[Caml-list] Ensemble release 1.31

[Caml-list] Ensemble release 1.31

[Ohad Rodeh]

A new Ensemble release is available. It is an experimental version that
has been primarily tested on Linux. It should work for other Unixes.
The code has been heavily tested, and I'm waiting for comments and bug reports.

All the best,
   Ohad.


     RELEASE_NOTES for Ensemble version 1.31

Author: Ohad Rodeh
Last updated: 13/11/2001

CHANGES

This experimental version contains an improved C-interface,
whose performance is identical to the native ML
interface. Zero-copying technology is used whereby
IO-vectors are passed between C and ML without copying.
This interface is named CE, short for "native Ensemble C interface".
CE has nothing to do with the old C interface, HOT.
An additional chapter was added to the tutorial explaining CE.

In detail:
1. The underlying routers, socket library, and reference counting
   code were completely rewritten.

2. The reference manual and tuturial were updated.

3. Removed 4-byte alignment restriction from the system.

4. No longer supported:
   - Mpi, Atm, and Sp2.
   - Threaded scheduling.
   - The Old ML interface.
   - Removed support for high-encryption, only RC4
     is natively supported. This is to abide by US encryption
     restrictions.
   -RPC code.

5. Small changes and bug fixes:
   - Fixed two race conditions in maestro (contributed by Mouna Seri)
   - Fixed the dbm demo.
   - Outboard now works only with TCP.
   - Catch EBUF error in HSYS.
   - Ported Tk to run on the new LablTk API. The life demo runs
     fine, the wbml does not (yet).
   - Pgp now works on WIN32.
   - A large number of small changes.

OCAML COMPILER VERSION
  We are using version 3.01 for this version.

PORTABILITY
  This version was tested on Linux only. It should run
fine on a Sparc/Solaris architecture, but will not run on WIN32 platforms.



-------------------
Bug reports: http://caml.inria.fr/bin/caml-bugs  FAQ: http://caml.inria.fr/FAQ/
To unsubscribe, mail caml-list-request@inria.fr  Archives: http://caml.inria.fr

[Caml-list] camlp4 + another preprocessor

[Christophe Raffalli]

For some of my code (my PhoX theorem prover) 
I have now to use camlp4o because of stream parser ...

But I still want to use sed as preprocessor (for instance replacing
CAMLSTREAMERROR by ? or ??
depending of the version of Caml people want to use.

The easiest way for me to stay compatible with all the versions of OCaml (from
2.02 I think)
would be that the compiler accept -pp "sed ..." --pp camlp4o as options ...

This mean I can use two preprocessors in a simple way ?

Someone see another solution ?
(I know I can probably use a pipe with -pp "set ... | camlp4o)

 
-- 
Christophe Raffalli
Université de Savoie
Batiment Le Chablais, bureau 21
73376 Le Bourget-du-Lac Cedex

tél: (33) 4 79 75 81 03
fax: (33) 4 79 75 87 42
mail: Christophe.Raffalli@univ-savoie.fr
www: http://www.lama.univ-savoie.fr/~RAFFALLI
-------------------
Bug reports: http://caml.inria.fr/bin/caml-bugs  FAQ: http://caml.inria.fr/FAQ/
To unsubscribe, mail caml-list-request@inria.fr  Archives: http://caml.inria.fr

Re: [Caml-list] camlp4 + another preprocessor

[Daniel de Rauglaudre]

Hi,

On Wed, Nov 28, 2001 at 01:53:48AM +0100, Christophe Raffalli wrote:

> would be that the compiler accept -pp "sed ..." --pp camlp4o as options ...
> This mean I can use two preprocessors in a simple way ?
> Someone see another solution ?

And using a shell script which combines the two preprocessors?

-- 
Daniel de RAUGLAUDRE
daniel.de_rauglaudre@inria.fr
http://cristal.inria.fr/~ddr/
-------------------
Bug reports: http://caml.inria.fr/bin/caml-bugs  FAQ: http://caml.inria.fr/FAQ/
To unsubscribe, mail caml-list-request@inria.fr  Archives: http://caml.inria.fr

Re: [Caml-list] camlp4 + another preprocessor

[Christophe Raffalli]

Daniel de Rauglaudre a écrit :
> 
> Salut,
> 
> > In fdact I failed to pipe ! when I do pipe, both preprocessor loops
> > ? I think camlp4o misses then end of the file. I do not understand !
> 
> Ça m'étonnerait que le "pipe" marche. Le préprocesseur est appelé en
> ajoutant le nom du fichier à la commande que tu mets. Alors si c'est:
>       ocamlc -pp "sed ... | camlp4o" toto.ml
> Ça exécutera:
>       sed ... | camlp4o toto.ml

ça pourrait marcher avec camlp4o -impl - qui lit l'entré standard 
(c'est pas documenté ? mais ça semble marcher ?)

> > I will move to one preprocessor, using a file config.ml to store the value of
> > the constants
> 
> Ou alors tu peux ajouter tes constantes dans une extension de syntaxe ou
> une quotation de Camlp4 et ça les remplacera syntaxiquement.
 
> > it remains the problem of ?? versus ? in stream patern ... why this syntax
> > changed ?
> 
> Oubli. Ça avait été changé de ? en ?? à cause des labels, mais je ne l'avais
> pas fait dans Camlp4. C'est corrigé dans la version courante.

Bon je pourrai (si je savais le faire) 
retirer le mail que je viens d'envoyer à la caml-list à ce sujet
car je ne m'en était pas aperçu dans un premier temps ...

En fait le pipe bouclait à cause de l'erreur ! Je pense que s'il il n'y a
pas d'erreurs,   ocamlc -pp "sed ... | camlp4o -impl -" toto.ml marche
mais que dès qu'il y a une erreur qqchose bloque tout ...

Maintenant ce n'est pas forcement idiot de vouloir faire du sed (ou
autre-chose)
pour des choses simples puis du camlp4 ... donc ocamlc et ocamlopt devrait
peut-être 
accepter plusieurs optinos -pp ... si c'est simple à implanter avec qqchose de 
raisonnable pour les messages d'erreurs ...

En tout cas, pour mon cas particulier, je n'utilise plus que camlp4 avec un
fichier
config.ml généré par le Makefile.


-- 
Christophe Raffalli
Université de Savoie
Batiment Le Chablais, bureau 21
73376 Le Bourget-du-Lac Cedex

tél: (33) 4 79 75 81 03
fax: (33) 4 79 75 87 42
mail: Christophe.Raffalli@univ-savoie.fr
www: http://www.lama.univ-savoie.fr/~RAFFALLI
-------------------
Bug reports: http://caml.inria.fr/bin/caml-bugs  FAQ: http://caml.inria.fr/FAQ/
To unsubscribe, mail caml-list-request@inria.fr  Archives: http://caml.inria.fr

Re: [Caml-list] camlp4 + another preprocessor

[Daniel de Rauglaudre]

Hi,

On Thu, Nov 29, 2001 at 03:31:32AM +0100, Christophe Raffalli wrote:

> ça pourrait marcher avec camlp4o -impl - qui lit l'entré standard 
> (c'est pas documenté ? mais ça semble marcher ?)

Yes, it works, but the problem is neither camlp4 nor sed, but the pipe:
       sed ... | camlp4o foo.ml
       camlp4o ... | sed foo.ml

In these commands, the input file is not applied to the first command,
the pipe is unuseful. And I am not sure that it is a good idea to ask
ocaml to look at the string parameter. What it does is:
       Sys.command (the_pp_string_parameter ^ " " ^ the_input_file)

If you need sed just to define constants, it is possible with camlp4
alone. Well, I can add a system of macros like #define of C. Like that,
you would not need to program it, but just adding the good predefined
camlp4 cmo file. I am going to think of it.

-- 
Daniel de RAUGLAUDRE
daniel.de_rauglaudre@inria.fr
http://cristal.inria.fr/~ddr/
-------------------
Bug reports: http://caml.inria.fr/bin/caml-bugs  FAQ: http://caml.inria.fr/FAQ/
To unsubscribe, mail caml-list-request@inria.fr  Archives: http://caml.inria.fr