net> | | * slime-repl.el (slime-repl-find-buffer-package): New function. | (slime-repl-init): Initialize slime-find-buffer-package-function. To use slime-repl-find-buffer-package this way, either slime-find-buffer-package-function should be made buffer local for "live" repl buffers, or much code needs to be redesigned to avoid redisplay errors, (when editing IN-PACKAGE forms, say), in a lisp buffer when slime is not connected to a lisp. The attached backtrace may help you understand what I'm talking about. -- Madhu backtrace() (or (slime-auto-connect) (backtrace) (error "Not connected.")) (cond ((and ... slime-net-processes) (or ... ...)) ((not conn) (or ... ... ...)) ((not ...) (error "Connection closed.")) (t conn)) (let ((conn ...)) (cond (... ...) (... ...) (... ...) (t conn))) slime-connection() (or process (slime-connection) (error "No connection")) (process-buffer (or process (slime-connection) (error "No connection"))) (set-buffer (process-buffer (or process ... ...))) (save-current-buffer (set-buffer (process-buffer ...)) slime-lisp-package:connlocal) (with-current-buffer (process-buffer (or process ... ...)) slime-lisp-package:connlocal) (slime-with-connection-buffer (process) slime-lisp-package:connlocal) slime-lisp-package() (or (slime-search-buffer-package) (slime-lisp-package)) slime-repl-find-buffer-package() funcall(slime-repl-find-buffer-package) slime-find-buffer-package() (save-restriction (widen) (slime-find-buffer-package)) (or slime-buffer-package (save-restriction (widen) (slime-find-buffer-package))) slime-current-package() (let* ((conn ...) (local ...) (pkg ...)) (cond (... ...) (...))) slime-modeline-string() eval((slime-modeline-string))

1 0

[slime-devel] Making SLIME ChangeLogs more useful
by Mark Harig 30 Jan '10

30 Jan '10

In the interest of making the SLIME ChangeLogs more useful, I have written the following notes concerning Change Log mode. 1. The major mode Change Log can be used with ChangeLog files to make them behave as hypertext that can then be used to review quickly the changes that are described. To do this, type C-c C-c when point is before or on a properly-formatted ChangeLog entry, and the changed file will displayed in another window, optionally, at the definition of a specified symbol. As the documentation for C-c C-c (change-log-goto-source) says: "Go to source location of "change log tag" near `point'. A change log tag is a symbol within a parenthesized, comma-separated list. If no suitable tag can be found nearby, try to visit the file for the change under `point' instead." A properly-formatted ChangeLog entry has the form: [white-space]* [file name] (symbol name,...): [descriptive text] [white-space](symbol name): [descriptive text] The trailing ':' is customary, but C-c C-c will attempt to find the definition of the parenthesized symbol(s) listed without it, or even if the symbol is listed after it. Examples: * slime.el (slime-edit-uses-xrefs): New variable. For contribs to * swank-cmucl.lisp (reset-sigio-handlers): New function. * swank-backend.lisp (socket-fd, make-fd-stream, dup, exec-image) * slime.el (slime-selector-other-window): New variable. (slime-selector): Bind it as need. (def-slime-selector-method): Use the other window if so indicated. Pressing C-c C-c while the cursor is anywhere on the first two entries will cause Emacs to locate those symbols' definitions (only) in those files. In the comma-delimited list of the third entry, C-c C-c will locate the definition of the symbol closest to point. In the multi-line fourth entry, C-c C-c will locate in slime.el the definition of the symbol closest to point. If the file or symbol name is misspelled or the parenthesized text is not (strictly) a symbol name, then Emacs will indicate this with a message rather than locating the symbol's definition. So, for example, * foo.lisp: Fix a typo (grammar error) will cause Emacs to look in the file foo.lisp for the symbol 'grammar' because it is parenthesized text following a file name. Authors of ChangeLog entries should be able to quickly check the validity of their entries by typing C-c C-c before committing them to CVS. 2. It is also possible to "walk through" a sequence of ChangeLog entries just as you do with compiler errors by typing C-x ` (backquote), which is mapped to `next-error'. This will: - open the file in another window at the definition specified by the ChangeLog entry - move the cursor to that definition - move point in the ChangeLog file to the next entry for when you return to the ChangeLog (Be aware that once the cursor is in the source file, `next-error' does not have the same meaning as it does in the ChangeLog file. After reviewing the change, you will need to move the cursor back to the ChangeLog file before typing C-x ` to examine the file/symbol-definition of the next ChangeLog entry.) 3. This capability of Change Log mode exists for Lisp code. It does not work with, for example, texinfo files, that is, it does not understand texinfo symbol definitions. The following entry will open "doc/slime.texi", but will not move point to the definition of the symbol: * doc/slime.texi (swank-loader:init): is also needed for loading swank. Similarly, non-symbol-name text inside parentheses is meaningless to C-c C-c. So, for example, this entry: * slime.el (compile-defun [test]): Add two cases. cannot be found using C-c C-c. But the following entry can locate the definition of `compile-defun', which can then be reviewed for the two test cases. * slime.el (compile-defun): Add two test cases. Also, non-definitions of symbols (that is, when you call a function or reference a variable) cannot be found by C-c C-c. So, the following entry: * swank.lisp: (decode-message): Adopted will cause Emacs to do nothing except display an informative message when C-c C-c is pressed because, while `decode-message' is *used* in swank.lisp, it is not defined there. "Source location of tag `decode-message' not found in file `swank.lisp'" An alternative entry that does not use the (symbol-name) syntax would have the benefit of opening swank.lisp, allowing the user to then search for references of `decode-message', for example: * swank.lisp: Adopted the function `decode-message' where appropriate. 4. Finally, on-the-fly spelling checking can be added to Change Log mode with, for example, the following: (add-hook 'change-log-mode-hook 'turn-on-flyspell) or, indirectly, with (add-hook 'text-mode-hook 'turn-on-flyspell) This could be of help to non-native English readers who may have a harder time guessing what a misspelled word should be. A misspelled word can be quickly repaired by cycling through possible corrections using repeated M-Tab key presses. Thank you again to the authors of SLIME.

3 4

[slime-devel] Suggested minor ecl-swank diff
by Matthew Mondor 30 Jan '10

30 Jan '10

Hello, I noticed that ecl-swank depends on ECL sb-bsd-sockets:socket-make-stream creating unbuffered streams by default (SBCL creates them fully buffered by default). Although this was not yet accepted upstream, in my local ECL tree, socket streams are created buffered by default like for SBCL (since ECL socket code was designed to be SBCL compatible, this was an annoyance for me). So I suggest the following change in case upstream ECL default behavior also eventually changes to be more SBCL-compatible. This would also prevent me from having to merge my minor local change when I update slime... In buffered mode, with the current state of ECL swank code, emacs hangs when loading slime. This simple patch fixes this. It'd ideally be best to fix the ECL swank code to use a buffered mode for performance (and for cleaner syscall trace reports, which are now encumbered by single-character I/O), but that would be a more complex change. Thanks, -- Matt

1 0

[slime-devel] Daily ChangeLog diff
by Helmut Eller 29 Jan '10

29 Jan '10

Index: slime/ChangeLog diff -u slime/ChangeLog:1.1972 slime/ChangeLog:1.1973 --- slime/ChangeLog:1.1972 Wed Jan 27 01:38:27 2010 +++ slime/ChangeLog Thu Jan 28 04:52:18 2010 @@ -1,3 +1,9 @@ +2010-01-28 Mark Evenson <evenson(a)dada.local> + + * swank-abcl.lisp (emacs-inspect): Make inspection of Java objects + toString() results dependent on explicit user request to avoid + potentially computationally expensive opreations. + 2010-01-27 Stas Boukarev <stassats(a)gmail.com> * swank-rpc.lisp: NIL -> nil (for unusual readtable settings). @@ -5,7 +11,7 @@ 2010-01-26 Mark Evenson <evenson(a)dada.local> - * swank-abcl.lisp (import-to-swank-mop): Import + * swank-abcl.lisp: Import MOP::COMPUTE-APPLICABLE-METHODS-USING-CLASSES if it exists in the ABCL implementation.

1 0

[slime-devel] Strange behavior with CVS access
by Mark Harig 27 Jan '10

27 Jan '10

SLIME maintainers, I am not able to update my copy of the source files for SLIME from the CVS repository since the 2010-01-19 changes. Has there been some change to the location of the CVS repository? $ mkdir new $ cd new $ cvs -d :pserver:anonymous:anonymous@common-lisp.net:/project/slime/cvsroot co slime slime cvs checkout: Updating slime (long list of files omitted) $ cd slime $ cvs status ChangeLog =================================================================== File: ChangeLog Status: Up-to-date Working revision: 1.1966 Repository revision: 1.1966 /project/slime/cvsroot/slime/ChangeLog,v Commit Identifier: F7aol0uPOJGkS0ku Sticky Tag: (none) Sticky Date: (none) Sticky Options: (none) So, CVS is reporting that revision 1.1966 is the latest revision to ChangeLog. CVS 'log' also reports the head revision as 1.1966, but many changes -- up to revision 1.1970 -- have been reported: Daily ChangeLog Diff: On 1/26/2010 3:33 AM, Helmut Eller wrote: Index: slime/ChangeLog diff -u slime/ChangeLog:1.1969 slime/ChangeLog:1.1970 --- slime/ChangeLog:1.1969 Thu Jan 21 19:02:08 2010 +++ slime/ChangeLog Mon Jan 25 05:50:10 2010 @@ -1,3 +1,9 @@ +2010-01-25 Tobias C. Rittweiler <tcr(a)freebits.de> + + * swank-allegro.lisp (redefinition-p, redefinition): New. + (handle-compiler-warning): Add :severity for redefinitions, + style-warnings, errors. +Is anyone else seeing this behavior? I have checked out the sources from two different computers running two different operating systems. I am using CVS version 1.11.23 and 1.12.13. The result is the same on both.

1 1