Hello,

Here is the latest Caml Weekly News, week 28 January to 04 February, 2003.

1) Wish: dynamic linking for Ocaml
2) question: "autoconfiguration" of Ocaml code (checking for libraries)
3) @, List.append, and tail recursion
4) XML-RPC server for OCaml
5) New release of Active-DVI
6) Finding the sign of a float
7) WDialog 2.00-test4 released

======================================================================
1) Wish: dynamic linking for Ocaml
----------------------------------------------------------------------
Basile Starynkevitch asked:

I am still wishing full dynamic linking support for Ocaml (at least on
the common x86/Linux plateform, which happens to be mine).

I am delighted to read on comp.lang.ml that such a patch exist (at
least for Ocaml 3.06 on x86) on http://algol.prosalg.no/~malc/scaml/

A big thanks to Malcy for providing it. It is at least a proof of
concept that it is doable!


And a question to our great Ocaml team: do they consider adding
dynamic linking to future Ocaml? As far as I am concerned, it is ok
for me if such features are supported only on few native plateforms
(provided they are available for the bytecode engine).

======================================================================
2) question: "autoconfiguration" of Ocaml code (checking for libraries)
----------------------------------------------------------------------
Basile Starynkevitch asked:

(while working on Poesia - see http://www.poesia-filter.org for more)

Does anyone have some tricks or code (usable in a GPL-ed opensource
project) to facilitate configuration (more precisely checking of
installed tools & libraries) of Ocaml code.

I need something which checks that

Ocaml is installed, with version >= 3.06

ocamlfind is installed (both for bytecode & native)

[actually my wish would be that ocamlfind is part of ocaml, like
camlp4 is]

xstr is installed (both bytecode & native)

netstring is installed (both bytecode & native)

ocamlnet is installed, at least version >= 0.92

etc..

(Unfortunately, I cannot do distribution specific stuff; I want a
script usable on Linux Redhat, Mandrake, Debian, ....)

Gerd Stolpmann answered:

I am normally using self-written shell scripts for that purpose.
They are simple to write, easy to understand, and work for all
Unices.

As an example, here is the configure script of wdialog.
You can find code snippets for all the tests you have
mentioned:

http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/wdialog/wdialog/code/configure?rev=3.6&content-type=text/vnd.viewcvs-markup

Maxence Guesdon added:

You could create a script to launch the ocaml toplevel on a
configuration script written in OCaml, like (without caring about the
shell syntax) :

if `ocaml my_conf_script.ml` then
  echo Type make to compile
else
  echo Damn, the configure script failed !

Then your my_conf_script.ml file could be something like :

let config () =
 let version = Sys.ocaml_version in
 if float_of_string version < 3.06 then
    (
     prerr_endline "Your ocaml is too old !" ;   
     exit 1
    )  
 else
   (
    Printf.printf "Using OCaml version %s is ok" version ;
    print_newline ()
   );
;
 (* other tests using Sys.command for example *)

let _ = config ()


If you change your mind and want to use autoconf, you can find
configure.in examples on the caml humps.

======================================================================
3) @, List.append, and tail recursion
----------------------------------------------------------------------
A very very long thread about stack size and tail recursion started
here: http://caml.inria.fr/archives/200301/msg00181.html
Here is a very little part of it.

Brian Hurt started the discussion:

I hit a bug recently wiith @ and List.append.  Since they're recursive, 
not tail-recursive, on long enough lists Ocaml thinks you've gone 
infinitely recursive and aborts.  The code:


let longlist len =
    let rec longlist_int v c acc =
        if (c == 0) then acc else longlist_int (v + 1) (c - 1) (v :: acc)
    in
    longlist_int 0 len []
;;

let x = longlist 65536 ;;

List.append x [] ;;

Exits with:

Stack overflow during evaluation (looping recursion?).

So does:
x @ [] ;;

You can work around this like:

let append' a b =
   List.rev_append (List.rev a) b
;;

Since both rev_append and rev are tail recursive (looping) and not 
recursive, this works.  But some ad-hoc testing says that this method is 
about 50% slower than normal append for lists short enough not to abort.

Thinking about this, I realized that my code is doing stuff like this all
over the place.  I'm basically doing sparse vector/matrix stuff, handling
(effectively) (colno * value) list for vectors, and (rowno * vector) list
for matrix.  And I may be hitting lists long enough to trip the problem.

Which means I'm currently doing a lot of recursion of the form:

let rec foo x = 
   match x with
       [] -> []
       | head :: tail -> (expr head) :: (foo tail)
;;

for various complexities.  And it has occured to me that all of these 
forms *should* be optimizable into loops.  The general case would work 
something like this in C:

struct list_t {
    void * datum;
    struct list_t * next_p;
}

struct list_t * foo (struct list_t * x) {
    struct list_t * retval = NULL;
    struct list_t ** ptr_pp = &retval;

    while (x != NULL) {
        struct list_t * temp = alloc(sizeof(struct list_t));
        *ptr_pp = temp;
        temp->datum = expr(x->datum);
        temp->next_p = NULL; /* be nice to the GC */
        ptr_pp = &(temp->next_p);
        x = x->next_p;
    }
    return retval;
}

If expr() returned a list, the only change necessary would be to find the 
end of the list before moving on, like:

struct list_t * foo (struct list_t * x) {
    struct list_t * retval = NULL;
    struct list_t ** ptr_pp = &retval;

    while (x != NULL) {
        *ptr_p = expr(x->datum); /* expr allocates the list */
        /* We assume the last element of the list expr() returned has
         * NULL for next_p.
         */
        while (*ptr_p != NULL) {
           ptr_p = &((*ptr_p)->next_p);
        }
        x = x->next_p;
    }
    return retval;
}

Rather than just looking at making @ an inline C function, I think we (the 
Ocaml community) should be looking at adding this more general 
optimization in.

So now we get to my two questions:
a) is anyone working on this/intending to work on this RSN?
b) if the answer to (a) is no, can anyone give me some pointers on where 
to start looking at code, so I can add it in?

Andrew Kennedy said:

The optimization you describe is sometimes known as
"tail modulo cons", and is an example of "destination-passing
style". In other words, the place to put the result (in
this case, the address of the tail of a just-constructed 
cons cell) is passed on in a tail-recursive call.

See "A Functional Representation of Data Structures with a Hole"
by Minamide in POPL'98.

http://www.score.is.tsukuba.ac.jp/~minamide/index.html

Although Minimide formalizes the problem in the context of
a typed intermediate language, it's probably quite easy to 
spot special cases quite far down the compiler pipeline.

Mattias Waldau remarked:

I agree that this is recurring problem, I myself often get bit by
List.map.

It makes it very easy to make non-scalable program, works for input less
that 1000 elements, and the when applied to a large problem it fails
without a trace. It is very difficult to find the location of the
problem if you use the native compiler, and most of these programs
doesn't even work using the byte-code compiler.

So one of my coding guidelines is:
- do not use List.map

I would like a prefer other solutions.

======================================================================
4) XML-RPC server for OCaml
----------------------------------------------------------------------
Hans Ole Rafaelsen announced:

I'm pleased to announce the release of the XmlRPCServer-0.1.0 package
for OCaml.  The package and documentation can be fetched from:

http://www.simula.no/~hans/ocaml/xmlrpc/

XmlRPCServer is a server side implementation for the OCaml
XML-RPC-0.1.0 package developed by Shawn Wagner.  The main motivation
for writing this software was to be able to call services written in 
OCaml from other languages.  Interoperability has been tested against 
XML-RPC implementations in Python and Java, see documentation.  The   
package has been developed and tested under Debian Linux, but will
probably work for other distributions.

XML-RPC and XmlRPCServer might serve as an RPC alternative between  
OCaml applications, at least as long as performance is not an issue.  
The xoridl tool provided with the package is able to create both
client stubs and server skeletons, based on an interface description,
shielding the application programmer from the details of marshaling
and unmarshaling the RPC calls.

If you find any bugs or have problems installing the package, please
let me know, and I'll try to help.

======================================================================
5) New release of Active-DVI
----------------------------------------------------------------------
Pierre Weis announced:

Active-DVI is a presenter and previewer for texts or slides written in LaTeX,
hence the presentation tool of choice for the discriminating hacker.

Version 1.4.0 is now available. As the release number shows, this is a
major improvement w.r.t version 1.2.0 of Active-DVI.

This new version has been developed by Jun Furuse, Didier Rémy and    
Pierre Weis with also contributions by Didier Le Botlan, Roberto Di
Cosmo, Xavier Leroy, and Alan Schmitt.

In addition to the regular features of the Active-DVI previewer:    

   * Encapsulated Postscript File inclusion
     (using the graphics LaTeX package)

   * Some effects for presentation (pause, delay, text color change)

this release introduces a lot of new features

   * Interactive demos, via launching of applications from within the slides

   * Background colors and background images for slides

   * Annotations of texts (visible when the mouse is on the annotated text)

   * Hyper references within the same document or to other dvi files

   * Animated transitions from slide to slide

   * Text movements within the slide

   * Recording of text and arbitrary playback of recorded elements

   * Superposition effects for included images (alpha channels and blending)

   * A manual

   * A lot of examples, including talks using popular LaTeX packages seminar
     and prosper (cf. http://prosper.sourceforge.net/).

   * A new user's interface in the redesigned LaTeX package advi.sty.

Play advi on the demonstration presentation demo.dvi that is in the test
directory of the distribution. Look at source code of various talks in
the directory ``examples''.

The source code, RPM packages, and information are available at

    http://pauillac.inria.fr/activedvi/

Contributions and comments are warmly welcome.

Mailing list advi-list-request@pauillac.inria.fr
Bug reports to advi-bugs@pauillac.inria.fr

======================================================================
6) Finding the sign of a float
----------------------------------------------------------------------
Shawn Wagner asked:

I'm looking for a way, in pure ocaml without having to bail out to C, to
tell if a float is negative or not.

Just using x < 0.0 won't work, as I need to be able to tell the difference
between -0.0 and +0.0. This is for ocaml versions of the C copysign and
signbit functions. Any suggestions?

Chris Hecker answered:

let is_neg v =
(Int64.shift_right_logical (Int64.bits_of_float v) 63) = Int64.one

Chris

PS.  My kingdom for overloaded arithmetic operators and Int32/64 constants.   

======================================================================
7) WDialog 2.00-test4 released
----------------------------------------------------------------------
Gerd Stolpmann announced:

I have just released the fourth test release of WDialog-2.00,
the web application framework.

The tarball is available on the sourceforge project page
http://sourceforge.net/projects/wdialog

This release improves features that were necessary to
write the web application "wtimer" that will be released soon.
Furthermore, it is now (again) possible to compile wdialog
with the development version of pxp.

======================================================================
Old cwn
----------------------------------------------------------------------

If you happen to miss a cwn, you can send me a message
(alan.schmitt@inria.fr) and I'll mail it to you, or go take a look at
the archive (http://pauillac.inria.fr/~aschmitt/cwn/). If you also wish
to receive it every week by mail, just tell me so.

======================================================================

Alan Schmitt