October 2003 - slime-devel - mailman3.common-lisp.net

[slime-devel] slime-space fix
by Hannu Koivisto 22 Oct '03

22 Oct '03

Greetings, The attached patch makes it possible to insert several spaces with slime-space, just like you can when SPC is bound to self-insert-command. Also this problem, like the one of not checking whether an inferior Lisp is connected, seems to have been plagiarized from an old version of ILISP ;) -- Hannu

1 0

[slime-devel] more stuff
by Luke Gorrie 21 Oct '03

21 Oct '03

Hi-diddly-ho, Very nice cleanup and extension work Helmut :-) I wonder - now that we have both input and output redirected to Emacs, would writing an `ielm'-style top-level just be a matter of some elisp hacking? I've hacked the connection setup to run asynchronously, so M-x slime will try to connect in the background using a timer to retry. There's a new command `slime-disconnect' that will abort connection retrying or, if already connected, close the connection. This makes it easier to see what's going on and do manual fixes if e.g. the backend fails to compile or something. Now I'm wondering about the next steps in the backend integration/factoring. There are two things I'd like to do: Have a well-defined interface that the backend-specific code implements. Perhaps a `defcallback' macro that we use in swank.lisp where we can put a docstring and possibly some assertions about return values. We should also detect/warn for unimplemented backend functions. This seems pretty easy. The other idea is not so clear, but I would like to factor all of the protocol-related code that sends and receives messages into common code. This is to make the protocol hackable without modifying/breaking all backends, and to get protocol invariants (like proper nesting of :debug and :debug-return) taken care of in one place. I'm also now leaning towards Helmut's idea of using send/receive asynchronous message passing as the lowest layer of communication. Currently we have a hybrid where Emacs always talks to Lisp with RPC but Lisp can send state-specific asynchronous messages to Emacs. This is a bit funky in some cases, e.g. invoking a restart from Emacs would be better done as an asynchronous message than as an RPC since we don't want a result. There's currently a bug in the protocol that you can see when trying to invoke restarts. In some cases the backends will want to send (:debug-condition <string>) messages to Emacs, but the Elisp code doesn't understand those messages anymore (due to my last "improvements"). In the automaton it would be nice to handle these messages in the `debugging' state, but since invoking a restart is done by RPC we end up in the `evaluating' state waiting for a result. So it seems nicer to control the debugger with asynchronous messages instead of RPCs. So, um, that's pretty vague. Mostly written to publicly recant my previous suggestion that RPC is the way to go for everything :-) One thing I did hacking-wise is to add more significant detection of protocol errors. If the Elisp code gets a message that it can't understand in its current state, it closes the connection and pops up a buffer with some debug info and a statement to the effect of "hey, you've found a bug!". Currently I'm getting this trying to invoke the `q'uit restart in the SBCL backend, due to the Elisp :debug-condition bug mentioned above (though I'm not sure why this is hitting a debug-condition in the first place - maybe a bug in the SBCL backend). Cheers, Luke

3 3

[slime-devel] inferior-slime-mode
by Luke Gorrie 21 Oct '03

21 Oct '03

Ahoyhoy, There is now an `inferior-slime-mode'. This is a minor mode to use with inferior-lisp-mode, i.e. in the *inferior-lisp* buffer. It's like the normal slime-mode but with fewer keybindings, e.g. it doesn't have M-p and M-n because a) there're no compiler notes in that buffer and b) comint uses those for input history. I was pretty generous in adding slime-mode bindings to this new mode, though some of them are shadowed by inferior-lisp-mode (probably rightly so). Some tweaking may be in order. The README is updated - what you need is: (add-hook 'inferior-lisp-mode-hook (lambda () (inferior-slime-mode t))) Thanks to everyone who pointed out the problems of using the regular slime-mode with the inf-lisp buffer. Cheers, Luke

1 0

[slime-devel] recent symbol completion change
by Andreas Fuchs 21 Oct '03

21 Oct '03

hi all, after today's "cvs up", I read in the changelog: * swank.lisp (completions): Slight change of semantics: when a prefix-designator is package-qualified, like "swank:", only match symbols whose home-package matches the one given - ignore inherited symbols. I think this leads to unexpected and not-so-useful behaviour. In short, as Krystof said on IRC, "if the symbol is exported from a given package, it's irrelevant what its home package is". To elaborate: This will break with interface packages, i.e. ones that re-export symbols that they import from other, more implementation-dependent packages. SB-MOP and CLIM could be examples of this. For example, CLIM:OUTPUT-STREAM-P is a symbol :IMPORTed-FROM #:CLIM-LISP in McCLIM. McCLIM imports it from COMMON-LISP. It is documented in the spec (http://www.stud.uni-karlsruhe.de/~unk6/clim-spec/D-1.html#_1927) but the programmer can't complete it (even though it's there); This will lead to the programmer losing hair or to SLIME losing users. OTOH, slime does complete symbols that are in a package that :USEs another, e.g. SB-MOP. This isn't all too consistent, IMHO. Now there. Please undo this change (-: Thanks, -- Andreas Fuchs, <asf(a)acm.org>, asf(a)jabber.at, antifuchs

2 2

[slime-devel] strange failure with the sbcl backend
by Andreas Fuchs 18 Oct '03

18 Oct '03

Hi, I was wondering lately why slime/sbcl did not want to compile swank-sbcl.lisp. When I tried it, I got the backtrace available at <URL:http://shell.iddqd.org:8081/paste/display/103>. As you can see, somehow sbcl fails to find the appropriate method for slime-output-stream-buffer. I can trigger this by executing C-c : (defclass slime-output-stream (sb-gray:fundamental-character-output-stream) ((buffer :initform (make-string-output-stream :element-type 'character) :accessor slime-output-stream-buffer))) RET (i.e. a complete 1:1 redefinition of the slime-output-stream class and the slime-output-stream-buffer GF) I'm not sure which is to blame: me, sbcl or swank. It also seems like I'm the only person for whom this is failing. On my machine, I can reproduce this with a clean slime, fresh sbcl (no .sbclrc), while running emacs -q; so it must be my karma or something (-: Thanks, -- Andreas Fuchs, <asf(a)acm.org>, asf(a)jabber.at, antifuchs

2 1

[slime-devel] stuff
by Luke Gorrie 17 Oct '03

17 Oct '03

Howdyo, Feels good to be hacking again :-) Helmut, list-callers is great! The inspector too! This backend integration is groovy too. I've been hacking on the Elisp source-path support a bit more just now. It turns out backquote is really hard (see ChangeLog for details). We probably should push resolving source-paths into Lisp to use their reader for compiler notes too. Maybe `forward-source-path's days are numbered? Also hacked the batch-mode test suite a bit. There's a little 'test.sh' in CVS for running it. I drive it with different lisp/emacs versions with this script: #!/bin/bash tester="/home/luke/hacking/slime/test.sh" emacsen="emacs21 emacs20" lispen="sbcl-cvs cmucl-snapshot cmucl-cvs" echo "(Removing old output.)" rm *.dribble* *.results* &>/dev/null for emacs in $emacsen; do for lisp in $lispen; do tag=${emacs}_${lisp} echo -en "Testing $tag:\t" $tester $emacs $lisp $tag.dribble $tag.results && echo "passed." done done which then prints output like: $ ./slime-test.sh (Removing old output.) Testing emacs21_sbcl-cvs: 6 test(s) failed. Testing emacs21_cmucl-snapshot: 1 test(s) failed. Testing emacs21_cmucl-cvs: 1 test(s) failed. Testing emacs20_sbcl-cvs: 6 test(s) failed. Testing emacs20_cmucl-snapshot: 1 test(s) failed. Testing emacs20_cmucl-cvs: 1 test(s) failed. The details/results and dribble output are saved in files. One test they all failed is placing a compiler note inside a backquote. The extra failed SBCL cases are on tests where we currently assume we're looking things up in CMUCL's codebase. Cheers, Luke

2 2

[slime-devel] OpenMCL Stuff and Swank Compilation
by James Bielman 17 Oct '03

17 Oct '03

Greetings, I've about finished merging my latest OpenMCL code with Slime CVS. However, the code that compiles the sys-dependent backend in an EVAL-WHEN in swank.lisp does not work in OpenMCL. (As I understand the problem, all the DEFMACROs and DEFVARs would need to have EVAL-WHENs around them as well...) Following Dan's suggestion in IRC I started on a loader that the Emacs Lisp calls to compile and load the backend---I've put it up at http://jamesjb.com/slime/swank-loader.lisp for comments. One thing we lose if we do it this way is the ability to give the user the nice "recompile swank?" prompt---we could instead have the elisp compile each of the backend files but then we cannot automagically figure the right backend based on *FEATURES*. (I tried to test the loader under MacOS X SBCL but Swank doesn't seem to run in my build of 0.8.4., thus only tested in OpenMCL). Speaking of testing, a few trivial runs of the unit tests with the OpenMCL backend resulted in 255 failures---after glancing at the code it looks like it would need a fair bit of work to port. In particular it doesn't seem to deal very well with unimplemented slimefuns---perhaps we need a better way to say 'this function is unimplemented' besides dumping the user into the debugger? James -- #:jamesjb

2 1

[slime-devel] STYLE-WARNING/NOTE distinction
by Christophe Rhodes 17 Oct '03

17 Oct '03

Hi, SBCL pays careful attention to the distinction between a STYLE-WARNING (a condition strong enough to warrant returning from COMPILE-FILE with WARNINGSP set) and a compiler note (generally an optimization note, but sometimes a code deletion note or something similar), which aren't counted as warnings. Attached is a patch to implement this. I know cmucl doesn't have this distinction; I'm unsure as to whether OpenMCL does, so I've just implemented it for SBCL; the behaviour of other lisps should be unchanged. Cheers, Christophe -- http://www-jcsu.jesus.cam.ac.uk/~csr21/ +44 1223 510 299/+44 7729 383 757 (set-pprint-dispatch 'number (lambda (s o) (declare (special b)) (format s b))) (defvar b "~&Just another Lisp hacker~%") (pprint #36rJesusCollegeCambridge)

1 0

[slime-devel] Re: [slime-cvs] CVS update: slime/swank-cmucl.lisp
by Daniel Barlow 16 Oct '03

16 Oct '03

Helmut Eller <heller(a)common-lisp.net> writes: > +(defun read-next-form () > + (handler-case > + (let* ((length (logior (ash (read-byte *emacs-io*) 16) > + (ash (read-byte *emacs-io*) 8) > + (read-byte *emacs-io*))) > + (string (make-string length))) > + (sys:read-n-bytes *emacs-io* string 0 length) > + (read-form string)) > + (condition (c) > + (throw 'serve-request-catcher c)))) OK, so what is the problem with CMUCL that means it can't use read-sequence here? I know it used to be broken (did partial reads) but I thought that was fixed ages ago. Can we not report this as a bug to the cmucl developers? Thanks for cleaning up my other mess. Now I have to look through your structure stuff and who-calls, see what I can steal for sbcl :-) -dan -- http://web.metacircles.com/cirCLe_CD - Free Software Lisp/Linux distro

2 2