[Scilab-users] ?==?utf-8?q? Replacing predef() with an actual varprot(): a top-5 priority for Scilab 6.1?==?utf-8?q? <= At last, protecting user variables one by one

Antoine Monmayrant amonmayr at laas.fr
Thu Jun 8 21:08:52 CEST 2017


Le Jeudi, Juin 08, 2017 19:51 CEST, Samuel Gougeon <sgougeon at free.fr> a écrit: 
 
> Hello,
> 
> Scilab 6.0 is released. With respect to previous major releases, the 
> storage and management of variables have been completely changed. This 
> was one of the major purposes for this release.
> 
> Yet, after 30 years of development of this high level software, we are 
> still unable to protect any user variable against clearing, and there 
> are no news on this side. There are many oitlying developments, whereas 
> this central feature is still unavailable. The pseudo function predef() 
> is still here, alone, never ever actually usable.
> 
> Indeed, predef() is unable to protect variables one by one, at any 
> moment after the session startup. As its name states it, predef() has 
> been designed exclusively to protect *predef*ined variables %i, %pi, etc 
> during the startup process. It is a startup *internal*, and we could 
> wonder why it has been documented.
> 
> One of the most disturbing and buggy consequences of this missing 
> feature is that it is still  impossible to protect a library that has 
> been loaded during the session, i.e. not autoloaded at startup. Any 
> /clear/ instruction -- so loved by many former Matlabers in their 
> scripts -- kills all such libraries.
> 
> The need for a true customizable function to protect/unprotect variables 
> was reported as soon as 2004 -- and likely even before --, so 13 years 
> ago. For the only bugzilla reports, please see for instance:
> 
> http://bugzilla.scilab.org/686
> http://bugzilla.scilab.org/8634
> http://bugzilla.scilab.org/10988
> 
> Scilab has still some central weaknesses like this one.
> IMO, implementing varprot() should be a top priority.
> 
> Don't you need this feature as well?

Oh, yes I do, like many of my colleagues!

Antoine

> 
> Best regards
> 
> Samuel Gougeon
>




More information about the users mailing list