[Scilab-users] weird bug with scicv: can you help me reproduce it & report it correctly?

classic Classic list List threaded Threaded
3 messages Options
Antoine Monmayrant Antoine Monmayrant
Reply | Threaded
Open this post in threaded view
|

[Scilab-users] weird bug with scicv: can you help me reproduce it & report it correctly?

Hi all,

I'm facing a weird bug with scicv atoms in scilab6.0.1.
Under certain conditions, trying to access an image that has been
deleted crashes scilab.
Here is a script to reproduce the bug (almost straight from imread help
page):

//////////////////////////////
scicv_Init();
img = imread(getSampleImage("lena.jpg"));
matplot(img);
delete_Mat(img);
img(:,:);//scilab crashes if this is typed in the console. No problem if
ran from scinotes!
//////////////////////////////

The weird thing is that running this script from scinotes (ie F5 on it)
does not cause any issue.
But running it and then typing :
     img(:,:);
in scilab console crashes scilab.

Is this expected to get two different results when running from the
console or from scinotes?
Can anyone reproduce this bug?
(I am under linux Ubuntu 18.04).

Thanks in advance for your help,

Cheers,


Antoine

_______________________________________________
users mailing list
[hidden email]
http://lists.scilab.org/mailman/listinfo/users
Samuel GOUGEON Samuel GOUGEON
Reply | Threaded
Open this post in threaded view
|

Re: weird bug with scicv: can you help me reproduce it & report it correctly?

Hello Antoine,

I do not think that this behavior is proper to sciCV or any very
specific function.
It can be reproduced for many situations, provided that executing from
Scinotes or from the console is done without echo (= in silent mode).
Typically,
echoing a bunch of commands into the console quite often crashes Scilab.

Now, as already discussed in private for the impact on uman(), sciCV had
an issue : The current version still distributed on ATOMS includes some
read() and write() implementations that overwrites Scilab's ones
and -- at least for write() -- are not compatible with the native versions.
This yields some errors, but does not make uman crashing:
https://atoms.scilab.org/toolboxes/scicv/0.4#comment2883
Simon told me that this bug is fixed, but the new version is not
yet pushed on ATOMS.

Regards
Samuel

Le 11/04/2018 à 10:15, antoine monmayrant a écrit :

> Hi all,
>
> I'm facing a weird bug with scicv atoms in scilab6.0.1.
> Under certain conditions, trying to access an image that has been
> deleted crashes scilab.
> Here is a script to reproduce the bug (almost straight from imread
> help page):
>
> //////////////////////////////
> scicv_Init();
> img = imread(getSampleImage("lena.jpg"));
> matplot(img);
> delete_Mat(img);
> img(:,:);//scilab crashes if this is typed in the console. No problem
> if ran from scinotes!
> //////////////////////////////
>
> The weird thing is that running this script from scinotes (ie F5 on
> it) does not cause any issue.
> But running it and then typing :
>     img(:,:);
> in scilab console crashes scilab.
>
> Is this expected to get two different results when running from the
> console or from scinotes?
> Can anyone reproduce this bug?
> (I am under linux Ubuntu 18.04).
>
> Thanks in advance for your help,
>
> Cheers,
>
>
> Antoine
>
> _______________________________________________
> users mailing list
> [hidden email]
> http://lists.scilab.org/mailman/listinfo/users


_______________________________________________
users mailing list
[hidden email]
http://lists.scilab.org/mailman/listinfo/users
Samuel GOUGEON Samuel GOUGEON
Reply | Threaded
Open this post in threaded view
|

Re: weird bug with scicv: can you help me reproduce it & report it correctly?

Le 11/04/2018 à 11:09, Samuel Gougeon a écrit :
> Hello Antoine,
>
> I do not think that this behavior is proper to sciCV or any very
> specific function.
> It can be reproduced for many situations, provided that executing from
> Scinotes or from the console is done without echo (= in silent mode).

Sorry: as written just on the next line: ... provided that executing is
done WITH echo,
while executing without echo then most often passes.

> Typically, echoing a bunch of commands into the console quite often
> crashes Scilab.
_______________________________________________
users mailing list
[hidden email]
http://lists.scilab.org/mailman/listinfo/users