cffi-devel December 2005

cffi-devel@common-lisp.net

17 participants
66 discussions

[cffi-devel] New patches: 29-Dec-2005
by loliveira＠common-lisp.net 29 Dec '05

29 Dec '05

Thu Dec 29 00:54:27 CET 2005 Luis Oliveira <loliveira(a)common-lisp.net> * Document FOREIGN-SLOT-NAMES and FOREIGN-SLOT-OFFSET. M ./TODO -1 +1 M ./doc/cffi-manual.texinfo -2 +95 M ./src/types.lisp -1 +1 Wed Dec 28 20:58:43 CET 2005 Luis Oliveira <loliveira(a)common-lisp.net> * FOREIGN-SLOT-NAMES and FOREIGN-SLOT-OFFSET - New functions: FOREIGN-SLOT-NAMES and FOREIGN-SLOT-OFFSET. - Test FOREIGN-SLOT-NAMES: STRUCT.NAMES - Removed respective TODO item and added a new one: document these. M ./TODO -1 +1 M ./src/early-types.lisp +9 M ./src/package.lisp +2 M ./src/types.lisp -5 +10 M ./tests/struct.lisp +4 An updated tarball of CFFI's source can be downloaded here: http://common-lisp.net/project/cffi/tarballs/cffi-051229.tar.gz

1 0

[cffi-devel] New patches: 28-Dec-2005
by loliveira＠common-lisp.net 28 Dec '05

28 Dec '05

Wed Dec 28 16:15:36 CET 2005 Luis Oliveira <loliveira(a)common-lisp.net> * Add some links to the TODO file. M ./TODO -1 +4 Wed Dec 28 16:04:24 CET 2005 Luis Oliveira <loliveira(a)common-lisp.net> * New TODO list and updated README. - New file: TODO. 46 items! - Removed the outdated paragraph about implementation support from the README file. M ./README -4 A ./TODO An updated tarball of CFFI's source can be downloaded here: http://common-lisp.net/project/cffi/tarballs/cffi-051228.tar.gz

1 0

[cffi-devel] New patches: 27-Dec-2005
by loliveira＠common-lisp.net 27 Dec '05

27 Dec '05

Tue Dec 27 16:47:53 CET 2005 Luis Oliveira <loliveira(a)common-lisp.net> * Fix compiler macros on foreign-slot-value/set - foreign-struct-slot-value-form and foreign-struct-slot-set-form were not quoting the type in the resulting mem-ref form. - struct.5, regression test. M ./src/types.lisp -2 +2 M ./tests/struct.lisp +13 An updated tarball of CFFI's source can be downloaded here: http://common-lisp.net/project/cffi/tarballs/cffi-051227.tar.gz

1 0

[cffi-devel] Calling function expecting char**
by JoeSB COE9 27 Dec '05

27 Dec '05

Hi, How do I call a function that expect char**? I test it with /** C Code */ #include <stdio.h> void expect_charpp(char** buffs){ printf("I got %s and %s", buffs[0], buffs[1]); } ;; Lisp side (cffi:with-foreign-object (argv :pointer 2) (setf (cffi:mem-ref argv :pointer 0) (cffi:foreign-string-alloc "Hello")) (setf (cffi:mem-ref argv :pointer 1) (cffi:foreign-string-alloc "Hello")) (cffi:foreign-funcall "expect_charpp" :pointer argv)) And I got a memory fault error in SBCL. If i comment out the code in C to make it not accessing the data, it won't err. So it looks like the error is in how I allocate the memory. Please advice. Pisin Bootvong,

2 1

[cffi-devel] New patches: 26-Dec-2005
by loliveira＠common-lisp.net 26 Dec '05

26 Dec '05

Mon Dec 26 14:58:06 CET 2005 Luis Oliveira <loliveira(a)common-lisp.net> * cffi-uffi-compat:allocate-foreign-object optimization Convert TYPE to the equivalent UFFI type at macro-expansion time when constant. M ./uffi-compat/uffi-compat.lisp -1 +4 Mon Dec 26 14:53:22 CET 2005 Luis Oliveira <loliveira(a)common-lisp.net> * Don't save (anonymous) parsed types in *foreign-types* Don't save parsed types in the *foreign-types* hashtable. This way they can be garbage collected and the hashtable won't grow indefinitely. Issue reported by Novikov Leonid <ln (at) bk.ru> M ./src/early-types.lisp -3 +1 An updated tarball of CFFI's source can be downloaded here: http://common-lisp.net/project/cffi/tarballs/cffi-051226.tar.gz

1 0

[cffi-devel] patch for uffi-compat
by Леонид Новиков 26 Dec '05

26 Dec '05

This patch accelerates work with array *** c:\temp\uffi-compat.lisp Wed Nov 9 21:23:12 2005 --- c:\temp\cffi-luis-051114-1956\uffi-compat\uffi-compat.lisp Fri Dec 16 13:53:55 2005 *************** *** 131,137 **** (:documentation "UFFI's :array type.")) (defmethod initialize-instance :after ((self uffi-array-type) &key) ! (setf (cffi::actual-type self) (cffi::find-type :pointer))) (defmethod cffi:foreign-type-size ((type uffi-array-type)) (or (* (cffi:foreign-type-size (element-type type)) (nelems type)) --- 131,144 ---- (:documentation "UFFI's :array type.")) (defmethod initialize-instance :after ((self uffi-array-type) &key) ! (setf (cffi::actual-type self) (cffi::find-type :pointer)) ! (setf (cffi::name self) ! (intern (format nil "~A-~A-~A-~D" ! (cffi::symbol-name '#:anon) ! (cffi::class-name (class-of self)) ! (element-type self) ! (nelems self)))) ! ) (defmethod cffi:foreign-type-size ((type uffi-array-type)) (or (* (cffi:foreign-type-size (element-type type)) (nelems type)) *************** *** 141,148 **** t) (cffi:define-type-spec-parser uffi-array (element-type count) ! (make-instance 'uffi-array-type :element-type element-type :nelems count)) ! ;; UFFI's :(unsigned-)char (cffi:define-foreign-type uffi-char (base-type) base-type) --- 148,158 ---- t) (cffi:define-type-spec-parser uffi-array (element-type count) ! (cond ((cffi::find-type (intern (format nil "ANON-UFFI-ARRAY-TYPE-~A-~D" ! element-type ! count)))) ! (t (make-instance 'uffi-array-type :element-type element-type :nelems count)))) ! ;; UFFI's :(unsigned-)char (cffi:define-foreign-type uffi-char (base-type) base-type)

3 3

[cffi-devel] New patches: 24-Dec-2005
by loliveira＠common-lisp.net 24 Dec '05

24 Dec '05

An updated tarball of CFFI's source can be downloaded here: http://common-lisp.net/project/cffi/tarballs/cffi-051224.tar.gz

1 0

[cffi-devel] Providing access to errno in CFFI-SYS
by James Bielman 24 Dec '05

24 Dec '05

Hi all, While working through my fancy gettimeofday CFFI example, I've added a type translator used as the return type for standard C functions that return a negative integer on error and set 'errno'. Since errno is often a macro expanding to a function, I can't very well define it as a foreign global. Most CL implementations seem to have a function used internally to safely get errno for the current thread. I propose adding such an interface to CFFI-SYS to safely get 'errno'. Here's what I added to CFFI-SYS for SBCL: (defun errno-value () (sb-alien:get-errno)) This should be equally simple for CMUCL, OpenMCL, and LispWorks. I haven't found anything in Allegro CL or CLISP (Jörg?) for this yet. I don't intend to do anything with the return value except call 'strerror' on it, since the values themselves are also defined as macros and not introspectable from Lisp either...argh! (I know errno isn't _completely_ portable---at least one OS I use regularly, Windows CE, doesn't have it at all, even though some functions are documented as setting it! I'm willing to put that off until we run into the problem though...) James

1 0

[cffi-devel] New patches: 23-Dec-2005
by loliveira＠common-lisp.net 23 Dec '05

23 Dec '05

Fri Dec 23 16:35:00 CET 2005 James Bielman <jamesjb(a)jamesjb.com> * remove bogus extra right parenthesis M ./examples/gettimeofday.lisp -2 +2 Fri Dec 23 15:47:49 CET 2005 James Bielman <jamesjb(a)jamesjb.com> * pull out and improve some cffi examples M ./cffi-examples.asd -1 +3 M ./examples/examples.lisp -70 +1 A ./examples/gethostname.lisp A ./examples/gettimeofday.lisp Fri Dec 23 15:47:01 CET 2005 James Bielman <jamesjb(a)jamesjb.com> * define type translators at compile time M ./src/types.lisp -1 +1 Fri Dec 23 14:53:21 CET 2005 Joerg-Cyril Hoehle <hoehle(a)users.sourceforge.net> * clisp %mem-set use unexported ffi::stuff to enforce left-to-right evaluation M ./src/cffi-clisp.lisp -5 +3 Fri Dec 23 14:14:08 CET 2005 Joerg-Cyril Hoehle <hoehle(a)users.sourceforge.net> * more tests: verify left to right evaluation rule M ./tests/memory.lisp +18 Fri Dec 23 14:13:05 CET 2005 Joerg-Cyril Hoehle <hoehle(a)users.sourceforge.net> * precision of wording (offset in bytes) & typo M ./doc/cffi-manual.texinfo -1 +1 M ./doc/cffi-sys-spec.texinfo -1 +1 M ./src/types.lisp -2 +2 Fri Dec 23 13:37:12 CET 2005 Joerg-Cyril Hoehle <hoehle(a)users.sourceforge.net> * clisp: (setf %mem-ref) is gone, optimize %mem-set M ./src/cffi-clisp.lisp +10 Fri Dec 23 13:35:49 CET 2005 Joerg-Cyril Hoehle <hoehle(a)users.sourceforge.net> * with-foreign-object: optimize constant type case + use eval M ./src/types.lisp -6 +7 Fri Dec 23 12:33:30 CET 2005 James Bielman <jamesjb(a)jamesjb.com> * inform rt of expected test failures under openmcl M ./tests/callbacks.lisp +3 M ./tests/foreign-globals.lisp +2 Fri Dec 23 12:35:23 CET 2005 James Bielman <jamesjb(a)jamesjb.com> * macrology to OAOO definition of %MEM-REF and %MEM-SET M ./src/cffi-cmucl.lisp -74 +42 M ./src/cffi-openmcl.lisp -94 +44 M ./src/cffi-sbcl.lisp -75 +42 Fri Dec 23 12:08:05 CET 2005 James Bielman <jamesjb(a)jamesjb.com> * export %MEM-SET from cffi-sys and use instead of (SETF %MEM-REF) M ./src/cffi-allegro.lisp -1 +3 M ./src/cffi-clisp.lisp -1 +2 M ./src/cffi-cmucl.lisp -25 +2 M ./src/cffi-corman.lisp -1 +3 M ./src/cffi-ecl.lisp -1 +2 M ./src/cffi-gcl.lisp -1 +3 M ./src/cffi-lispworks.lisp -1 +2 M ./src/cffi-openmcl.lisp -25 +2 M ./src/cffi-sbcl.lisp -25 +2 M ./src/strings.lisp -2 +3 M ./src/types.lisp -3 +3 Fri Dec 23 10:12:03 CET 2005 Luis Oliveira <loliveira(a)common-lisp.net> * uffi-compat optimization on deref-array Determine the element type at macro-expansion time when the array type is constant. M ./uffi-compat/uffi-compat.lisp -3 +6 Fri Dec 23 10:11:39 CET 2005 Luis Oliveira <loliveira(a)common-lisp.net> * Free old callbacks on clisp. M ./src/cffi-clisp.lisp +3 Fri Dec 23 11:21:06 CET 2005 Luis Oliveira <loliveira(a)common-lisp.net> * Revert brainfart. What missing value? M ./src/types.lisp -1 +1 Fri Dec 23 11:17:48 CET 2005 Luis Oliveira <loliveira(a)common-lisp.net> * Fix conflict and missing "value". Oops. M ./src/types.lisp -4 +4 Thu Dec 22 16:13:02 CET 2005 Joerg-Cyril Hoehle <hoehle(a)users.sourceforge.net> * fix foreign-struct-set-form & enable foreign-slot-set optimization M! ./src/types.lisp -8 +6 Thu Dec 22 15:19:31 CET 2005 Joerg-Cyril Hoehle <hoehle(a)users.sourceforge.net> * clisp: inline %mem-ref when given constant type M ./src/cffi-clisp.lisp +6 Thu Dec 22 15:15:06 CET 2005 Joerg-Cyril Hoehle <hoehle(a)users.sourceforge.net> * fix foreign-struct-*-form thus enable foreign-slot-value optimization M ./src/types.lisp -5 +3 Fri Dec 23 02:43:50 CET 2005 James Bielman <jamesjb(a)jamesjb.com> * fix bugs in example implementation of mem-vector functions M ./doc/mem-vector.txt -3 +4

1 0

[cffi-devel] Branch: New Type Translator Interface
by James Bielman 23 Dec '05

23 Dec '05

Hi all, I've done yet-another rewrite of the type translator interface. It does everything through generic functions at run-time, which allows us to specialize on both the Lisp object being converted, and the foreign type we are converting to. DEFINE-TYPE-TRANSLATOR is gone---instead you simply specialize the generic functions as needed. For example (taken from examples/gettimeofday.lisp): ;; A NULL-POINTER is a :POINTER that must always be NULL. (defctype null-pointer :pointer) ;; Translator that ensures VALUE is a null pointer. NIL is ;; also converted to a null pointer for convenience. (defmethod translate-to-foreign (value type (name (eql 'null-pointer))) (declare (ignore type name)) (cond ((null value) (null-pointer)) ((null-pointer-p value) value) (t (error "~A is not a null pointer." value)))) This method on TRANSLATE-TO-FOREIGN replaces the :TO-C case of DEFINE-TYPE-TRANSLATOR. Specializing the other GFs is similar---I've also added a FREE-TRANSLATED-OBJECT generic function. There isn't yet a mechanism for optimizing the case where we are converting a Lisp object to a C object with dynamic extent---I will add this next, with some sort of compiler-macro-like interface. The branch can be downloaded via: darcs get http://slacknet.com/~jamesjb/cffi-new-translators Please give it a spin and let me know what you think! Documentation updates coming soon... James

1 0

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

cffi-devel December 2005