Hello.
I'm using Debian GNU/Linux (jessie) on amd64 architecture.
Usually, I do not have any problems with starting slime, but, now, whenever I issue the command "M-x slime" in Emacs, I get lots of STYLE-WARNING messages from sbcl showing redefinitions of swank functions, followed by "Polling" messages suggesting that I issue the command "M-x slime-abort-connection".
I have uploaded the part of my ~/.emacs.d/init.el file concerning the Slime connection to http://paste.lisp.org/+31M2.
Could someone look at that paste and tell me what I'm doing wrong?
Sian Mountbatten poenikatu@fastmail.co.uk writes:
Hello.
I'm using Debian GNU/Linux (jessie) on amd64 architecture.
Usually, I do not have any problems with starting slime, but, now, whenever I issue the command "M-x slime" in Emacs, I get lots of STYLE-WARNING messages from sbcl showing redefinitions of swank functions, followed by "Polling" messages suggesting that I issue the command "M-x slime-abort-connection".
I have uploaded the part of my ~/.emacs.d/init.el file concerning the Slime connection to http://paste.lisp.org/+31M2.
This is really not enough to investigate your problem since I don't know what you have"/opt/ql/slime-helper.el" and I also don't know what kind of code you have compiled in your /opt/sbcl/lib/sbcl/sbcl.slime-core file
Can you paste slime-helper.el if it's not too big of a file?
Have you tried launching slime with the official installation recipe, meaning the following code and nothing else?
;; setup load-path and autoloads (add-to-list 'load-path "~/dir/to/cloned/slime") (require 'slime-autoloads)
;; Set your lisp system and, optionally, some contribs (setq inferior-lisp-program "/opt/sbcl/bin/sbcl") (setq slime-contribs '(slime-fancy))
This should work, then try replacing the penultimate line with
(setq inferior-lisp-program "/opt/sbcl/bin/sbcl --noinform --core /opt/sbcl/lib/sbcl/sbcl.slime-core")
João
joaotavora@gmail.com (João Távora) writes:
Sian Mountbatten poenikatu@fastmail.co.uk writes:
Hello.
I'm using Debian GNU/Linux (jessie) on amd64 architecture.
Usually, I do not have any problems with starting slime, but, now, whenever I issue the command "M-x slime" in Emacs, I get lots of STYLE-WARNING messages from sbcl showing redefinitions of swank functions, followed by "Polling" messages suggesting that I issue the command "M-x slime-abort-connection".
I have uploaded the part of my ~/.emacs.d/init.el file concerning the Slime connection to http://paste.lisp.org/+31M2.
This is really not enough to investigate your problem since I don't know what you have"/opt/ql/slime-helper.el" and I also don't know what kind of code you have compiled in your /opt/sbcl/lib/sbcl/sbcl.slime-core file
Can you paste slime-helper.el if it's not too big of a file?
For reference, slime-helper.el is part of quicklisp-slime-helper. It sets up emacs to point to the version of slime that is installed by Quicklisp. It looks like this:
https://github.com/quicklisp/quicklisp-slime-helper/blob/master/slime-helper...
Zach
Zach Beane xach@xach.com writes:
For reference, slime-helper.el is part of quicklisp-slime-helper. It
I figure it would probably be something, but still wanted to rule out any external interference.
It looks like this:
https://github.com/quicklisp/quicklisp-slime-helper/blob/master/slime-helper...
Thanks, this is not the offical recipe, but is a de facto standard.
By the way in pull request
https://github.com/slime/slime/pull/83
there is a proposal for making ASDF the default swank-loading backend in SLIME, with slime-loader.lisp still being the fallback and fully supported (but doesn't need to load contribs upfront).
I consider it mostly finished and would appreciate it if any of the users/developers on this mailing list could spare a review. I'm not going to merge it without some feedback, since it's too big of a change.
I mention this in this thread, Zach, because it might make sense to use it to quicklisp's advantage, which already uses ASDF. Taking advantage of the new feature might simplify slime-helper.el.
If the goal of providing SLIME and its contribs as ELPA packages is additionally achieved, then maybe the need for slime-quicklisp-helper could be reevaluated.
What exactly is the use case that (ql:quickload :slime-quicklisp-helper) is good at solving? And how would it be easier/better/different than
M-x install-package RET slime-fancy RET
(if that were possible)?
João
Je Wed, 16 Apr 2014 18:14:21 +0100 joaotavora@gmail.com (João Távora) skribis:
Zach Beane xach@xach.com writes:
For reference, slime-helper.el is part of quicklisp-slime-helper. It
I figure it would probably be something, but still wanted to rule out any external interference.
It looks like this:
https://github.com/quicklisp/quicklisp-slime-helper/blob/master/slime-helper...
Thanks, this is not the offical recipe, but is a de facto standard.
By the way in pull request
https://github.com/slime/slime/pull/83
there is a proposal for making ASDF the default swank-loading backend in SLIME, with slime-loader.lisp still being the fallback and fully supported (but doesn't need to load contribs upfront).
I consider it mostly finished and would appreciate it if any of the users/developers on this mailing list could spare a review. I'm not going to merge it without some feedback, since it's too big of a change.
I mention this in this thread, Zach, because it might make sense to use it to quicklisp's advantage, which already uses ASDF. Taking advantage of the new feature might simplify slime-helper.el.
If the goal of providing SLIME and its contribs as ELPA packages is additionally achieved, then maybe the need for slime-quicklisp-helper could be reevaluated.
What exactly is the use case that (ql:quickload :slime-quicklisp-helper) is good at solving? And how would it be easier/better/different than
M-x install-package RET slime-fancy RET
(if that were possible)?
João
I've solved the problem by following the installation instructions in the Slime manual. No problems now.
joaotavora@gmail.com (João Távora) writes:
What exactly is the use case that (ql:quickload :slime-quicklisp-helper) is good at solving? And how would it be easier/better/different than
M-x install-package RET slime-fancy RET
(if that were possible)?
quicklisp-slime-helper was designed to improve two things.
First, installation used to be a real pain, what with using CVS, setting up load paths for both ASDF and elisp, and configuring stuff. It was intended to cut out all the fetching work and most of the configuring work. If M-x install-package can do that now, that's one problem solved.
Second, it was meant to make it easy to go back in time (http://blog.quicklisp.org/2011/08/going-back-in-dist-time.html) to work with a swank from a particular timeframe, and load the right slime to be compatible with it. I'm not sure if the emacs packaging system can help with that, but if it can, that's cool.
Zach