[Scilab-users] graphics .uid : usage (& display)

[Samuel Gougeon]

Dear co-scilabers,

I am wondering about the hidden "uid" property of all graphical objects :
--> plot()
--> gcf().uid
  ans  =
    581.
--> gca().uid
  ans  =
    628.
--> gce().uid
  ans  =
    675.

It is a protected property, and it's OK:
--> gca().uid=500
Unknown property: uid.

*Question #1*: this property is not displayed. Is it on purpose?

*Question #2*: What can we do with this property? In particular, can we 
call directly some java routines with it, and then: How (example)?

I was digging around this uid, because i am searching for a robust -- 
and if possible built-in -- way to declare that a graphical component 
relies on another one. Examples :

  * a label of a contour line is "ascribed" to this line
  * a colorbar (that is an axes in its own) is related to a given other axes
  * etc.

Of course, it is possible to use the .tag property of the "parent" 
object (the line, the colored axes,..), and register it in the .userdata 
property of the "child" (the label, the colorbar..). And may be i will 
finally do things in this way.

But since there is this ".uid" that has the main advantage to be unique, 
with which we are sure that there was no id collision, i was wondering 
about it.

 From here to there, i was wondering about the robustness of these uid 
values when saving in a file a figure or a graphical component, and then 
reloading it 2 months later. It comes that .uid values ascribed to the 
reloaded objects are NOT the original ones. So, unfortunately, referring 
to uid to link objects is not robust...

So, if anyone knows

  * something about questions #1 and #2
  * how to avoid using the .tag to declare robust links between
    graphical objects

Hope reading you.

Best regards
Samuel


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.scilab.org/pipermail/users/attachments/20180213/60f158a2/attachment.htm>