[Scilab-Dev] Bug 2789 (debugger broken)

[Vincent COUVERT]

Hi François,

To be short and clear about the current development of Scilab and your 
ceaseless criticisms about the work of operational team.

 * About the rewriting of the TCL loop management: yes some 
functionalities have been broken, and particularly the functionalities 
used by Scipad debugger (reentrant Tcl constructs). These reentrant Tcl 
calls can be "quite easily" rewritten by separating calls to ScilabEval 
and calls to TCL_* as we did for the error handling for Scipad "Load 
into Scilab..." menu. We have to update the Wiki page you talk about to 
give workarounds for these special use cases of the TCL interface.

But please stop naming a person of the operational team in your mails! 
When a development is done, it is done after dialogue with the whole 
operational team or at least a part of it. So you can invoke the whole 
team responsibility but please stop front attacks!

 * About Scilab 5.0 BETA version and Release: we are late and we have 
some more developments and improvements to do before a BETA version can 
be released and by the way for Scilab 5.0 release.

 * About the bugs in Bugzilla: we cleaned the database during the last 
past weeks and we now have a good idea of all the remaining bugs and we 
try to attach the same importance to each bug and not only the ones you 
want us (operational team) to fix.

Do not hesitate to send us your encouragement instead of your 
criticisms, we are working hard here however we are still late...

Vincent

François Vogel a écrit :
> Hi all,
>
> Having just fixed two Scipad bugs tonight, I was wondering what you 
> operational team intend to do with bug 2789 (debugger broken in trunk)?
>
> I have seen that wiki page: http://wiki.scilab.org/Tcl_Thread
> and it explains what is the main problem. The new Tcl thread 
> management in Scilab 5 does not allow reentrant calls. And Scipad is 
> full of such reentrant Tcl constructs.
>
> Now, what? What's the next move?
>
> Is the Tcl thread management now finished in your opinion? (I didn't 
> see any activity on this subject for two weeks).
>
> Are you perhaps waiting for someone to rewrite the entire debugger in 
> order to work around the new limitation? On this the answer must be 
> no, since Bruno Jofret stated many times on this list that he didn't 
> want to break anything. But you're perhaps in a dead end?
>
> Or do you plan to allow for reentrant calls, so that the debugger will 
> work again with little or no change in Scipad?
>
> Given that Scilab 5 beta was supposed to happen in March, and that the 
> stable 5.0 is still scheduled in April, I'm getting a bit worried. 
> Could you please explain how you see the next future with respect to 
> that bug 2789 (and also 2596)?
>
> Thanks for your answers to come.
> Francois
>
>

-- 
==============================================
Vincent COUVERT
Centre de Recherche INRIA Paris-Rocquencourt
Domaine de Voluceau - B.P. 105
78153 Le Chesnay Cedex
==============================================
Equipe Projet SCILAB
Bâtiment 1B - Bureau 013
Email : vincent.couvert at inria.fr
Tél : +33 (0)1 39 63 54 46
Fax : +33 (0)1 39 63 55 94
==============================================