Install GSLL on SBCL with Quicklisp - common-lisp

I am trying to install the GNU Scientific Library for Lisp (GSLL).
I saw there are answers to a similar question already, but I didn't find the connection to what is happening in my case.
Could anyone help to move on from here?
CL-USER> (lisp-implementation-type)
"SBCL"
CL-USER> (lisp-implementation-version)
"1.2.4.debian"
CL-USER> (ql:quickload "gsll")
; cc -m64 -I/lrde/home/alandi/quicklisp/dists/quicklisp/software/cffi_0.14.0/ -o /lrde/home/alandi/.cache/common-lisp/sbcl-1.2.4.debian-linux-x64/lrde/home/alandi/quicklisp/dists/quicklisp/software/cffi_0.14.0/libffi/libffi-unix /lrde/home/alandi/.cache/common-lisp/sbcl-1.2.4.debian-linux-x64/lrde/home/alandi/quicklisp/dists/quicklisp/software/cffi_0.14.0/libffi/libffi-unix.c
;
; compilation unit aborted
; caught 2 fatal ERROR conditions
To load "gsll":
Load 8 ASDF systems:
alexandria antik asdf-system-connections cffi-grovel
cffi-libffi metabang-bind osicat trivial-garbage
Install 1 Quicklisp gsll
release:
; Loading "gsll"
; cc -m64 -I/lrde/home/alandi/quicklisp/dists/quicklisp/software/cffi_0.14.0/ -o /lrde/home/alandi/.cache/common-lisp/sbcl-1.2.4.debian-linux-x64/lrde/home/alandi/quicklisp/dists/quicklisp/software/cffi_0.14.0/libffi/libffi-unix /lrde/home/alandi/.cache/common-lisp/sbcl-1.2.4.debian-linux-x64/lrde/home/alandi/quicklisp/dists/quicklisp/software/cffi_0.14.0/libffi/libffi-unix.c
;
; compilation unit aborted
; caught 2 fatal ERROR conditions
; Evaluation aborted on Component "gsll" not found.

I am on OS X, but maybe this helps you anyway. I could load the lib by (ql:quickload "gsll") but I had to make an adjustment:
The path to GSLs dynamic libraries is found by a shell call gsl-config --prefix.
For me this call returns /usr/local/Cellar/gsl/1.16 but my dynlibs are located in /usr/local/Cellar/gsl/1.16/lib. So I just had to make some adjustments to the pathnames in the file ~/quicklisp/dists/quicklisp/software/gsll-20140211-git/init/init.lisp.
Seach for (cffi:define-foreign-library libgslcblas ...) and (cffi:define-foreign-library libgsl ...) and adjust your paths there.
Maybe that helps.

Related

Don't know how to require sb-cltl2

I'm trying to run an executable with an Hunchentoot server and I'm getting (after an unusual high CPU usage):
<INFO> [17:55:14] weblocks/server server.lisp (start) -
Starting weblocks WEBLOCKS/SERVER::PORT: 40001
WEBLOCKS/SERVER::SERVER-TYPE: :HUNCHENTOOT DEBUG: T
debugger invoked on a SB-INT:EXTENSION-FAILURE in thread
#<THREAD "main thread" RUNNING {1008C1EA13}>:
Don't know how to REQUIRE sb-cltl2.
Do you have any idea what's going on ? It works correctly on Slime where I use the start function.
(sbcl manual: http://www.sbcl.org/manual/#Customization-Hooks-for-Users)
In the main entry point, I try to capture the running thread so that I keep the server running on the foreground (my notes). This pattern worked with another clack-based web framework.
(defun start ()
(weblocks/debug:on)
(weblocks/server:start :port *port*))
(defun main ()
(defvar *port* (find-port:find-port))
(start)
(handler-case (bt:join-thread (find-if (lambda (th)
(search "hunchentoot" (bt:thread-name th)))
(bt:all-threads)))
(#+sbcl sb-sys:interactive-interrupt
#+ccl ccl:interrupt-signal-condition
#+clisp system::simple-interrupt-condition
#+ecl ext:interactive-interrupt
#+allegro excl:interrupt-signal
() (progn
(format *error-output* "Aborting.~&")
(uiop:quit 1))
;; for others, unhandled errors (we might want to do the same).
(error (c) (format t "Woops, an unknown error occured:~&~a~&" c)))))
Or any indication of what could be the cause ?
Thanks again.
(I'm using 40ants' weblocks:reblocks branch)
SBCL Debian 1.2.4
edit I tried
export SBCL_HOME=/usr/local/lib/sbcl/
I build with
build:
$(LISP) --quit \
--eval '(ql:quickload "foo")' \
--eval '(require :sb-cltl2)' \
--eval '(asdf:make :foo)'
=>
fatal error encountered in SBCL pid 25248(tid 140737353910016):
can't load .core for different runtime, sorry
Welcome to LDB, a low-level debugger for the Lisp runtime environment.
ldb>
The following environment failed:
export SBCL_HOME=/usr/local/lib/sbcl/
The error message tells us the following:
can't load .core for different runtime, sorry
Apparently the SBCL you ran used the given SBCL_HOME to find its core file, but failed due to the core being generated by different version SBCL.
A quick look at SBCL's source (starting from require) shows that the underlying function #'SB-INT:SBCL-HOMEDIR-PATHNAME is called to determine the installation path.
It looks like the one installed from the Debian package was installed in /usr/lib/sbcl/. The core file was easily found when starting Slime. You also had another version of SBCL in ~/.roswell/, but I guess you also ran ros install which installed it under /usr/local/lib/sbcl (/usr/local/ is for software that is not managed by the system).
Starting the roswell one when setting SBCL_HOME to the directory of the Debian one provoked the error about the incompatible core file (I guess).
What remains suprising is that (SB-INT:SBCL-HOMEDIR-PATHNAME) returns nil when starting your main function.

Building an executable with ECL: missing dependency or can not find make-build in asdf/interface

I'm trying to build an executable with ECL. I looked at the doc and this other SO question, where we learn that with ECL v <= 16.1.3 we must add a
(require 'adsf)
Then I push my project to the asdf registry:
(pushnew "~/projects/my-project/" asdf:*central-registry* :test #'equal)
I load it:
(load "my-project.asd")
My program obviously has Quicklisp dependencies.
I copied the initialization of Quicklisp from my .sbclrc to ~/.eclrc:
#-quicklisp
(let ((quicklisp-init (merge-pathnames "quicklisp/setup.lisp"
(user-homedir-pathname))))
(when (probe-file quicklisp-init)
(load quicklisp-init)))
So than I can Quickload my project: (this step does not appear in the doc but this allows to find the project dependencies)
(ql:quickload "my-project")
I also
(use-package :my-project)
but now, as I want to build the executable with
(asdf:make-build :my-project
:type :program
:move-here #P"./"
:epilogue-code '(progn (my-project:main)
(si:exit)))
I get the error
Cannot find the external symbol MAKE-BUILD in #<"ASDF/INTERFACE" package>.
Can someone help ? Thanks !
ECL 16.1.3
ps: and it seems that, in the REPL, to enter a restart nb has no effect.
My makefile:
ecl-build:
~/.roswell/impls/x86-64/linux/ecl/16.1.3/bin/ecl \
--eval "(require 'asdf)" \
--eval '(format t "--- ASDF version: ~a~&" (asdf:asdf-version))' \
--eval "(pushnew \"~/projets/cl-torrents/\" asdf:*central-registry* :test 'equal)" \
--eval '(load "cl-torrents.asd")' \
--eval '(asdf:make-build :cl-torrents :type :program :move-here #P"./" :epilogue-code "(progn (torrents "matrix") (si:exit))")'
Result:
;;; Loading "/home/vince/quicklisp/setup.lisp"
;;; Loading #P"/home/vince/.roswell/impls/x86-64/linux/ecl/16.1.3/lib/ecl-16.1.3/asdf.fas"
--- ASDF version: 3.2.1
An error occurred during initialization:
Cannot find the external symbol MAKE-BUILD in #<"ASDF/INTERFACE" package>..
Makefile:22: recipe for target 'ecl-build' failed
make: *** [ecl-build] Error 1
edit 27, oct - progress
The simplest path now seems to unsure to use ASDF bundled into ECL, which will have the make-build command.
I had a newer ASDF version because of my Quicklisp's initialization code into ~/.eclrc. I removed it and now I have ASDF v3.1.8.2, which is ok. But now, ECL doesn't know any dependencies:
Component :DEXADOR not found, required by #.
Indeed, because there is no Quicklisp any more.
ECL doc says to include all libraries path into asdf:*central-registry*. By chance, Quicklisp installed all libraries into ~/quicklisp/dists/quicklisp/software/.
First, I wonder how this will work in a CI pipeline where Quicklisp has not run before…
Then I see that adding only this directory is not enough, Dexador's dependencies will in turn not be found, I had to add dexador's directory precisely:
--eval "(pushnew \"~/quicklisp/dists/quicklisp/software/dexador-20170830-git/\" asdf:*central-registry* :test 'equal)" \
So do I really have to write code to include every directory here ?
How to make this work when Quicklisp has not run once before ?
update: With Ecl installed with Roswell: requireing asdf before everything in my .eclrc gives the version 3.1.8.2, after Quicklisp initialization 3.2.1 and make-build symbol unknown.
With Debian's Ecl: first v2.33.10 and then v3.2.1 likewise.
Looks like a dead end.
update november: I waited for the release of ASDF 3.3.0 to manually update. It was buggy. Message on the mailing list: waiting for 3.3.1: buggy again.
Went with the lisp-devel Docker image, shipping ECL 16.1.3. (I don't want to compile ECL myself on each and every VPS). I could build an executable (52Mo in weight), but got a runtime error:
Condition of type: SIMPLE-ERROR
Package ((UIOP/STREAM . #)) referenced in > compiled file
NIL
but has not been created
It seems a problem with comtibility with ASDF and ECL this is solved in last commit take a look here,
Add back make-build on ECL
This provides for backward compatibility with ECL, whose current
maintainer Daniel K. has decided to keep supporting the make-build
interface and has forked ASDF for that.
you can install/use the last ASDF from this repo
You may use either bundled ASDF (which is a frozen 3.1.7 version with some fixes backported) – then you load ASDF with (require asdf), or you may use upstream 3.3 version.
If you are interested in using upstream ASDF, download asdf.lisp file and call:
(load (compile-file "/path/to/asdf.lisp")) instead of (require 'asdf).
I Went with the lisp-devel Docker image, shipping ECL 16.1.3. (I don't want to compile ECL myself on each and every VPS). I could build an executable (52Mo in weight VS 78Mo with SBCL), so I was able to deliver it with Gitlab CI.
For reference, to launch Docker and mount your sources:
service docker start
docker run --rm -it -v /home/you/projets/project:/usr/local/share/common-lisp/source daewok/lisp-devel:latest bash
Unfortunately I got a runtime error:
Condition of type: SIMPLE-ERROR
Package ((UIOP/STREAM . #)) referenced in > compiled file
NIL
but has not been created
It also seems I had to ql:quickload :closer-mop manually before loading my app.
I won't investigate this.

Cannot install lparallel alongside CL21 (?) Failing with CL21.CORE.FUNCTION

I have a weird mix of errors.
I was using CL21, I was in my package, and I wanted to install lparallel. Not possible:
(ql:quickload :lparallel)
To load "lparallel":
Load 1 ASDF system:
lparallel
; Loading "lparallel"
;
; caught ERROR:
; DYNAMIC-EXTENT on a weird thing: (CL21.CORE.FUNCTION:FUNCTION #:BODY-FN1)
;
; compilation unit aborted
; caught 2 fatal ERROR conditions
; caught 1 ERROR condition
; Evaluation aborted on #<UIOP/LISP-BUILD:COMPILE-FILE-ERROR {1008956C13}>.
I can reproduce it in a new session but it's a bit weird: if I quickload lparallel in cl-user, it complains on not finding the symbol CL21.CORE.FUNCTION, even if I didn't do nothing with CL21 yet:
The name "CL21.CORE.FUNCTION" does not designate any package.
So I ql:quickload CL21 and then on retrying to load lparallel. I get the first error.
But, I tried in Portacle for a fresh image and… I couldn't reproduce this.
Any idea ? Is that an issue with cl21, lparallel, quicklisp or asdf??
cl21 seems to replace the standard reader macro #' with its own version that isn't compatible with lparallel. When you tried to load lparallel in a fresh image, ASDF will load it from the .fasl-files that were compiled with cl21 loaded, so you must either delete those files or force recompilation with
(asdf:operate 'asdf:load-op :lparallel :force t)
Loading cl21 after lparallel is compiled with the standard language shouldn't cause the same problem.

Quicklisp: Manually add library to dist?

Some days ago I updated SBCL to 1.2.1 and I thought it couldn't hurt to update the quick lisp dist/client as well.
WRONG!
After firing up slime in Emacs, I got this error:
; caught ERROR:
; READ error during COMPILE-FILE:
;
; Symbol "CODE-TRACE-TABLE-OFFSET-SLOT" not found in the SB-VM package.
;
; Line: 1507, Column: 70, File-Position: 60197
;
; Stream: #<SB-SYS:FD-STREAM
; for "file /Users/fyi/quicklisp/dists/quicklisp/software/slime-2.7/swank-sbcl.lisp"
; {1008B07E43}>
;
; compilation unit aborted
; caught 1 fatal ERROR condition
; caught 1 ERROR condition
; printed 1 note
;;
;; Error while compiling /Users/fyi/quicklisp/dists/quicklisp/software/slime-2.7/swank-sbcl.lisp:
;; COMPILE-FILE returned NIL.
;; Aborting.
If you're using the quicklisp-slime-helper, so as I, the path to slime comes form the ~/quicklisp/dists/quicklisp/installed/systems/swank.txt file. Slime 2.8 should fix that error, but there is no quicklisp distribution that contains slime 2.8.
Coming down to my actual question. Is there a way to specify a library location outside of a quicklisp distribution? Or will I have to wait unit the next distribution has been made available and is there a guaranty a library will be updated in the next distribution?
I am sure I'm missing something, but the whole distribution stuff seems to me more of a hindrance than a help.
The next Quicklisp dist update is coming July 12 or 13. It will include a new SLIME that is compatible with SBCL 1.2.1. In the meantime, the easiest fix is to install SBCL 1.2.0.
Another option is to comment out the quicklisp-slime-helper lines in ~/.emacs and install slime according to its own procedure.

Unable to load sdl-gfx in quicklisp

I've been installing lispbuilder-sdl family with quicklisp and encountered error in sdl-gfx:
CL-USER> (ql:quickload "lispbuilder-sdl-gfx")
To load "lispbuilder-sdl-gfx":
Load 1 ASDF system:
lispbuilder-sdl-gfx
; Loading "lispbuilder-sdl-gfx"
...........;
; compilation unit aborted
; caught 1 fatal ERROR condition
Unable to load any of the alternatives:
("libSDL_gfx.dylib" (:FRAMEWORK "SDL_gfx"))
[Condition of type CFFI:LOAD-FOREIGN-LIBRARY-ERROR]
I use slime+emacs+SBCL under macosx 10.6. I've installed SDL.framework from this link:
http://thirdcog.eu/apps/frameworks#glew
a file called sdl_with_friends.zip and put things under /Library/Frameworks. However it still complains about cannot find framework.
Any idea on this?
The easiest way to install lispbuilder-sdl on Mac OSX is to fire up SBCL and use Quicklisp:
(ql:quickload "lispbuilder-sdl")
It will probably fail, but you can then compile the OS-X specific helper library located in "~quicklisp/dists/quicklisp/software/lispbuilder-20110619-svn/lispbuilder-sdl/cocoahelper"; just cd to this directory and type "make"
To verify that worked, try this:
(ql:quickload "lispbuilder-sdl-examples")
(lispbuilder-sdl-examples:bezier)
Another common gotcha is when you are using Emacs / SLIME. The cocoa bits must run on the primary thread, so you have to invoke things like this:
#+darwin #+sb-thread
(let ((thread (first (last (sb-thread:list-all-threads)))))
(sb-thread:interrupt-thread thread #'(lambda () (ql:quickload "lispbuilder-sdl-examples")))
(sb-thread:interrupt-thread thread #'(lambda () (lispbuilder-sdl-examples:bezier))))
The error is complaining that it can't load the dynamic library for SDL. I'm sorry - I don't know much about the directory layout on macs (which appears to be where you're working), so I can't exactly tell you how to fix this. But somewhere you should have installed a file called libSDL_gfx.dylib (probably this extension) and the error message means that CFFI is failing to find it.
using homebrew you can brew install sdl_gfx

Resources