cffi-devel September 2005

cffi-devel@common-lisp.net

9 participants
35 discussions

[cffi-devel] running examples - errors?
by Surendra Singhi 15 Dec '05

15 Dec '05

Hello, I am having trouble running the cffi examples on clisp/allegrocl on windows xp. (cffi-luis-050925-2059) First of all do I need some dll or library to run the examples? The tar ball doesn't contain anything. When I load the file through slime I get the error FFI::FOREIGN-LIBRARY-FUNCTION: no dynamic object named "sqrtf" in library :DEFAULT but loading on the command line works fine. But if from the command line I execute, [3]> (cffi-examples:sqrtf 5) *** - FFI::FOREIGN-LIBRARY-FUNCTION: no dynamic object named "sqrtf" in library :DEFAULT The following restarts are available: USE-VALUE :R1 You may input a value to be used instead. ABORT :R2 ABORT Break 1 [4]> :R2 [6]> (cffi-examples:gettimeofday) *** - FFI::FOREIGN-LIBRARY-FUNCTION: no dynamic object named "gettimeofday" in library :DEFAULT The following restarts are available: USE-VALUE :R1 You may input a value to be used instead. ABORT :R2 ABORT Break 1 [7]> Only gethostname function seems to work. Thanks. -- Surendra Singhi http://www.public.asu.edu/~sksinghi/index.html "All animals are equal, but some animals are more equal than others." - Orwell, Animal Farm, 1945

3 6

Re: [cffi-devel] CFFI and SBCL
by Luis Oliveira 27 Sep '05

27 Sep '05

On 27/set/2005, at 14:04, Immanuel Litzroth wrote: >> Hmm, I would assume SBCL/CMUCL translate sb-alien:int and c-call:int >> to a fixnum already but I can't say that I checked. Is my assumption >> incorrect? > > Not if the int is 32 bits. You get a efficiency not about this when > compiling > with speed set to 3. If the int is only 30 bit the note goes away and > I have > some code where this was important (ffi to portmidi). Ah, right... that makes sense. Anyway this is a pretty specific optimization, right? I mean, this makes sense only for int and (long) long. Well, and when one is using long's it's very likely that one will be handling values bigger than most-positive-fixnum anyway. Did I miss any other cases? > I think the two different design decisions are best exemplified in > cmucl > (lists of objects with finalizers, but when the finalizer is run your > object > has already been garbage collected) and lispworks where the object is > passed > to the finalizer. Cool, thanks. I'll check those out. -- Luís Oliveira http://student.dei.uc.pt/~lmoliv/ Equipa Portuguesa do Translation Project http://www.iro.umontreal.ca/translation/registry.cgi?team=pt

2 1

[cffi-devel] CFFI and SBCL
by Immanuel Litzroth 27 Sep '05

27 Sep '05

Dear List, I tried out CFFI on SBCL linux yesterday and would like to give some feedback. 1) On SBCL the rt package is named sb-rt. 2) The string callback test does not compile nor does the qsort callback test. 3) All the other tests run. I also have the following observations. * SBCL and CMUCL have alien integer types that can specify a number of bits. I have had success in optimizing lisp ffi by restricting values to xx bits so they can be immediately translated to fixnums. It seems a shame to lose this possiblity in a cffi. * I often need to allocate foreign objects and put them in a lisp struct. I would then want to run a finalizer if the struct gets garbage collected, to free the allocated memory. A nice interface to the finalizer functionality on different systems would be a great feature for cffi. Greetz Immanuel *************************************************************************** I can, I can't. Tubbs Tattsyrup -- Immanuel Litzroth Software Development Engineer Enfocus Software Antwerpsesteenweg 41-45 9000 Gent Belgium Voice: +32 9 269 23 90 Fax : +32 9 269 16 91 Email: Immanuell(a)enfocus.be web : www.enfocus.be ***************************************************************************

3 2

[cffi-devel] setf-expander of FOREIGN-SLOT-VALUE
by Marco Gidde 26 Sep '05

26 Sep '05

The setf-expansion for nested FOREIGN-SLOT-VALUE calls ist not quite correct: CFFI> (macroexpand '(setf (foreign-slot-value (foreign-slot-value ptr `struct1 'slot1) `struct2 'slot2) val)) (LET* () (MULTIPLE-VALUE-BIND (#:G1997) VAL (PROGN (FOREIGN-SLOT-SET #:G1997 (FOREIGN-SLOT-VALUE PTR) 'STRUCT2 'SLOT2) #:G1997))) T ^^^^^^^^^^^^^^^^^^^^^^^^ Defining the expander like (define-setf-expander foreign-slot-value (ptr type slot-name &environment env) (multiple-value-bind (temps vals stores store-form access-form) (get-setf-expansion ptr env) (declare (ignore store-form stores)) (let ((type-tmp (gensym "TYPE-")) (slot-name-tmp (gensym "SLOT-NAME-")) (store (gensym "STORE-"))) (values `(,type-tmp ,slot-name-tmp ,@temps) `(,type ,slot-name ,@vals) `(,store) `(progn (foreign-slot-set ,store ,access-form ,type-tmp ,slot-name-tmp) ,store) `(foreign-slot-value ,access-form ,type-tmp ,slot-name-tmp))))) should do the job a bit better, though it might be optimized in some way (this is my first try to define a setf expander). Btw. the manual contains an example use of FOREIGN-SLOT-VALUE with the simplified syntax (foreign-slot-value ptr 'struct 'slot-1 'slot-2) I think CFFI has all the informations at hand to handle this, so will it be implemented one day? Regards, Marco Gidde

2 4

[cffi-devel] Good news regarding CLISP
by Luis Oliveira 26 Sep '05

26 Sep '05

Hello, I just noticed that CLISP 2.35 has integrated the RTLD_GLOBAL patch and consequently CFFI now works with CLISP out of the box. Passes all tests on Win32, Linux/x86 and Darwin/ppc32. Updated CFFI's webpage, the implementation status table is getting greener and greener. :-) -- Luís Oliveira http://student.dei.uc.pt/~lmoliv/ Equipa Portuguesa do Translation Project http://www.iro.umontreal.ca/translation/registry.cgi?team=pt

1 0

[cffi-devel] Recent changes, new operator foreign-symbol-ptr
by Luis Oliveira 24 Sep '05

24 Sep '05

Hello, Here are the latest changes in cffi-luis. The new operator foreign-symbol-ptr is documented here: http://common-lisp.net/project/cffi/manual/html_node/ foreign_002dsymbol_002dptr.html defcvar is now more dynamic (the way Kenny likes), it now looks up the address of the foreign var at run-time when accessing it. Sat Sep 24 21:13:05 WEST 2005 Luis Oliveira <loliveira(a)common-lisp.net> * foreign-symbol-ptr and defcvar, etc.. - defcvar's expansion will now lookup the foreign symbol at runtime. get-var-ptr too. - Changed a few implementations to return NIL when foreign symbol doesn't exist (instead of, say, raising an error): clisp, cmucl, lispworks and sbcl. - Exported foreign-symbol-ptr from the CFFI package and added it to the user manual. - Also added missing word to foreign-free's description. - Fixed typo in cffi-corman-lisp. Sat Sep 24 21:01:46 WEST 2005 Luis Oliveira <loliveira(a)common-lisp.net> * Support for my asdf setup. Well, support for asdf setups that don't put fasl files next to the source files. The Makefile will delete the .fasls directories on make clean (that's where my setup puts the fasls) and bindings.lisp will save the shared library's location at compile-time. Tue Sep 20 18:35:32 WEST 2005 Luis Oliveira <loliveira(a)common-lisp.net-ptr - new cffi-sys primitive: foreign-symbol-ptr (still needs semantics to be defined when dealing with undefined foreign symbols) - defcvar now uses this /at load-time/. Fri Sep 16 01:47:00 WEST 2005 Luis Oliveira <loliveira(a)common-lisp.net> * Don't free lisp allocated memory on the C side. Thu Sep 15 15:22:38 WEST 2005 Luis Oliveira <loliveira(a)common-lisp.net> * Fix bug with arrays in uffi-compat. -- Luís Oliveira http://student.dei.uc.pt/~lmoliv/ Equipa Portuguesa do Translation Project http://www.iro.umontreal.ca/translation/registry.cgi?team=pt

1 0

[cffi-devel] Shareable Byte Vectors for SBCL
by Faré 22 Sep '05

22 Sep '05

Here's a code snippet I've used in my patch to sb-bsd-sockets, that implements something similar to your "Shareable Byte Vectors" for SBCL. Tested with recvfrom and sendto. [ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ] The ultimate result of shielding men from the effects of folly is to fill the world with fools. -- Herbert Spencer ;;; This little protocol could and should be used in a whole lot of I/O functions. ;;; It silently assumes that buffers are big enough to fit the length required, ;;; and are either subtypes of characters or integer, and that data will fit. (deftype iobuffer () "buffer for direct I/O" '(simple-array (unsigned-byte 8) (*))) (defun iobufferp (x) "check if X is a proper iobuffer" (typep x '(iobuffer))) (defun iobuffer-to-data (buffer data &optional (length (length buffer))) "copy contents of iobuffer to data array" (if (subtypep (array-element-type data) 'integer) (replace data buffer :end1 length) (loop for i from 0 below length do (setf (char data i) (code-char (aref buffer i)))))) (defun data-to-iobuffer (data buffer &optional (length (length data))) "copy contents of data array to iobuffer" (if (subtypep (array-element-type data) 'integer) (replace buffer data :end1 length) (loop for i from 0 below length do (setf (aref buffer i) (char-code (char data i)))))) (defun ensure-iobuffer-from-data (data &optional (length (length data))) "if data is not an iobuffer already, create a new one with similar contents" (if (iobufferp data) data (let ((buffer (make-array (list length) :element-type '(unsigned-byte 8)))) (data-to-iobuffer data buffer length) buffer))) (defun prepare-iobuffer-for-data (data &optional length) "if data is not an iobuffer already, create a new one with same length" (if (iobufferp data) data (make-array (list length) :element-type '(unsigned-byte 8)))) (defun commit-iobuffer-to-data (buffer data &optional length) "if data is not an iobuffer already, replace its contents with those of given buffer" (unless (eq buffer data) (iobuffer-to-data buffer data length))) (defmacro with-iobuffer-sap ((buffer data &key (direction :input) length) &body body) "bind buffer with an iobuffer suitable for I/O of given direction, to be committed to/from the contents of data array data" (let ((data-tn (gensym)) (length-tn (gensym))) `(let* ((,data-tn ,data) (,length-tn ,length) (,buffer (ecase ,direction (:input (prepare-iobuffer-for-data ,data-tn ,length-tn)) ((:output :io) (ensure-iobuffer-from-data ,data-tn ,length-tn))))) (prog1 (sb-sys:with-pinned-objects (,buffer) (let ((,buffer (sb-sys:vector-sap ,buffer))) ,@body)) (case ,direction ((:input :io) (commit-iobuffer-to-data ,buffer ,data-tn ,length-tn))))))) ;;; Should I add a :start argument for the data? ;;; Have :start and :end instead of :start and :length?

1 0

[cffi-devel] pointer issue and SoC eval
by Kenny Tilton 20 Sep '05

20 Sep '05

OK, my opengl demo now runs via CFFI bindings on clisp/win32. (a) I had to rearrange things so libraries get loaded during the ASDF build, before any bindings for said libraries do. (b) The CLisp equivalent of Lispworks's make-pointer is unsigned-foreign-address, which was necessary to pass an enum as a pointer argument (odd Glut/win32 issue). Congrats, I will now do a (favorable) Google mentor evaluation for CFFI-luis. -- Kenny Why Lisp? http://wiki.alu.org/RtL_Highlight_Film "I've wrestled with reality for 35 years, Doctor, and I'm happy to state I finally won out over it." Elwood P. Dowd, "Harvey", 1950

2 1

[cffi-devel] Interesting problem passing faux pointers
by Kenny Tilton 20 Sep '05

20 Sep '05

Background ======== The OpenGL Glut utility has an oddity. Font utilities such as glutStrokeChar take as their first argument a pointer. The oddity is that on X11 an actual pointer to font info is expected, but on win32 an enum is expected. The enum indicates which of a half dozen built-in fonts to use. So on win32 I need to pass something like 0 or 1 as a pointer. Under UFFI, for ACL and LW I was forced to make a special pointer object to pass thru the FFI. Here is the Lispworks code: (defun enum-to-pointer (enum) (fli:make-pointer :address enum :pointer-type '(:pointer :void))) There was something similar for ACL. Problem: with CFFI, I could (and had to) eliminate the ACL wrapping. But Lispworks still needs it, and I just ran into what looks like the same thing from CLisp. Is this oddity something that is not covered by the test suite? Or did I miss some new CFFI make-pointer? The CFFI reference chapter on pointers is a little sparse. :) -- Kenny Why Lisp? http://wiki.alu.org/RtL_Highlight_Film "I've wrestled with reality for 35 years, Doctor, and I'm happy to state I finally won out over it." Elwood P. Dowd, "Harvey", 1950

2 1

[cffi-devel] Clisp load foreign library
by Kenny Tilton 19 Sep '05

19 Sep '05

I passed a pathname to load-foreign-library and Clisp complained that it was not a string. Maybe I missed something. Anyway, if CLisp indeed has a limitation in re accepting a pathname for a library, maybe the CFFI wrapper could add that flexibility. -- Kenny Why Lisp? http://wiki.alu.org/RtL_Highlight_Film "I've wrestled with reality for 35 years, Doctor, and I'm happy to state I finally won out over it." Elwood P. Dowd, "Harvey", 1950

2 7

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

cffi-devel September 2005