cffi-devel January 2006

cffi-devel@common-lisp.net

34 participants
97 discussions

[cffi-devel] Patch: FOREIGN-FUNCALL for LispWorks
by Edi Weitz 26 Jan '06

26 Jan '06

Hi! Attached is a hasty patch to enable CFFI-SYS:%FOREIGN-FUNCALL and CFFI-SYS:%FOREIGN-FUNCALL-POINTER for LispWorks. The idea is to create one foreign funcallable per signature and cache those in a hash table. I took a quick glance at the source code of CMUCL's ALIEN-FUNCALL and my impression (not really understanding the details) was that CMUCL actually does something similar internally - it creates a "stub" if necessary, depending on the types. As you can observe at the REPL it compiles /something/ when ALIEN-FUNCALL is called. Granted, it is kind of ugly to call FLI:DEFINE-FOREIGN-FUNCALLABLE in non-top-level position but I think this solution is better than nothing. It looks to me as if FOREIGN-FUNCALL is mostly intended for interactive use and for "one-shot" invocations - correct me if I'm wrong. I don't know if CMUCL's "stubs" will ever be garbage-collected, the foreign funcallables in my patch certainly won't. I don't expect this to be a big problem, though. I tried the CFFI test suite and it basically ran through with three exceptions - FUNCALL.FLOAT, DEREF.FLOAT.2, and DEREF.FLOAT.3. My guess is that these are due to LispWorks identifying FLOAT and DOUBLE-FLOAT and are not related to my patch. Two minor quibbles, unrelated: 1. The documentation says it is only about the cffi-luis branch but it seems to me that the main branch and cffi-luis are almost identical - or did I miss something? 2. This form (define-foreign-library libcurl (:unix (:or "libcurl.so.3" "libcurl.so")) (t (:default "libcurl"))) in the (very nice, BTW) tutorial didn't work for me on Linux, neither with LispWorks nor with CMUCL. I had to remove the second (:UNIX) line to make it work. edi@vmware:~$ ls -l /usr/lib/libcurl* lrwxrwxrwx 1 root root 16 Dec 28 19:29 /usr/lib/libcurl.so.3 -> libcurl.so.3.0.0 -rw-r--r-- 1 root root 196632 Dec 7 12:39 /usr/lib/libcurl.so.3.0.0 Cheers, Edi.

3 5

Re: [cffi-devel] defenum proposal
by Hoehle, Joerg-Cyril 25 Jan '06

25 Jan '06

Justin Heyes-Jones writes: >The problem I am looking at is that SDL uses un-named enums quite a lot. >In addition it does arithmetic on them to create other enums... >#define MAKE_MASK (n) (1<<n) I can't find this. I grep'ed for "enum", "MAKE_MASK" (0 hits) and "<<" (5 hits) in /usr/include/SDL/*.h and found: o plenty of typedef enum { ... } SDL_audiostatus; which does not sound anonymous at all, even though its not enum <name> { ... }; o a single #define SDL_EVENTMASK(X) (1<<(X)) used by enum SDL_EventMasks, referring to enum SDL_Events o and a single really anonymous one: enum { SDL_PRESSED = 0x01, SDL_RELEASED = 0x00 }; where SDL_PRESSED is reused via #define SDL_BUTTON(X) (SDL_PRESSED<<(X-1)) I'd say SDL_PRESSED is worth a defconstant. Did I look at the wrong place or not understand anything at all? [on a Ubuntu Breezy Debian i686 machine, with libsdl-1.2] BTW, does anybody know which package holds the gl* Manpages? I have the SDL_*.3.gz ones now, from package libsdl1.2-dev. Sadly, packages.debian.org (really valuable site) is down. Regards, Jörg Höhle.

3 2

[cffi-devel] cl+ssl & shareable-vector alternatives
by Hoehle, Joerg-Cyril 25 Jan '06

25 Jan '06

Hi, I wrote: >Lisp vectors at non-moving locations. I haven't yet investigated how >one could make transparent use of such a feature, even though any >Allegro user would find it suboptimal if cl+ssl (or any other library) >would not make best use of her/his Lisp implementation. Here's a possible solution path: Let the cl+ssl's stream object have slots for both a Lisp vector and a foreign entity. The latter is either a pointer to the Lisp vector, if it could be allocated with a fixed location (Allegro (unverified)), or a foreign-alloc'ed memory area in all other cases. Then read-ssl uses this pointer (or array). In the non-Allegro case, CFFI's block copying routine is invoked to copy from foreign to Lisp memory. So this does not go without #+/#- in the cl+ssl library. To me, this is not surprising. It corroborates what Duane Rettig said some time ago in comp.lang.lisp: there won't be a unification of CL's many FFI. He did not say that because UFFI did not exist at that time, but because beyond the common features that can be captured by UFFI and CFFI, there's a great variability among implementations. Similarly, all implementations feature some extensions beyond ANSI-CL! Another consequence of all I said is that I consider the shareable- vector.txt specification not mature or possibly to be rejected. o :in and/or :out are missing, and BTW would lead to what James Bielman dislikes: inconsistency, because some implementations would ignore it, being able to bypass copying. o without :in/out specification, the double copying penalty that I talked about, and that the specification's last example contains, is not acceptable. o Implementating this using si::without-gc is not acceptable (e.g. cmucl) because of a too long and uncontrolled lifetime. Better use the vector copying API. Be safe, not disabled. o It's somewhat superfluous given a good vector copying API. So what's left? Not enough, IMHO. Something that solely works on Allegro (IIRC they can allocate Lisp vectors at a fixed address). IMHO there's not enough benefit for other implementors to add that. I myself would prefer, if any, the converse: allow all SEQUENCE functions to work on foreign arrays... Regards, Jörg Höhle

1 0

[cffi-devel] Unloading a library
by Justin Heyes-Jones 25 Jan '06

25 Jan '06

I saw some discussion in the archives about extending the library loading functionality. There's a couple of things I'd like that I didn't see, and this may be an oversight on my part... 1) CFFI keep track of which foreign modules it had loaded and allow you to iterate through them 2) Have CFFI allow you to unload a foriegn library by name, which should fail if it is not loaded, otherwise return true. I'd find this handy both for debugging and for cleaning up while a program runs.

2 1

[cffi-devel] New patches: 24-Jan-2006
by loliveira＠common-lisp.net 25 Jan '06

25 Jan '06

Tue Jan 24 18:42:15 CST 2006 Luis Oliveira <loliveira(a)common-lisp.net> * more doc changes; renamed features - change "docs" to be the default target in doc/Makefile. - add Stephen the list of authors in cffi-manual - rename "Index" to "Comprehensive Index" so that Index.html doesn't clash with index.html on OSX. - add varargs example for defcfun - rename features (again) foreign-funcall -> no-foreign-funcall, long-long -> no-long-long. M ./doc/Makefile -4 +4 M ./doc/cffi-manual.texinfo -65 +21 M ./src/cffi-allegro.lisp -2 +2 M ./src/cffi-clisp.lisp -2 +2 M ./src/cffi-cmucl.lisp -4 +1 M ./src/cffi-corman.lisp -2 +3 M ./src/cffi-ecl.lisp -2 +2 M ./src/cffi-lispworks.lisp -2 +2 M ./src/cffi-openmcl.lisp -4 +1 M ./src/cffi-sbcl.lisp -4 +1 M ./src/features.lisp -2 +3 M ./src/types.lisp -4 +4 M ./tests/callbacks.lisp -5 +5 M ./tests/defcfun.lisp -1 +1 M ./tests/foreign-globals.lisp -3 +3 M ./tests/funcall.lisp -3 +3 M ./tests/misc-types.lisp -4 +4 M ./tests/struct.lisp -1 +1 Tue Jan 24 10:55:02 CST 2006 Stephen Compall <scompall(a)nocandysw.com> * stabilize the last few doc changes - display Luis's name with an accent - @sc the warranty - remove @code from section headings and @seealso - reword no-long-long note again - hey, define-foreign-library supports multiple libraries - clarify what *darwin-framework-directories* is actually used for - reorganize *foreign-library-directories* description - in gendocs.sh, fix libtool links as well M ./doc/cffi-manual.texinfo -94 +100 M ./doc/gendocs.sh -3 +19 Mon Jan 23 10:55:41 CST 2006 Luis Oliveira <loliveira(a)common-lisp.net> * documentation changes - removed known issue about modern mode as it should work now. - removed note about callbacks being linux/x86 only. - fixed typo: supported -> unsupported. - remove note about [u]int{8,16,32,64} not being implemented. - add documentation about uint, uchar, ulong, ullong. - fix foreign-alloc's function signature. - update info regarding lispworks's new foreign-funcall. - add example for *darwin-framework-directories*. - fixed define-foreign-library's syntax. - comment out sentence about use-foreign-library setting the current foreign library as it doesn't do that yet. - remove the (setf (callback ...)) example. this isn't allowed anymore. - rename foreign-funcall and long-long to no-foreign-funcall and no-long-long. - long-long is a standard C99 type, remove mention of GNU there. - removed TODO item: there's already a tutorial. - added TODO item: document defcfun's new capabilities M ./TODO -2 +1 M ./doc/cffi-manual.texinfo -300 +257 Sun Jan 22 22:06:05 CST 2006 Stephen Compall <scompall(a)nocandysw.com> * document new library interface, some other manual stuff - New macro @Variable. - Fix 2 @end examples. - In "Arguments and Values" headings, use @var for tables of variables, as they act as metasyntactic variables in this situation. - Document new library interface. - Use @unnumberedsec in dictionary nodes instead of @heading, and @code the names. - Platform-specific features: new appendix. - define-curl-option-setter: curl-code is not a keyword - Use new @seealso macro for "See Also" xrefs. - Merge all indexes into cp. M ./doc/cffi-manual.texinfo -215 +730 Thu Jan 19 12:59:20 CST 2006 Stephen Compall <scompall(a)nocandysw.com> * update manual Introduction and Foreign types, texinfo idioms throughout - Remove note on cffi-luis branch. - Write introductory chapter. - Acronym/sc-ify CL implementation names. - Use @cffi{} instead of CFFI. - Use @impnote for notes from luis and james. - Split existing type translator explanation to make "Defining typedefs" and rewrite the former. - Invent an example for defcstruct, and explain its limitations. - Write "Allocating Foreign Objects". - Move with-foreign-object(s) to Pointers chapter, to group it with foreign-alloc. - Use @sc and @result to make REPL examples pretty, and move them into @example sections. - Remove commented-out sections; they can be readded later from darcs repo if need be. M ./doc/cffi-manual.texinfo -442 +544 An updated tarball of CFFI's source can be downloaded here: http://common-lisp.net/project/cffi/tarballs/cffi-060124.tar.gz Darcsweb URL: http://common-lisp.net/cgi-bin/darcsweb/darcsweb.cgi?r=cffi;a=summary

1 0

Re: [cffi-devel] defenum proposal
by Justin Heyes-Jones 24 Jan '06

24 Jan '06

This may be a repost as I got a bounce... The problem I am looking at is that SDL uses un-named enums quite a lot. In addition it does arithmetic on them to create other enums... enum { a_value, another_value, } then it does #define MAKE_MASK (n) (1<<n) enum { MAKE_MASK(a_value), MAKE_MASK(another_value) } In order to do this I can define a lisp function in my swig interface which does the same thing, but for that to work you need someway of getting to the value of the first enums. Justin

1 0

[cffi-devel] defenum proposal
by Frank Buss 24 Jan '06

24 Jan '06

When translating anonymous enums to Lisp, it might be easier to use something like the defcenum: (defmacro defenum (&body enums) `(progn ,@(loop for value in enums for index = 0 then (1+ index) when (listp value) do (setf index (second value) value (first value)) collect `(defconstant ,value ,index)))) This will be expanded to a block of defconstants: (defenum a (b 10) c) -> (PROGN (DEFCONSTANT A 0) (DEFCONSTANT B 10) (DEFCONSTANT C 11)) It is not necessary to use keywords like for the named enums, because unnamed enums in C are just syntatic sugar for a list of "#define"s, which is mapped to a list of defconstants. I don't know, if this macro fits in the philosophy of CFFI, perhaps in some utility class, and maybe with another name "defanonenum". -- Frank Buss, fb(a)frank-buss.de http://www.frank-buss.de, http://www.it4-systems.de

4 3

AW: [cffi-devel] defenum proposal
by Hoehle, Joerg-Cyril 24 Jan '06

24 Jan '06

Luís Oliveira wrote: >I suppose defcenum shouldn't force the user to use keywords though? Certainly not! Where would you get their value from if they are interned into the keyword package. BTW, I think a SWITCH macro would be a useful companion to ENUM. It's CASE where you can put such things into the selector clauses. Instead of (case event-type (#.SDL-keyboard-event[-enum] (do-this)) ... ... or (cond ((if (eql event-type SDL-keyboard-event[-enum]) ... ((if ... or (cond ((if (enum-eq? event-type SDL-keyboard-event) ...) ... BTW, I hate enums in Lisp and the explosion of constants etc. that they cause. I'm not convinced DEFCONSTANT is appropriate. Maybe all is needed is a mapping from name to values (and vice-versa, the CLISP FFI has it), and nice useable forms like the above SWITCH to hide ugliness. BTW, DEFCONSTANT won't help with types like (:array :character MAX_PATH) which once again, shows that it's not a good translation of C. It's just widespread. (:array :character #.MAX_PATH) is also ugly, but I've never thought about finding something better. FWIW, remember to wrap such uses of defconstant into EVAL-when also compile. Regards, Jörg Höhle

3 5

Re: [cffi-devel] Re: defenum proposal
by Hoehle, Joerg-Cyril 23 Jan '06

23 Jan '06

Luis Oliveira writes: >With FOREIGN-ENUM-VALUE, see >http://common-lisp.net/project/cffi/manual/html_node/foreign_002denum_002dv… >Right, that's what CFFI has, see >http://common-lisp.net/project/cffi/manual/html_node/defcenum.html Thanks for pointing me to the CFFI enum stuff. I forgot that the converters already convert to a symbol, so indeed, the SWITCH is superfluous. BTW, w.r.t. to the CFFI enum API, it's ok to use keywords, not other symbols, since the functions all take a "context" parameter, i.e. the enum type. So the keyword is always relative to the type. The manual does not say what happens a) when multiple enum identifier want to map to the same int. b) how foreign-enum-keyword behaves in this case. I've come across a few .h files which had that, can't remember where. >(case (sdl:get-event-type ...) ; or whatever > (:keyboard ...) > (:mouse ...)) Very, very nice! The best I've encountered. Regards, Jörg Höhle.

1 0

[cffi-devel] [patch] lots of new manual changes
by Stephen Compall 23 Jan '06

23 Jan '06

Most notably, documenting the library interface. Thanks especially to Jonathan `slyphon' Simms on #lisp for pointing out a bug in the tutorial code. I added two more broken cross-references, one to libtool, and one to asdf. I'll fix gendocs.sh for libtool later, but would appreciate suggestions (other than host an asdf manual myself) for asdf. Details follow; patch (140k) at http://csserver.evansville.edu/~sc87/cffi/cffi-manual-2s11.darcs.patch Thu Jan 19 12:59:20 CST 2006 Stephen Compall <scompall(a)nocandysw.com> * update manual Introduction and Foreign types, texinfo idioms throughout - Remove note on cffi-luis branch. - Write introductory chapter. - Acronym/sc-ify CL implementation names. - Use @cffi{} instead of CFFI. - Use @impnote for notes from luis and james. - Split existing type translator explanation to make "Defining typedefs" and rewrite the former. - Invent an example for defcstruct, and explain its limitations. - Write "Allocating Foreign Objects". - Move with-foreign-object(s) to Pointers chapter, to group it with foreign-alloc. - Use @sc and @result to make REPL examples pretty, and move them into @example sections. - Remove commented-out sections; they can be readded later from darcs repo if need be. Sun Jan 22 22:06:05 CST 2006 Stephen Compall <scompall(a)nocandysw.com> * document new library interface, some other manual stuff - New macro @Variable. - Fix 2 @end examples. - In "Arguments and Values" headings, use @var for tables of variables, as they act as metasyntactic variables in this situation. - Document new library interface. - Use @unnumberedsec in dictionary nodes instead of @heading, and @code the names. - Platform-specific features: new appendix. - define-curl-option-setter: curl-code is not a keyword - Use new @seealso macro for "See Also" xrefs. - Merge all indexes into cp. -- Stephen Compall http://scompall.nocandysw.com/blog

3 2

← Newer
1
2
3
4
5
6
7
8
9
10
Older →

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

cffi-devel January 2006