Re: Interface between OCaml and Cduce

Hello Serge,

Serge Leblanc <serge.leblanc <at> wanadoo.fr> writes:

> Here the continuation,  I have just coded the save functions.

I'm sorry, I have not yet had the opportunity to test your code. It is
probably the worst attitude I could have for a free software
project. The end of november and december months are quite full for
me. At the very least, I'll look at it during december holidays.

My first feedback after reading your code:

 1. lines should be shorter than 80 characters;

 2. do you know about an emacs mode suitable to edit cduce code?

 3. I would like to stop at first error (rationale: it is not useful to
    load an apparently corrupted database. The XML file should not be
    written by hand but by a first saving of the bases) so raise an
    exception. However, for the error to be useful, I would like to
    provide XML file line and column number of the error (much like
    OCaml errors). Do you know if it is possible?

 4. in the definition of "type Xml_participant", it should be
    "Xml_group*" and not "Xml_group+": the groups are optional.

Let me know your opinion on these points.

Yours,

Re: Interface between OCaml and Cduce

Re: Interface between OCaml and Cduce

Re: Interface between OCaml and Cduce

Hello Serge,

Serge Leblanc <serge.leblanc <at> wanadoo.fr> writes:

> Here a patch.

As a general remark, I prefer much "diff -u" for patches, it makes them
more readable.

Thanks for the patch.

Yours,
d.
--

-- 
pub  1024D/A3AD7A2A 2004-10-03 David MENTRE <dmentre <at> linux-france.org>
 5996 CC46 4612 9CA4 3562  D7AC 6C67 9E96 A3AD 7A2A

Re: Re: Interface between OCaml and Cduce

Hello,

Serge Leblanc <serge.leblanc <at> wanadoo.fr> writes:

> 2 possible solutions: 
>
>      1. either the exceptions are caught and managed in the CDuce code, 
>      2. or the exceptions are transmitted and managed to the level of
>         the Ocaml code.
>
>
> Solution 1 is easily realizable.  Solution 2 raises today a little more
> difficulty, there is currently loss of information at the time of the
> passage of a CDuce exception to a Ocaml exception.  The detail of the
> exception unfortunately is left fallen. 
>
> I think that the next versions of Ocaml and CDuce will integrate
> improvements into this level because the Cristal project takes part now
> in the effort of research which CDuce requests.

Ok. If solution 1 can provides line numbers, I'm in favor of solution 1.

Maybe a third solution is possible :

 1. CDuce code raises an exception with precise location;

 2. CDuce code catches the exception and prints an error message. It
    then raises an OCaml exception;

 3. OCaml code catches the exception and properly closes the server.

What do you think of it?

Yours,
d.
--

-- 
pub  1024D/A3AD7A2A 2004-10-03 David MENTRE <dmentre <at> linux-france.org>
 5996 CC46 4612 9CA4 3562  D7AC 6C67 9E96 A3AD 7A2A

Re: Re: Interface between OCaml and Cduce

Re: [demexp-fr] Re: Licence de CDuce.

Re: Re: [demexp-fr] Re: Licence de CDuce.

(Follow-up)

On Fri, Dec 17, 2004 at 12:21:54AM +0100, lehobey wrote:
> Hi, (english for the -dev list)
> 
> On Thu, Dec 16, 2004 at 11:33:47PM +0100, Serge Leblanc wrote:
> > On Thu, 2004-12-16 at 23:06 +0100, David MENTRE wrote:
> 
> > > > Mme. Benzaken m'a confirmé l'avancement de ses travaux.
> 
> How far are they towards the relicensing of CDUCE code in a Debian
> compatible DFSG-free license?

  Actually the corresponding complete message in french (demexp-fr
mailing list) was saying they were to relicense their code with the
next release.  Eager to see it.  Thanks to Cduce team.

Experimentally,
Frédéric

Re: Re: [demexp-fr] Re: Licence de CDuce.

Hi, (english for the -dev list)

On Thu, Dec 16, 2004 at 11:33:47PM +0100, Serge Leblanc wrote:
> On Thu, 2004-12-16 at 23:06 +0100, David MENTRE wrote:

> > > Mme. Benzaken m'a confirmé l'avancement de ses travaux.

How far are they towards the relicensing of CDUCE code in a Debian
compatible DFSG-free license?

Experimentally,
Frédéric

Compilation report for Win32 platform (fwd)

It seems a .qmail-dev is missing on ras.eu.org, as demexp-dev <at> ras.eu.org 
fails (qmail daemon reply on request). So, hoping this one works...

I've previously sent the forwarded mail below to demexp-contact which 
obviously wasn't the proper choice, so here comes my report on the proper 
list this time. (I could have bounced the mail, but some explainations 
first seemed better to me)

Nota: glib 2.6 / GTK+ 2.6 and Pango 1.8 are out. They should be
binary-compatible with previous versions, so there may be some interesting
things ahead... It all depends on whether they took into account the fact 
that gtk+-2.4 fails to build on Cygwin.

Cordialy,
-- 
Christophe GISQUET
Hello! This is a signature virus! Please copy me into your .signature

---------- Forwarded message ----------
Date: Thu, 9 Dec 2004 22:42:13 +0100 (CET)
From: GISQUET Christophe <gisquet <at> inforezo.org>
To: demexp.contact <at> ras.eu.org
Subject: Compilation report for Win32 platform

Hello,

(please Cc: me as I'm not suscribed - even if traffic isn't that high, I know 
;)

As already discussed with D. MENTRE and other people in this project, I gave a 
shot at compiling DemExp under Windows. For this, I used both Cygwin and Msys 
environments, that mimic more or less accurately a GNU environment.

You can find a more or less detailed report for each environement in the 
attached file. The detail level fluctuates a lot, because at some point I 
thought it was almost useless to keep reporting. If those attachments get 
filtered out or some 56K'er flames me (sorry in advance), I can find a webhost.

I'm not sure in which direction I would continue, as no road seems successful. 
I can try precompiled win32 package for lablgtk, but I yet have to know how to 
get them recognized.

I hope you'll find the reports interesting (sorry for the text formatting and 
what not).

Still trying,
--

-- 
Christophe GISQUET
Hello! This is a signature virus! Please copy me into your .signature