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

[slime-devel] cmucl source-location and slime cvs problem
by Raymond Toy 17 Apr '04

17 Apr '04

As Luke knows, I've been having some problems with the latest cmucl cvs and slime. Whenever emacs connects to the swank server, cmucl immediately gets a bus error. (It shouldn't do that, but I don't know why.) I think I finally tracked this down. In cmucl cvs, Helmut's source-location patches have been incorporated. However, swank-cmucl.lisp references some unknown functions such as file-source-location-tlf-number. There's also a bug in resolve-file-source-location: (defun resolve-file-source-location (location) (let ((filename (c::file-source-location-pathname location)) (tlf-number (c::file-source-location-tlf-number location)) (form-number (c::file-source-location-tlf-number location))) Presumably the tlf-number and form-number should not be the same; form-number should be from file-source-location-form-number. Anyway, attached is a simple hack patch that fixes these issues. At least slime is now working again with cmucl cvs. :-1 Thanks for the great tool! Ray Index: swank-cmucl.lisp =================================================================== RCS file: /project/slime/cvsroot/slime/swank-cmucl.lisp,v retrieving revision 1.94 diff -u -r1.94 swank-cmucl.lisp --- swank-cmucl.lisp 6 Apr 2004 10:42:53 -0000 1.94 +++ swank-cmucl.lisp 17 Apr 2004 01:41:54 -0000 @@ -712,10 +712,22 @@ `(:function-name ,string))) (t (list :error (princ-to-string c)))))) +(defun file-source-location-tlf-number (location) + (nth-value 0 (decode-form-numbers (c::file-source-location-form-numbers location)))) + +(defun file-source-location-form-number (location) + (nth-value 1 (decode-form-numbers (c::file-source-location-form-numbers location)))) + +(defun source-location-tlf-number (location) + (nth-value 0 (decode-form-numbers (c::file-source-location-form-numbers location)))) + +(defun source-location-form-number (location) + (nth-value 1 (decode-form-numbers (c::file-source-location-form-numbers location)))) + (defun resolve-file-source-location (location) (let ((filename (c::file-source-location-pathname location)) - (tlf-number (c::file-source-location-tlf-number location)) - (form-number (c::file-source-location-tlf-number location))) + (tlf-number (file-source-location-tlf-number location)) + (form-number (file-source-location-form-number location))) (with-open-file (s filename) (let ((pos (form-number-stream-position tlf-number form-number s))) (make-location `(:file ,(unix-truename filename)) @@ -723,8 +735,8 @@ (defun resolve-source-location (location) (let ((info (c::source-location-user-info location)) - (tlf-number (c::source-location-tlf-number location)) - (form-number (c::source-location-tlf-number location))) + (tlf-number (source-location-tlf-number location)) + (form-number (source-location-form-number location))) ;; XXX duplication in frame-source-location (assert (info-from-emacs-buffer-p info)) (destructuring-bind (&key emacs-buffer emacs-buffer-string

2 2

[slime-devel] Daily ChangeLog diff
by Luke Gorrie 17 Apr '04

17 Apr '04

Index: slime/ChangeLog diff -u slime/ChangeLog:1.336 slime/ChangeLog:1.338 --- slime/ChangeLog:1.336 Thu Apr 15 10:49:07 2004 +++ slime/ChangeLog Fri Apr 16 15:33:56 2004 @@ -1,3 +1,14 @@ +2004-04-16 Luke Gorrie <luke(a)bluetail.com> + + * doc/Makefile: Added 'install' and 'uninstall' targets for the + Info manual. It may be necessary to tweak `infodir' in the + Makefile to suit the local system before installing. (Patch from + from Richard M Kreuter.) + + * doc/slime.texi (Top): The Top node is now smaller, with details + moved into Introduction. This makes the Info front page easier to + navigate. (Patch from Richard M Kreuter.) + 2004-04-15 Ivan Boldyrev <email@secret-by-request> * slime.el (slime-handle-repl-shortcut): Call `completing-read'

1 0

[slime-devel] Not getting list mail
by David Steuber 17 Apr '04

17 Apr '04

Mail from slime-devel doesn't seem to be reaching me. I'm not sure what the problem is. I didn't even realize I had gotten replies to my last post until I saw mention of the gmane.org having an archive of the list on c.l.l. Does anyone have any ideas about what the problem might be? PS, thanks for the responses to my previous post. -- I wouldn't mind the rat race so much if it wasn't for all the damn cats.

2 2

[slime-devel] Re: Not getting list mail
by David Steuber 17 Apr '04

17 Apr '04

Luke Gorrie <luke <at> bluetail.com> writes: >David Steuber <david <at> david-steuber.com> writes: > >> Mail from slime-devel doesn't seem to be reaching me. I'm not sure >> what the problem is. I didn't even realize I had gotten replies to >> my last post until I saw mention of the gmane.org having an archive >> of the list on c.l.l. Does anyone have any ideas about what the >> problem might be? > >No. Currently you are listed as subscribed and with delivery >enabled. I can't find any reason why you shouldn't be getting the >mails. > >If you only see this message in Gmane, and don't get it in the mail, >please let me know with Cc: to admin <at> common-lisp.net. > >-Luke I'm using postini as a spam filter and checked with that. No slime-devel mail there. I even added a rule to make sure that comnon-lisp.net can send me mail. I don't know what the weirdness is. slime-devel-request was able to send me confirmation that I joined the list. According to the headers, that mail came through the postini filter. This was before I checked with postini. I only looked there when I found out about gmane and wondered why I hadn't seen any of the mail. Please let me know what other information you need. I hate computers. Thanks. -- I wouldn't mind the rat race so much if it wasn't for all the damn cats.

1 0

[slime-devel] [PATCH] slime.texi: fixups for info
by Richard M Kreuter 16 Apr '04

16 Apr '04

Hello, Here are two patches for slime.texi. The first is tiny, and just fixes node references for the new Shortcuts node. The second moves the long passage that had been the contents of the Top node down one level to a node called Introduction, and adds a 3 line overview to the Top node. This is so that info users can see the manual's menu easily from the top of the Top node. These changes don't affect the ps/pdf output. Thanks, Richard

4 7

[slime-devel] Daily ChangeLog diff
by Luke Gorrie 16 Apr '04

16 Apr '04

Index: slime/ChangeLog diff -u slime/ChangeLog:1.335 slime/ChangeLog:1.336 --- slime/ChangeLog:1.335 Wed Apr 14 15:15:13 2004 +++ slime/ChangeLog Thu Apr 15 10:49:07 2004 @@ -1,3 +1,8 @@ +2004-04-15 Ivan Boldyrev <email@secret-by-request> + + * slime.el (slime-handle-repl-shortcut): Call `completing-read' + with an alist as expected, using `slime-bogus-completion-alist'. + 2004-04-14 Luke Gorrie <luke(a)bluetail.com> * doc/slime.texi (Shortcuts): Described REPL shortcuts.

2 1

[slime-devel] repl scrolling
by Brian Downing 15 Apr '04

15 Apr '04

I'm trying to get slime-repl to scroll in a certain way, and it's not cooperating. The way I'd ideally like emacs to work is to, when my cursor gets to the bottom or top of the screen, to scroll by 5 lines or less. I'd prefer it didn't scroll line-by-line, as my machine is kind of slow, but I /NEVER/ want it to recenter the cursor as is the default - that drives me crazy. For normal editing, these settings do this for me: (set-default 'scroll-up-aggressively 0.0) (set-default 'scroll-down-aggressively 0.0) (setq scroll-step 5 scroll-conservatively 0) However, large amounts of output into slime-repl will still recenter the cursor (hiss!). To try and get around this, I put this on slime-repl-mode's hook: ;; attempt (ha!) to fix scrolling (defun scroll-line-by-line () (make-local-variable 'scroll-step) (make-local-variable 'scroll-conservatively) (setq scroll-up-aggressively 0.0 scroll-down-aggressively 0.0 scroll-step 1 scroll-conservatively 100000)) This makes it scroll a line at a time (which isn't ideal), but seemed to prevent it from ever recentering the screen (which I /REALLY/ don't want.) However, I just updated slime from CVS today, and now slime-repl seems to want to recenter the screen again anyway. If the prompt is at the last line of the screen, and I enter "'hello", it will print "HELLO" and the next prompt will still be at the end of the screen. However, if I say "(prin1 'hello)", it will recenter the screen and I'll have a bunch of useless whitespace at the bottom of the screen. Is there any way I can make it stop this? I really never want it to recenter unless I explicitly call M-x recenter. As a disclaimer, I'm convinced that emacs's model of scrolling is completely broken - none of the variables that control scrolling work the way they are documented, and they all interact in incredibly confusing and broken ways. So perhaps slime-devel is the wrong place to come with this. However, scrolling in slime-repl is what annoys me the most. I'm using GNU Emacs from CVS on Mac OS X (built for X11, not Carbon). -bcd

1 0

[slime-devel] SIGIO vs. FD-HANDLER
by Luke Gorrie 15 Apr '04

15 Apr '04

Ahoy, In CMUCL and non-threads SBCL we default to using a 'SIGIO' signal handler to read and execute requests from Emacs. This means that Lisp is interrupted to immediately serve requests as soon as Emacs sends them. It is convenient, but it has potential to cause unpredictable problems. For me this has been working perfectly, but I've been hearing reports that some people are having stability problems. If you're using CMUCL or non-threads SBCL and are finding Lisp unstable under SLIME, please try adding this to your ~/.swank.lisp: (setq swank:*communication-style* :fd-handler) That way we use the select()-loop instead of signals. It ought to be very safe, though Lisp will only process requests from Emacs when it's otherwise idle. If this solves a problem for you then please let us know. Currently I'm not sure whether SIGIO or FD-HANDLER makes the better default. NB: This stuff is covered in the manual under Lisp-Side configuration. Cheers, Luke

1 0

[slime-devel] Causing slime to enter the debugger
by Thomas F. Burdick 15 Apr '04

15 Apr '04

I'm adding some signal handlers to an application I'm writing, so it acts like a well-behaved Unix daemon. It would be nice to do something like: (defun handle-sighup (&rest ignore) (declare (ignore ignore)) #+debug (cerror "Restart server" "Signal request to restart") (shutdown) (read-configuration) (startup)) The problem is that if I call break from a signal handler, the normal SBCL toplevel enters the debugger. With the following patch, I can set *debugger-hook* to swank:swank-debugger-hook at the toplevel: --- swank.lisp.~1.166.~ Thu Apr 8 16:41:05 2004 +++ swank.lisp Wed Apr 14 18:37:54 2004 @@ -14,6 +14,7 @@ #:create-swank-server #:ed-in-emacs #:print-indentation-lossage + #:swank-debugger-hook ;; configurables #:*sldb-pprint-frames* #:*communication-style* @@ -882,7 +883,8 @@ *package*)) (*sldb-level* (1+ *sldb-level*)) (*swank-state-stack* (cons :swank-debugger-hook *swank-state-stack*)) - (*print-readably* nil)) + (*print-readably* nil) + (*emacs-connection* (or *emacs-connection* (default-connection)))) (force-user-output) (call-with-debugging-environment (lambda () (sldb-loop *sldb-level*)))))

2 1

[slime-devel] [BUG] Completions for "," commands doesn't work
by Ivan Boldyrev 15 Apr '04

15 Apr '04

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 First, when I press comma just after Slime start, it works as self-insert char, not as command. When I press <ENTER> and press comma in new REPL line, Emacs asks me a command. But autocompletion doesn't work: for example, when I press <TAB> to see a list of available completions, I see error message: PC-do-completion: Wrong type argument: listp, "force-compile-system" Here is stack trace: Debugger entered--Lisp error: (wrong-type-argument listp "force-compile-system") all-completions("" ("force-compile-system" "compile-system" "load-system" "force-load-system" "compile-and-load" "defparameter" "!" "sayoonara" "resend-form" "pop-package" "-p" "push-package" "+p" "change-package" "!p" "pop-directory" "-d" "push-directory" "+d" "change-directory" "!d" "help" "?") nil) PC-do-completion(nil) PC-complete() call-interactively(PC-complete) completing-read("Command: " ("force-compile-system" "compile-system" "load-system" "force-load-system" "compile-and-load" "defparameter" "!" "sayoonara" "resend-form" "pop-package" "-p" "push-package" "+p" "change-package" "!p" "pop-directory" "-d" "push-directory" "+d" "change-directory" "!d" "help" "?") nil t nil slime-repl-shortcut-history) (slime-lookup-shortcut (completing-read "Command: " (slime-list-all-repl-shortcuts) nil t nil (quote slime-repl-shortcut-history))) (let ((command-spec ...)) (call-interactively (cdr ...))) (if (save-excursion (goto-char slime-repl-input-start-mark) (looking-at " *$")) (let (...) (call-interactively ...)) (insert (string slime-repl-shortcut-dispatch-char))) slime-handle-repl-shortcut() call-interactively(slime-handle-repl-shortcut) As I understand, second argument of all-completions is assoc. list, not just list of strings (and documentation string for this function states the same). GNU Emacs 21.3.1 (i686-pc-linux-gnu, X toolkit, Xaw3d scroll bars) of 2003-09-20 on localhost Slime was fetched from CVS last night (HEAD revision). - -- Ivan Boldyrev Violets are red, Roses are blue. // I'm schizophrenic, And so am I. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.3.5 (GNU/Linux) iQEVAwUBQH4kBA4ALcwzZFpVAQI7aQf8Cqo5wAlUnL6Mxpx85yBSOFY/027iOaLe TsbKT9CFL3qFoyC+lyIHEGLagM+BhMy6HjY1Vx7fwZvIT82etZyq/PWW1ji+Mv/+ 7F0Jqd1ILL17Nfso+gsgG/ayslwGNIKwGO1H2jgEZ72E5H9pbVmwhrSbHPTFDIFg wAfA2FzCrTwvA5Q5wtFuuOOrCYl54e9ZxpbA1Cn1x3Z2ojeE4yd6SFFc5Ah3lea2 WGfyohcxEdT5UdtgEntfp4t0oDBid0YcsWwcv30icclgC0ODq7p9ol3WrL6HPN9n SNFPqXGzRoLGKYQkhmu46Yg6zG9gkWAqW9Dmlq2ssnTpgGydIjHbZA== =m5E+ -----END PGP SIGNATURE-----

2 1