Previous week Up Next week


Here is the latest Caml Weekly News, for the week of 15 to 22 November, 2005.

  1. OCaml-MySQL
  2. Sudoku
  3. THREADED_CODE: Why CODE32 defined on x86-64?
  4. bytecode apps without stdlib/pervasives
  5. LablTk Deuglification Tutorial
  6. new bug tracking system
  7. Objective CAML for Scientists: Chapter 1
  8. "The Whitespace Thing" for OCaml



Daniel Margolis asked and Richard Jones answered:
> With some help, I've managed to get OCaml-MySQL working. But if  
> anyone has any other recommendations of database libraries, still  
> feel free to share them with me.

ocamldbi has a MySQL interface.
David Baelde then said:
It may be worth pointing that ocaml-mysql is not thread-ready. I
reported that a few months ago and nothing was done. The problem is
that the bindings do not declare blocking sections, so when one thread
calls a mysql function (which can take time) the other threads won't
be scheduled.

I'm glad to hear about ocamldbi. Is it well maintained ? has blocking
sections stuff ? If so, we ( should swicth to it.


The editor says:
There were several threads about Sudoku, you may find them here:

Next are some short extracts, with most of the discussions removed.
Alain Frisch said:
Funny. My father told me about the game last sunday and I was willing to
write a solver too :)

Here it is:
Diego Olivier Fernandez Pons said:
There is a nice paper by Helmut Simonis "Sudoku as a constraint problem"
with reference to the relevant paper for the algorithms.
David Thomas said:
Mostly to teach myself Ocaml and Tk, I put together a
simple Sudoku game.  If anyone wants to play with it,
modify it, etc, feel free.

The puzzle generation at present is entirely
braindead, but if someone produces a nice solution,
I'd love to see it included.
Alain Frisch said:
Follow-up to the recent discussion here: I've improved my Sudoku solver 
and released it with some documentation

It was able to deal in reasonnable time with all the 3x3, 4x4 and 5x5 
grids I could find (often <10ms for 3x3 grids). Some difficult puzzles 
are included in the distribution.
Jon Harrop said:
In the face of stiff competition, I have updated my Sudoku solver in OCaml 
with a new implementation that is shorter, simpler and faster (although 
nothing like as fast as Alain's):

I have also included links to Sudoku solvers written in OCaml by Alain Frisch, 
Pascal Brisset and David Thomas.
Arthur Chargueraud said:
Well, if you are interested, you may wish to have a look at my:
- thirteen lines of Caml (though my lines are not so short)
- of a rather efficient sudoku solver
- that gives all solution for an input grid
- and uses only modules Pervasive and Array.
Thanks Caml !

You may find it here:

I haven't done any advanced benchmarking, but it looks like my code isn't too slow.
Marco Maggesi said:
include Set.Make(struct type t = (int * int) * int let compare = compare end)

let (@) g f x = g (f x) and id x = x and sw f x y = f y x and zip x y = (x, y)

let fold9 f = let rec loop i = if i>8 then id else loop (i+1) @ f i in loop 0

let fold81 f = fold9 (fold9 @ (@) f @ zip)

let mark ((i,j),x as e) : t -> t =
  add e @ fold9 (fun k -> remove ((i/3*3 + k/3, j/3*3 + k mod 3), x) @
    remove ((i,j),k) @ remove ((i,k),x) @ remove ((k,j),x))

let search =
  let g p f s = fold (f @ sw mark s) (filter ((=) p @ fst) s) in
  fold81 g

let read () =
  let f p = Scanf.scanf "%d " (fun x -> if x>0 then mark (p,x-1) else
  id) in
  fold81 f (fold81 (fold9 @ ((@) add @ zip)) empty)

let print s () =
  let pr ((i,j),x) = Printf.printf "%d%c" (x+1) (if j=8 then '\n' else ' ') in
  iter pr s; print_newline ();;

search print (read ()) ();;

THREADED_CODE: Why CODE32 defined on x86-64?


Richard Jones said:
Why is CODE32 defined on the x86-64 architecture?  This architecture
has 64 bit pointers, and it seems like it's only by luck that the
bytecode interpreter normally works.

In particular, if you have the interpreter in a shared library, or (I
assume) address-space randomization, then the jump table is located
above the 4GB boundary, and the bytecode interpreter segfaults as soon
as it tries to jump to the first instruction.

I had to apply the following patch to allow the bytecode interpreter
to work in a shared library:

--- ocaml-3.09.0.orig/configure 2005-09-24 10:19:09.000000000 +0100
+++ ocaml-3.09.0/configure      2005-11-16 14:59:56.000000000 +0000
@@ -291,9 +291,7 @@
-    bytecccompopts="-fno-defer-pop $gcc_warnings"
-    # Tell gcc that we can use 32-bit code addresses for threaded code
-    echo "#define ARCH_CODE32" >> m.h;;
+    bytecccompopts="-fno-defer-pop $gcc_warnings";;
     bytecccompopts="-fno-defer-pop $gcc_warnings";;
Xavier Leroy answered:
No, it's not by luck: the x86-64 ABI defines several "code models",
one of which -- the "small code model" -- guarantees that code and
static data is in the lower 2G of the address space, and is the
default for gcc (and I suspect other compilers as well).

I would have expected the dynamic loader to implement the small model
like the static linker does, but from what you say it's not the case.

bytecode apps without stdlib/pervasives


Jonathan Roewen asked and Damien Doligez answered:
> I have a simple ML file (
> external raise : exn -> 'a = "%raise";;
> raise End_of_file;;
> so I can check how the compiler does things (like if predefined ocaml
> symbols are still present without stdlib present), and compile as:
> ocamlc -nostdlib -nopervasives, which generates a.out.
> when I run it, I get error: "-bash: ./a.out: cannot execute binary
> file" (btw, generates expected behabiour without those flags)
> why can't it execute it? can we only use -nostdlib as long as we
> provide our own pervasives module?

Note that a.out is still a valid bytecode file:
   ocamlrun ./a.out

What happens is that ocamlc tries to create an executable bytecode file,
by prepending the standard header (#!/usr/local/bin/ocamlrun) to your
bytecode file.  But it can't find this header because you have specified
-nostdlib, thus preventing it from looking in its library directory.

It explicitely ignores the "file not found" error, I'm not sure why.

What you can do is provide your own header file, named "camlheader", and
placed in one of the directories searched by ocamlc.  If you don't give
any -I option, this has to be the current directory:

$ echo '#!/bin/echo foo' >camlheader
$ ocamlc -nostdlib -nopervasives
$ ./a.out
foo ./a.out

LablTk Deuglification Tutorial


Matt Gushee announced:
Some time ago on this list, there was a discussion of the pros and cons
of the Tk GUI toolkit. In the course of that discussion I claimed that
Tk could be made significantly less ugly ... someone asked how, but I
never answered--sorry!

At any rate, here, finally, is the answer:

new bug tracking system


Xavier Leroy announced:
The bug tracking system for Caml now uses Mantis in replacement for
Jitterbug, which was increasingly showing its age.  The new BTS should
make it much easier to classify bug reports, add follow-ups and
attachments, and be notified of actions taken by the Caml maintainers.

However, bug submission can no longer be done by e-mail; please use
the Web interface exclusively.

Many thanks to Maxence Guesdon for installing Mantis and re-importing
the old bug reports.

Objective CAML for Scientists: Chapter 1


Jon Harrop announced:
Chapter 1 of "Objective CAML for Scientists" is now available on-line as a web 

"The Whitespace Thing" for OCaml


Mike Lin announced:
Brace yourselves for "The Whitespace Thing". This is an OCaml
preprocessor that uses your indentation to group expressions, like in
Python and Haskell. Effectively, the preprocessor auto-parenthesizes
expressions that are split over multiple lines, using your indentation
as clues. This eliminates syntax clutter such as multi-line
parenthesizations, the sequencing operators ; and ;; , and the
keywords done, end, and begin.

Using folding to read the cwn in vim 6+

Here is a quick trick to help you read this CWN if you are viewing it using vim (version 6 or greater).

:set foldmethod=expr
:set foldexpr=getline(v:lnum)=~'^=\\{78}$'?'&lt;1':1

If you know of a better way, please let me know.

Old cwn

If you happen to miss a CWN, you can send me a message and I'll mail it to you, or go take a look at the archive or the RSS feed of the archives.

If you also wish to receive it every week by mail, you may subscribe online.

Alan Schmitt