FW: Re: problems with "sync" execution, explanations and questions

[Enrico Segre]

I definitely think that not running this kind of discussion trough the mailing list is foul play as usual and incompetent development management. Do what you like, cover your asses closing bugs reports, whatever, please just leave me out of this game. I have better things to do. Enrico

-----Original Message-----
>From Vincent COUVERT <vincent.couvert at scilab.org>
Sent Thu 3/26/2009 12:44 PM
To fvogelnew1 at free.fr
Cc Bruno JOFRET <bruno.jofret at scilab.org>; Enrico Segre <enrico.segre at weizmann.ac.il>; Serge <serge.steer at inria.fr>; Sylvestre Ledru <sylvestre.ledru at scilab.org>
Subject Re: problems with "sync" execution, explanations and questions

Hi François,

More and more bugs are reported about Scilab crashes when the user uses 
"Load into Scilab" menu in Scipad.

We (Operational Team) are going to implement the solution proposed by 
Bruno in the email below and modify Scipad sources to make "Load into 
Scilab" work. These modifications will be done in 5.1 branch and then 
merged in master. We will then release a Scilab 5.1.1 version as soon as 
possible.

The aim of these modifications is to make Scilab more stable and avoid 
crashes when using "Load into Scilab" menu. We know that they will not 
make Scipad debugger work.

Best regards,
Vincent

fvogelnew1 at free.fr a écrit :
> Good morning Vincent,
>
> Selon Vincent COUVERT <vincent.couvert at scilab.org>:
>   
>> Of course we are really serious about fixing these Tcl-related problems.
>> Many bugs have been reported about Scilab freezes when using the "Load
>> into Scilab" menu. All these bugs can be fixed with the solution
>> proposed here by Bruno, Serge & al. We are aware that this solution
>> breaks some Scipad functionalities
>>     
>
> Sorry, which Scipad functionalities would be newly broken by your solution? What
> are you talking about?
>
>   
>> and does not make Scipad debugger
>> work. However this solution increases Scilab/Scipad interaction
>> stability and can be a great improvement for Scilab 5.1.* versions.
>>     
>
> Yes, perhaps. You could certainly make "Load into Scilab" work. If you say the
> "torpedo" solution is good enough for this, I trust you but I didn't check in
> depth. What I told you in past emails is that this solution is making ScilabEval
> sync no longer synchronous, which is not consistent with the help file, but we
> already know this from the broken debugger lesson. Up to you to add another
> backward incompatibility.
>
> So now I understand. Your goal is NOT to fix this broken Tcl interface /
> ScilabEval design. It is NEITHER to have a working debugger again.
> You just want to fix the bugs related to Tcl that make the users complain a bit
> too loud for your taste (and as a matter of fact I agree with your taste on
> this). And patching this will buy you the time you need to write a new editor in
> java.
>
> So your approach is to patch the patches over the existing design. Well, that's
> just fine for me.
>
> That is your choice and it can even be understood as a logical and reasonable
> choice given that you want to throw Scipad away completely in the same bin as
> Tcl: shut down users complaints by piling yet another patch, and in the meantime
> write from scratch a full java editor that will mend your life and make everyone
> happier. Fine for me.
>
> As already stated, all this plan will be without me. I trink you are making a
> big mistake with all your java, and you certainly know all this but perhaps
> there was not yet enough complaints about java crashes in 5.x graphics. You are
> stuck with your java choice made by others, I know that's a sad thing. What I
> definitely do not understand is why you stubbornly want to repeat this same
> mistake by removing Tcl and Tk. This is really a mystery to me but as I'm
> writing this email I have to say that the answer matters less and less to me.
>
>   
>> We have planned to write a Java editor for Scilab 5.2 (specifications
>> are not yet written).
>>     
>
> Wow, you will write real specs? Or do yo mean just a SEP?
>
>   
>> However, it would be great to have Scipad still
>> working even if we have an other editor for Scilab
>> and your experience
>> as Scipad developer could help use to work faster...
>>     
>
> You're kidding, right? Tell me you are not expecting me to join your plan for
> what I think will be a failure?
>
> Francois
>   


-- 
==============================================
Vincent COUVERT
Responsable du développement
                    ------                    
Fondation DIGITEO - Consortium Scilab
Domaine de Voluceau - B.P. 105
78153 Le Chesnay Cedex
Bâtiment 23 - Bureau 22
                    ------                    
Email : vincent.couvert at scilab.org
Tél : +33 (0)1 39 63 54 46
Fax : +33 (0)1 39 63 55 94
==============================================