April 2014 - cffi-devel - mailman3.common-lisp.net

[Cffi-devel] FSBV question
by Cyrus Harmon 01 Jun '14

01 Jun '14

So I’ve been trying to get FSBV working on ABCL and I’ve run into a problem with returning structs. The problem is that we allocate memory on the stack for the return struct, successfully call the foreign function, writing the return value into the appropriate place, then we free the memory and return a pointer to the now defunct memory location. I’m curious as to how this is supposed to work in other implementations. It obviously does work in, at least, SBCL, so I must be doing something wrong. But looking at the ffcall-body-libffi function it looks like we setup the foreign memory for RESULT in with-foreign-objects and then if we have a translatable-foreign-type, we return RESULT, which of course gets freed by with-foreign-objects. What am I missing here? thanks, Cyrus (defun ffcall-body-libffi (function symbols return-type argument-types &optional pointerp (abi :default-abi)) "A body of foreign-funcall calling the libffi function #'call (ffi_call)." (let ((number-of-arguments (length argument-types))) `(with-foreign-objects ((argvalues :pointer ,number-of-arguments) ,@(unless (eql return-type :void) `((result ',return-type)))) (loop :for arg :in (list ,@symbols) :for count :from 0 :do (setf (mem-aref argvalues :pointer count) arg)) (call (prepare-function ,function ',return-type ',argument-types ',abi) ,(if pointerp function `(foreign-symbol-pointer ,function)) ,(if (eql return-type :void) '(null-pointer) 'result) argvalues) ,(if (eql return-type :void) '(values) (if (typep (parse-type return-type) 'translatable-foreign-type) ;; just return the pointer so that expand-from-foreign ;; can apply translate-from-foreign 'result ;; built-in types won't be translated by ;; expand-from-foreign, we have to do it here `(mem-aref result ',return-type))))))

2 4

[Cffi-devel] Help perfecting my GC finalizer/constructor
by Joeish W 30 Apr '14

30 Apr '14

Thanks for your replies to my question, I was planning to go the trivial-garbage route and someone from CFFI-DEVEL kindly gave me this code to do GC automatically using my mat and del-mat functions (defstruct (cvmatrix (:constructor %make-cvmatrix)) (sap (mat) :type sb-sys:system-area-pointer :read-only t)) (defun make-cvmatrix () (let* ((matrix (%make-cvmatrix)) (sap (cvmatrix-sap matrix))) (tg:finalize matrix (lambda (x) (del-mat sap))) matrix)) When I run the make-cvmatrix function to create a matrix it outputs a struct instead of a pointer like i needed. I changed the last line of the make-cvmatrix function to be "sap" as below: (tg:finalize matrix (lambda (x) (del-mat sap))) sap)) ...and it appears to run poorly with only major variations in my ram levels which i suppose is just a result of lisps GC in process. Since I'm new to GC w/ trivial-garbage, I was hoping someone can verify first that the change to sap was an ok move. I was hoping someone could edit my code so I have a good example of Lisp GC with TG to work from and if I write a defcfun for this C wrapper : Mat* cv_create_Mat_typed(int rows, int cols, int type) { return new Mat(rows, cols, type); } eg: (defcfun ("cv_create_Mat_typed" mat-typed) (:pointer mat) "MAT constructor with a row, column and type parameter." (rows :int) (cols :int) (type :int)) show me where the rows cols type params get placed in an edited finalizer/constructor above..I would definately appreciate greatly:), concrete examples using the code I posted.

3 16

[Cffi-devel] Lists of METAOBJECTS and FOREIGN-ALLOC
by Joeish W 30 Apr '14

30 Apr '14

I have this conditional from a cond statement I need to satisfy ((listp (first args)) (c-arr-to-vector-point (first args))) The (first args) will be this but it should be able to have an infinite number of (point i j) in it. (foreign-alloc :pointer :initial-contents (LIST (POINT 1 2) (POINT 3 4))) The thing is that the output of (point i j) is a metaobject e.g. #<CV-POINT {10038888E3}> . That is how I have my return for point defined. When I need to use foreign-alloc to define an array of point I have to do it like this... (foreign-alloc :pointer :initial-contents (LIST (c-pointer (POINT 1 2)) (c-pointer (POINT 3 4)))) ...because my defclass is this: (defclass std-vector-point () ((c-pointer :reader c-pointer :initarg :c-pointer))) Here is my function I need to change so a when it is run I can just enter (LIST (POINT 1 2) (POINT 3 4)) for the (first args) in the above mentioned conditional block and have the %c-arr-to-vector-point function be satisfied, because it accepts a list of pointers as its input. But I need help because I don't want the user to see the c-pointer at all let alone have to wrap each (point i j) like this (c-pointer (point i j)) that is too much typing. How would I do this without it making my function any slower. (defun vector-point (&rest args) (cond ((fourth args) (error "odd number of args to VECTOR-POINT")) ((null (first args)) (%vector-point)) ((listp (first args)) (%c-arr-to-vector-point (foreign-alloc :pointer :initial-contents (first args)))) ((eq :size (second args)) (%vector-point-size (first args))) ((and (eq (type-of (first args)) 'std-vector-point) (second args) (not (third args))) (mem-aref (c-pointer (%vector-point-to-c-array (first args))) :pointer (second args))) ((and (eq (type-of (first args)) 'std-vector-point) (second args) (third args)) (mem-aref (c-pointer (mem-aref (c-pointer (%vector-point-to-c-array (first args))) :pointer (second args))) :int (third args))) ((not (eq (type-of (first args)) 'std-vector-point)) (error "The value ~a is not of type (STD-VECTOR-POINT)." (first args)))))

2 2

[cffi-devel] cffi-grovel: signal GROVEL-ERROR instead of CL:SIMPLE-ERROR
by Anton Vodonosov 29 Apr '14

29 Apr '14

Hello. In the cl-test-grid project we test how every library in Quicklisp is compiled and also run some test suites. Then we build reports, review regressions and submit bugs to library authors. It is very often that machine where tests are run don't have some foreign library, of don't have a C compiler. For libraries using cffi-grovel it results in cl:simple-error: External process exited with code 1. (example log: http://cl-test-grid.appspot.com/blob?key=1sisosse6j) This kind of errors are not regressions, so we don't want to report them. It would be very helpful to distinguish such errors automatically, because they are not regressions and we don't need to report them to authors. As this type of errors happens all the time during testing on various machines, automation will save us significant efforts. We could easily distinguish these errors if cffi-grovel signaled specific condition instead of generic cl:simple-error. Possible simple fix is to introduce new condition cffi-grovel:grovel-error and shadow cl:error in the cffi-grovel package, instead writing an own cff-givel:error function with the same prototype as cl:error, but with the condition type it signals will be cffi-grovel:grovel-error. If you agree to this proposal I am willing to provide a patch. Best regards, - Anton

2 2

[Cffi-devel] How do I define a CFFI foreign-type so it can be used in a function that accepts a different foreign type as a parameter
by Joeish W 29 Apr '14

29 Apr '14

I have thes 3 defcfuns, for OpenCV functions (C++) which have been wrapped in C(C and C+++ function definitions have been included), the top 2, %brisk and surf0 can both can be used as the (self feature-detector) param in the bottom one, %feat-detect-detect If the output of the top two and the self param of feat-detect-detect are just :pointer, But I need to have all three of their types defined similiarly to the ones on the bottom of the page so I can use finalizers. All of the finalizers and types for the three are the same, just substitute in brisk and surf for feature-detector in the four functions at the bottom of the page and you will know what each looks like. Should I just call the types for the output of brisk and surf and the input type offeat-detect-detect by the same name, and make them all use the same 4 function type/finalizer combo at the bottom of the page, or is there a better way. ;; BRISK::BRISK(int thresh=30, int octaves=3, float patternScale=1.0f);; BRISK* cv_create_BRISK(int thresh, int octaves, float patternScale)(defcfun ("cv_create_BRISK"%brisk)brisk (thresh :int)(octaves :int)(pattern-scale :float));; SURF::SURF();; SURF* cv_create_SURF() (defcfun ("cv_create_SURF"surf0)surf);; void FeatureDetector::detect(const Mat& image, vector<KeyPoint>& keypoints, const Mat& mask=Mat() ) const;; void cv_FeatureDetector_detect3(FeatureDetector* self, Mat* image, vector_KeyPoint* keypoints, Mat* mask)(defcfun ("cv_FeatureDetector_detect3"%feat-detector-detect):void (self feature-detector)(image mat)(keypoints (:pointer vector-keypoint))(mask mat))(define-foreign-type feature-detector ()((garbage-collect :reader garbage-collect :initform nil:initarg :garbage-collect))(:actual-type :pointer)(:simple-parser feature-detector))(defclass cv-feature-detector ()((c-pointer :reader c-pointer :initarg :c-pointer)))(defmethod translate-to-foreign ((lisp-value cv-feature-detector)(c-type feature-detector))(c-pointer lisp-value))(defmethod translate-from-foreign (c-pointer (c-type feature-detector))(let((feature-detector (make-instance 'cv-feature-detector:c-pointer c-pointer)))(when (garbage-collect c-type)(tg:finalize feature-detector (lambda()(del-feature-detector c-pointer))))feature-detector))ound matches of keypoints from two images." (%draw-matches img1 keypoints1 img2 keypoints2 matches1to2 outimg match-color single-point-color matches-mask flags))

1 0

[Cffi-devel] readded broken link from last post
by Joeish W 26 Apr '14

26 Apr '14

Lisp code is here https://gist.github.com/W-Net-AI/11205892

2 5

[Cffi-devel] Fw: Possible mem-aref bug or maybe my error on accessing member of a pointer to a struct
by Joeish W 26 Apr '14

26 Apr '14

On Saturday, April 26, 2014 5:39 AM, Joeish W <joeish80829(a)yahoo.com> wrote: Thank you very much, I should have went back to my C ways of doing things, but it was so odd to see a struct in the OpenCv C++ interface I must have got a little cocky. Again I really appreciate all your attention to these 3 matters today > > > >On Saturday, April 26, 2014 4:32 AM, Willem Rein Oudshoorn <woudshoo(a)xs4all.nl> wrote: > >Joeish W <joeish80829(a)yahoo.com> writes: >> >>> I have this C wrapper >>> >>> >>> DMatch* cv_create_DMatch3(int _queryIdx, int _trainIdx, float _distance) { >>> return new DMatch(_queryIdx, _trainIdx, _distance); >>> } >>> >>> >>> it is a wrapper for these struct members >>> >>> http://docs.opencv.org/modules/features2d/doc/common_interfaces_of_descript… >>> >>> >>> here is my lisp wrapper for it: >>> >>> (defctype dmatch :pointer) >>> >>> (defcfun ("cv_create_DMatch3" dmatch3) (:pointer dmatch) >>> "DMatch constructor" >>> (-query-idx :int) >>> (-train-idx :int) >>> (-distance :float)) >>> >>> In C I can do this: >>> >>> DMatch* a = cv_create_DMatch3(1, 2, 3.0); >>> cout << a->distance ; >>> >>> and the output is 3 >>> >>> but in lisp when I do this: >>> >>> (defparameter a (dmatch3 1 2 3f0)) >>> >>> LCV> (mem-aref a :int 0) >>> 1 >>> LCV> (mem-aref a :int 1) >>> 2 >>> LCV> (mem-aref a :float 2) >>> #<SINGLE-FLOAT quiet NaN> <--- this is output >>> >>> >>> Any help on figuring out the reason would be much appreciated >> >> >>This is not really suprising. You are treating the resulting pointer >>`a` as an untyped pointer. >>When you do `(mem-aref a :float 2)` you are basically saying: >> >>1. Assume a is a pointer to an array of floats >>2. Give me the third element (element at index 2). >> >>Unfortunately `a` is not a pointer to an array of floats, but a pointer >>to a struct DMatch. >> >>You should lookup `defcstruct` in the CFFI documentation: >> >>http://common-lisp.net/project/cffi/manual/html_node/defcstruct.html#defcst… >> >>Wim Oudshoorn. >> >> >>_______________________________________________ >>Cffi-devel mailing list >>Cffi-devel(a)common-lisp.net >>http://common-lisp.net/cgi-bin/mailman/listinfo/cffi-devel >> >> >> > >

1 0

[Cffi-devel] Possible mem-aref bug or maybe my error on accessing member of a pointer to a struct
by Joeish W 26 Apr '14

26 Apr '14

I have this C wrapper DMatch* cv_create_DMatch3(int _queryIdx, int _trainIdx, float _distance) { return new DMatch(_queryIdx, _trainIdx, _distance); } it is a wrapper for these struct members http://docs.opencv.org/modules/features2d/doc/common_interfaces_of_descript… here is my lisp wrapper for it: (defctype dmatch :pointer) (defcfun ("cv_create_DMatch3" dmatch3) (:pointer dmatch) "DMatch constructor" (-query-idx :int) (-train-idx :int) (-distance :float)) In C I can do this: DMatch* a = cv_create_DMatch3(1, 2, 3.0); cout << a->distance ; and the output is 3 but in lisp when I do this: (defparameter a (dmatch3 1 2 3f0)) LCV> (mem-aref a :int 0) 1 LCV> (mem-aref a :int 1) 2 LCV> (mem-aref a :float 2) #<SINGLE-FLOAT quiet NaN> <--- this is output Any help on figuring out the reason would be much appreciated

2 1

[Cffi-devel] How to convert a c++ vector to lisp vector
by Joeish W 25 Apr '14

25 Apr '14

I have these C bindings for the C++ vector which I have wrapped in CFFI. I know how to create vectors with std_carrayTovector and convert the data back to a int pointer with std_vectorToCArray so that I can retrieve data from it in Lisp using the CFFI function MEM-AREF. I have correct defcfuns written for the below . My question is how to I convert the output of my defcfun for std_vectorToCArray at the bottom of the page, into a Lisp vector eg #(1 2 3) and make it a 0(1) operation/all data copied at the same time. vector_int*std_carrayTovector(int*a,size_tlen){vector<int>*v =newvector<int>;for(size_ti =0;i <len;i++)v->push_back(a[i]);returnv;}int*std_vectorToCArray(vector_int*s){returns->data();}(defcfun ("std_vectori_to_carray"%vector-int-to-c-array):pointer (s (:pointer vector-int)))

1 0

[Cffi-devel] Inconsistencies between c and lisp code
by Joeish W 23 Apr '14

23 Apr '14

This takes a little bit of code to explain but I posted the code in gist and notated it thoroughly first ..I'm writing a lisp wrapper for OpenCV c++ interface and my lisp code wraps around c wrappers for the c++ so in my code used for testing there will be my c wrappers and c++ OpenCV code intermixed. Here is an example of what I can do in C https://gist.github.com/W-Net-AI/11205899 that I can't in Lisp, the Lisp code is here https://gist.github.com/W-Net-AI/11205892 The question I'm asking hinges on the output of the detectMultiScale c wrapper for the c++ OpenCV function detectMultiScale here:http://docs.opencv.org/modules/objdetect/doc/cascade_classification.ht… that's why my code examples are long because it takes a few lines to set up being able to use detectMultiScale. Pls let me know and I can supply more info/ such as my wrappers etc.

1 0