[Scilab-users] xclick bug on Mac OS ?

classic Classic list List threaded Threaded
16 messages Options
Philippe Roux Philippe Roux
Reply | Threaded
Open this post in threaded view
|

[Scilab-users] xclick bug on Mac OS ?

Hi  to all,

Today I saw a strange bug with xclick and scilab 6.0.0 on Mac OS :
there is a shift between the position in the graphic window and
coordinates  recovered by xclick (see the screen-shot with version
information : https://photos.app.goo.gl/wywgRFAH5wSdszGa2 ). I can't
reproduce the bug cause I don't have a Mac, it was generated by a piece
of scilab code like :

    clf()
    plot2d(0,0,rect=[0 0 3 3])
    xgrid(1)
    x=[];y=[]
    ibutton=0
    while ibutton<>5
        [ibutton,xcoord,ycoord]=xclick()
        x($+1)=xcoord
        y($+1)=ycoord
        plot(x,y,'-o')
    end

I don't find something similar in the bugzilla list :

http://bugzilla.scilab.org/buglist.cgi?quicksearch=xclick&list_id=43893

I would appreciate if someone could confirm (or not) the bug  on
MacOS+scilab6 , and report on bugzilla if necessary.

Best regards,

Philippe

_______________________________________________
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: xclick bug on Mac OS ?

Hello Philippe,

Sorry, but i don't catch clearly the circumstances of this bug.
Are you the user that experimented this bug, or someone else did?
If you are not the user, where did you find the primary information?
If you are the user, why do you say that "you can't reproduce it because
you don't have a Mac" while you experimented it on a Mac? Would you
mean that the bug is a transient not fully reproducible one?

Unfortunately, the screenshot does not show the bug.

When providing the piece of "code like..", i am afraid that the "like" is risky.
So, is it actually this code that yielded the bug, or you don't know?

Such a bug was already met to set a legend: http://bugzilla.scilab.org/13592
So , it could be the same for xclick, whether it it used in a figure homed in a uicontrol frame.

In the piece of code that you provide, we don't know how the homing figure is generated.

Version information may be useful, but less than a minimal self-sufficient copy/pastable
piece of code that, in a fresh scilab session, yields the bug.

Thanks

Best regards
Samuel


Le 27/10/2017 à 23:16, philippe a écrit :
Hi  to all,

Today I saw a strange bug with xclick and scilab 6.0.0 on Mac OS :
there is a shift between the position in the graphic window and
coordinates  recovered by xclick (see the screen-shot with version
information : https://photos.app.goo.gl/wywgRFAH5wSdszGa2 ). I can't
reproduce the bug cause I don't have a Mac, it was generated by a piece
of scilab code like :

    clf()
    plot2d(0,0,rect=[0 0 3 3])
    xgrid(1)
    x=[];y=[]
    ibutton=0
    while ibutton<>5
        [ibutton,xcoord,ycoord]=xclick()
        x($+1)=xcoord
        y($+1)=ycoord
        plot(x,y,'-o')
    end

I don't find something similar in the bugzilla list :

http://bugzilla.scilab.org/buglist.cgi?quicksearch=xclick&list_id=43893

I would appreciate if someone could confirm (or not) the bug  on
MacOS+scilab6 , and report on bugzilla if necessary.

Best regards,

Philippe

_______________________________________________
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: xclick bug on Mac OS ?

Le 28/10/2017 à 12:22, Samuel Gougeon a écrit :
Hello Philippe,

Sorry, but i don't catch clearly the circumstances of this bug.
Are you the user that experimented this bug, or someone else did?
If you are not the user, where did you find the primary information?
If you are the user, why do you say that "you can't reproduce it because
you don't have a Mac" while you experimented it on a Mac? Would you
mean that the bug is a transient not fully reproducible one?

Unfortunately, the screenshot does not show the bug.

When providing the piece of "code like..", i am afraid that the "like" is risky.
So, is it actually this code that yielded the bug, or you don't know?

Such a bug was already met to set a legend: http://bugzilla.scilab.org/13592
So , it could be the same for xclick, whether it it used in a figure homed in a uicontrol frame.

In the piece of code that you provide, we don't know how the homing figure is generated.

Sorry: the clf() tells it (if it was actually run): it was in a (almost) fresh figure, not in a frame.

The clf page https://help.scilab.org/docs/6.0.0/en_US/clf.html misses telling that
uicontrols are cleared as well, including frames (i guess).

Samuel


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

Re: xclick bug on Mac OS ?

In reply to this post by Samuel GOUGEON
Hello Samuel,


Le 28/10/2017 à 12:22, Samuel Gougeon a écrit :
> Hello Philippe,
>
> Sorry, but i don't catch clearly the circumstances of this bug.
> Are you the user that experimented this bug, or someone else did?
> If you are not the user, where did you find the primary information?

the user was one of my student during a course :-)


> If you are the user, why do you say that "you can't reproduce it because
> you don't have a Mac" while you experimented it on a Mac? Would you
> mean that the bug is a transient not fully reproducible one?

that's it : the bug appeared only with Mac OS X  and scilab 6.0.0


>
> Unfortunately, the screenshot does not show the bug.

It can't, the user click in the graphics windows , xclick retrieves the
position, and there is a plot of this position which is very different
from the position of the mousse.

>
> When providing the piece of "code like..", i am afraid that the "like"
> is risky.
> So, is it actually/this/ code that yielded the bug, or you don't know?

yes it is this part of the code (which does not appear in the screen
shot because it is embedded in one of the function call.

>
> Such a bug was already met to set a legend:
> http://bugzilla.scilab.org/13592 <http://bugzilla.scilab.org/13592>
> So , it could be the same for xclick, whether it it used in a figure
> /homed in a uicontrol frame/.

yes it looks close to what happens with xclick  and legend with optional
argument "5" puts the legends at mousse position so both bugs should be
related.
>
> Version information may be useful, but less than a minimal
> self-sufficient copy/pastable
> piece of code that, in a fresh scilab session, yields the bug.

the code below should be the minimal example :


clf; plot2d(0,0,rect=[0,0,1,1]);xgrid(1);// set graphic windows
[ibutton,x,y]=xclick()// +  click in graphic window
plot(x,y,'xr') // ->the position differs from the mousse position


Best regards,

Philippe


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

Re: xclick bug on Mac OS ?

In reply to this post by Samuel GOUGEON
Le 28/10/2017 à 12:45, Samuel Gougeon a écrit :
> [...]
> Sorry: the clf() tells it (if it was /actually/ run): it was in a
> (almost) fresh figure, not in a frame.

yes , the figure is cleared at begining,  but the graphic windows can be
a new one (if no graphic window is already open) or an existing one
(perhaps already docked in scilab main window). Those different
configurations where tested by the student .

>
> The clf page https://help.scilab.org/docs/6.0.0/en_US/clf.html misses
> telling that uicontrols are cleared as well, including frames (i guess).

I don't understand the link between this bug and uicontrols .

Philippe

_______________________________________________
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: xclick bug on Mac OS ?

Le 28/10/2017 à 13:48, philippe a écrit :
The clf page https://help.scilab.org/docs/6.0.0/en_US/clf.html misses
telling that uicontrols are cleared as well, including frames (i guess).
I don't understand the link between this bug and uicontrols .

If you are sure that the student did not define any uicontrol("style","frame") to plot in it,
indeed, there is no link.

uicontrol frames have their own system of coordinates, a bit like "a figure in a figure".
This is what made the shift between the legend handle and the mouse, when setting
the legend position interactively in a frame.


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

Re: xclick bug on Mac OS ?

In reply to this post by Philippe Roux
Hi Philippe and Samuel,

  I experiment exactly the problem described by Philippe on my Mac (MacBook Pro Retina 13 pouces running MacOS Sierra 10.12.6) with Scilab 6.0
I wrote to the mailing list on the 10 mai 2017 about that :

—— begin quote
Hi,

My name is Philippe, I work in the field of physics and signal processing, I use scilab a lot in my work and I want to thanks very much all the the people involve in the development of this great software. 

 I updated my scilab to 6.0.0 two weeks ago on my power book (retina) running macOS 10.12.4. Since this time I experiment two problems with graphic figures. The first issue is that the thickness of the lines is really small and the text for the axis labels, curves label and title are almost unreadable (way smaller than in the preceding version). The second problem is about the zoom, when I try to zoom on a graphic figure using the rectangular selection with the mouse, I get an offset between the mouse pointer and the actual selection, making it impossible to correctly define de rectangular region. 

Is anybody else experimenting the same issues ? Does somebody have a fix or a workaround for that ?
— — end quote

So i’m glad to know that I’m not the only one who experiment this problem. In order to try to help I ran the following piece of code

clf()
   plot2d(0,0,rect=[0 0 1 1])
   xgrid(1)
   x=[];y=[]
   ibutton=0
   while ibutton<>5
       [ibutton,xcoord,ycoord]=xclick()
       x($+1)=xcoord
       y($+1)=ycoord
       plot(x,y,'-o')
   end
(thank you Philippe) on my Mac and clic with the mouse on the point (1,0), (0.9,0.1), (0.8,0.2), (0.7,0.3), (0.6,0.4) and (0.5,0.5) which result to plot circles on points :
x  = 

   0.4169399
   0.3666667
   0.3163934
   0.2661202
   0.215847
   0.1677596
   0.1677596
 y  = 

   0.5811594
   0.6318841
   0.6782609
   0.7289855
   0.7768116
   0.8304348
   0.8304348

you can download the png of the figure here : https://we.tl/VXmrnlsI2V


I take the opportunity to report again that this problem affect the interactive zooming in graphic windows and, I thing, is related to the fact that the graphics windows (with default settings) are almost unreadable on My MacBook Retina (very small window with thin lines and fonts). Those issues where new with version 6.0 I didn’t have those problems with the 5.5.2 version.

I hope it helps.

One more time thanks a lot to all the people working on the development of Scilab, it is a very powerful software that I use a lot.

Philippe.

  

Le 28 oct. 2017 à 13:48, philippe <[hidden email]> a écrit :

Le 28/10/2017 à 12:45, Samuel Gougeon a écrit :
[...]
Sorry: the clf() tells it (if it was /actually/ run): it was in a
(almost) fresh figure, not in a frame.

yes , the figure is cleared at begining,  but the graphic windows can be
a new one (if no graphic window is already open) or an existing one
(perhaps already docked in scilab main window). Those different
configurations where tested by the student .


The clf page https://help.scilab.org/docs/6.0.0/en_US/clf.html misses
telling that uicontrols are cleared as well, including frames (i guess).

I don't understand the link between this bug and uicontrols .

Philippe

_______________________________________________
users mailing list
[hidden email]
http://lists.scilab.org/mailman/listinfo/users


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

Re: xclick bug on Mac OS ?

Le 28/10/2017 à 14:26, Philippe Dérogis a écrit :
> Hi Philippe and Samuel,
>
>   I experiment exactly the problem described by Philippe on my Mac
> (MacBook Pro Retina 13 pouces running MacOS Sierra 10.12.6) with Scilab 6.0
> I wrote to the mailing list on the 10 mai 2017 about that :

thanks for your feedback, so it is actually a bug that affect only Mac
OS with scilab 6.0.0 . It would be nice if you report the bug on
bugzilla ( http://bugzilla.scilab.org/ ) you need to create an account
to report bugs. If you don't want to create an account tell me, I will
report the bug , but it would be better than it comes from a Mac OS user.


Best regards,

Philippe

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

Re: xclick bug on Mac OS ?

In reply to this post by Samuel GOUGEON
Le 28/10/2017 à 14:00, Samuel Gougeon a écrit :
>
> If you are /sure/ that the student did not define any
> uicontrol("style","frame") to plot in it,
> indeed, there is no link.

by the infinite monkey theorem* it would be possible that a student
create a uincontrol from scratch but the probability of this event is
close to 0 ;-) but the bug has been confirmed  by Philippe Derogis ...
>
> uicontrol frames have their own system of coordinates, a bit like "a
> figure in a figure".
> This is what made the shift between the legend handle and the mouse,
> when setting  the legend position interactively in a frame.

well the problem I saw was very close to the one described for legend.

Best regards,

Philippe


*https://en.wikipedia.org/wiki/Infinite_monkey_theorem

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

Re: xclick bug on Mac OS ?

In reply to this post by Philippe Roux
        Hi Samuel,

  Yes It should be nice if you could register the bug in Scilab bug tracking system.

Sincerely,

Philippe.

> Le 28 oct. 2017 à 22:44, philippe <[hidden email]> a écrit :
>
> Le 28/10/2017 à 14:26, Philippe Dérogis a écrit :
>> Hi Philippe and Samuel,
>>
>>   I experiment exactly the problem described by Philippe on my Mac
>> (MacBook Pro Retina 13 pouces running MacOS Sierra 10.12.6) with Scilab 6.0
>> I wrote to the mailing list on the 10 mai 2017 about that :
>
> thanks for your feedback, so it is actually a bug that affect only Mac
> OS with scilab 6.0.0 . It would be nice if you report the bug on
> bugzilla ( http://bugzilla.scilab.org/ ) you need to create an account
> to report bugs. If you don't want to create an account tell me, I will
> report the bug , but it would be better than it comes from a Mac OS user.
>
>
> Best regards,
>
> Philippe
>
> _______________________________________________
> users mailing list
> [hidden email]
> http://lists.scilab.org/mailman/listinfo/users

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

Re: xclick bug on Mac OS ?

Le 29/10/2017 à 13:16, Philippe Dérogis a écrit :
>
>
>   Yes It should be nice if you could register the bug in Scilab bug tracking system.

done : http://bugzilla.scilab.org/show_bug.cgi?id=15316

Best regards,

Philippe

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

Re: xclick bug on Mac OS ?

Le 29/10/2017 à 16:00, philippe a écrit :
> Le 29/10/2017 à 13:16, Philippe Dérogis a écrit :
>>
>>
>>   Yes It should be nice if you could register the bug in Scilab bug tracking system.
>
> done : http://bugzilla.scilab.org/show_bug.cgi?id=15316
>


good new the bug status has changed to "resolved fixed" ! Now you can
see how it is important to report bugs in bugzilla, and not only on
mailing list. Every one can do it, of course you have to give a minimal
and complete example to reproduce the bug (it is not so easy for GUI
problems). I've started doing this 10 years ago, most of the bugs I
reported have been solved, some of them quickly, other required to wait
for a scilab release. Even reporting duplicates of already existing bugs
can help to improve scilab ...

best regards,

Philippe

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

 Paul BIGNIER changed bug 15316

---------- Bug Summary -----------
shift between mouse position and xclick() coordinnates


What      Removed Added
Status      NEW RESOLVED
CC          [hidden email]
Resolution --- FIXED

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

_______________________________________________
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: xclick bug on Mac OS ?

Hello Philippe,

Le 31/10/2017 à 10:58, philippe a écrit :
> Even reporting duplicates of already existing bugs
> can help to improve scilab ...

Sorry, but i strongly disagree with this last assertion.
Please contribute to process bugs reported on bugzilla, and you will
understand why.
On the contrary, your piece of advices is the best way to sink reports.
This is somewhat spamming bugzilla.
By the way, this is just a basic rule in net tools like forums to search
for anteriority
before addressing an item.

Bugzilla has a builtin vote system. This one is disabled. This setting
has already been addressed.
This is a choice from the dev team, not a default.
But if anyone just want to stress on the fact that a bug is really
handicaping for him/her,
it is always possible to add a comment to an existing report, or just
add "+1" in comment
as a vote for fixing it.

A more extensive rational against spamming bugzilla with duplicates is
available here:
http://bugzilla.scilab.org/show_bug.cgi?id=15031#c3

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

Re: xclick bug on Mac OS ?

        Happy to know that the bug is fixed, thanks you for that. I will check the Nighty Builds.

Regards,

Philippe.


> Le 31 oct. 2017 à 11:23, Samuel Gougeon <[hidden email]> a écrit :
>
> Hello Philippe,
>
> Le 31/10/2017 à 10:58, philippe a écrit :
>> Even reporting duplicates of already existing bugs
>> can help to improve scilab ...
>
> Sorry, but i strongly disagree with this last assertion.
> Please contribute to process bugs reported on bugzilla, and you will understand why.
> On the contrary, your piece of advices is the best way to sink reports.
> This is somewhat spamming bugzilla.
> By the way, this is just a basic rule in net tools like forums to search for anteriority
> before addressing an item.
>
> Bugzilla has a builtin vote system. This one is disabled. This setting has already been addressed.
> This is a choice from the dev team, not a default.
> But if anyone just want to stress on the fact that a bug is really handicaping for him/her,
> it is always possible to add a comment to an existing report, or just add "+1" in comment
> as a vote for fixing it.
>
> A more extensive rational against spamming bugzilla with duplicates is available here:
> http://bugzilla.scilab.org/show_bug.cgi?id=15031#c3
>
> Regards
> Samuel
> _______________________________________________
> users mailing list
> [hidden email]
> http://lists.scilab.org/mailman/listinfo/users

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

Re: xclick bug on Mac OS ?

In reply to this post by Samuel GOUGEON
Le 31/10/2017 à 11:23, Samuel Gougeon a écrit :
> Hello Philippe,
>
> Le 31/10/2017 à 10:58, philippe a écrit :
>> Even reporting duplicates of already existing bugs
>> can help to improve scilab ...
>
> Sorry, but i strongly disagree with this last assertion.

probably this last assertion was blundering :-( I want to encourage
users reporting bugs but not to spam of course. Unfortunately for
beginners it is hard  to clearly identify if a bug is already reported
or not (I made a lot of duplicates while starting reporting bugs) so we
have to accept that new users  make such errors if we want that more
persons get involved.


> [...]
> But if anyone just want to stress on the fact that a bug is really
> handicaping for him/her,it is always possible to add a comment to
> an existing report, or just add "+1" in comment as a vote for fixing it.

yes, I absolutely agree with this.

>
> A more extensive rational against spamming bugzilla with duplicates is
> available here:
> http://bugzilla.scilab.org/show_bug.cgi?id=15031#c3

I understand your point of view, I know you do a lot of work so seeing
the amount of unresolved bugs can be discouraging . But from my user
point of view the system doesn't work so bad : I see scilab is improving
year after years.

have a nice day!

Philippe

_______________________________________________
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: xclick bug on Mac OS ?

Le 31/10/2017 à 13:53, philippe a écrit :

> Le 31/10/2017 à 11:23, Samuel Gougeon a écrit :
>> Hello Philippe,
>>
>> Le 31/10/2017 à 10:58, philippe a écrit :
>>> Even reporting duplicates of already existing bugs
>>> can help to improve scilab ...
>> Sorry, but i strongly disagree with this last assertion.
> probably this last assertion was blundering :-( I want to encourage
> users reporting bugs but not to spam of course. Unfortunately for
> beginners it is hard  to clearly identify if a bug is already reported
> or not (I made a lot of duplicates while starting reporting bugs) so we
> have to accept that new users  make such errors if we want that more
> persons get involved.

The search engine of bugzilla is powerful and easy to use. It is normal
to expect from any user
to take care when targeting and describing things. Moreover, it is even
easier and gets trivial
to get the list of bugs already reported about an item from Scilab 5.5
and 6 with
--> uman <item> wb

The point that slows down improving Scilab is NOT missing reports. It is
missing contributions.
The first elemental contribution is to take care when reporting bugs.
When it is not easy to know, the first step is to post questions on the
mailing list, to be
sure that the issue is not in one's own code rather than in Scilab.
Bugzilla is not a forum.

An easy way to get involved is to answer to bugs, possibly in the day,
rather than
asking to the poster "Hey, could you post a minimal test case?".. 6
months after he/she posted.
It would be actually very great to increase this kind of involvement!
But still, Bugzilla is not a forum.

When encouraging to post on bugzilla without really encouraging fixing
bugs or
answering to posters mainly increases frustrations of posters, or/and
enforces the
feeling that we can get something for free without doing  anything else
than vaguely
complaining. This is just the opposite of  involvement.
About frustration, please see the relevant post from Ross, to which i
answered in link.

So yes, thanks to every user to get involved!

Best regards
Samuel

_______________________________________________
users mailing list
[hidden email]
http://lists.scilab.org/mailman/listinfo/users