[Scilab-Dev] Strange bug? in CUTEr

[Clément David]

Hello Jean-Pierre,

This seems to be a bug. Can you report an issue on the forge (and with a
testcase) please ? 

--
Clément

Le vendredi 06 mars 2015 à 18:32 +0000, Jean-Pierre Dussault a écrit :
> Hi,
> 
> my script, which tests and compares optimization algorithms on a list
> of examples from the CUTEr toolbox, on one specific problem, crashes.
> In the terminal where scilab was started, the last message is:
>         INITW: The size of array IWK must be increased by at least
>         1213615
>          setcuterr
> 
> Here is the resulting strange behavior:
>         -->ls '/tmp/SCI_TMP_10021_oQNGWQ/NONMSQRT'
>          ans  =
>          
>         !RANGE.f      !
>         !             !
>         !OUTSDIF.d    !
>         !             !
>         !cleaner.sce  !
>         !             !
>         !ELFUN.f      !
>         !             !
>         !loader.sce   !
>         !             !
>         !EXTER.f      !
>         !             !
>         !GROUP.f      !
>         !             !
>         !OUTMESS      !
>         !             !
>         !librange.so  !
>          
>         -->ptst = '/tmp/SCI_TMP_10021_oQNGWQ/NONMSQRT'
>          ptst  =
>          
>          /tmp/SCI_TMP_10021_oQNGWQ/NONMSQRT   
>          
>         -->ls ptst
>          ans  =
>          
>              []
> Fore some strange reason, the path in a string variable ptst points to
> an empty location. The actual error message was:
>         -->[x0] = usetup(probpath+'/OUTSDIF.d'); 
>                                              !--error 998 
>         probpath                /OUTSDIF.d'); 
>          
>         -->probpath
>          probpath  =
>          
>          /tmp/SCI_TMP_10021_oQNGWQ/NONMSQRT   
> caused by the empty location.
> 
> Any idea on this strange problem?
> 
> Thx,
> 
> JPD
> 
> 
> _______________________________________________
> dev mailing list
> dev at lists.scilab.org
> http://lists.scilab.org/mailman/listinfo/dev