NEWT Canvas Overlap on MacOSX with Multiple RCP perspectives

I have several perspectives in my Eclipse RCP application. Two perspectives using JOGL.

When I open a JOGL perspective with a NEWT view and then switch back to another perspective
(NEWTcanvas free - no view with a JOGL NEWT canvas) the NEWT canvas still overlaps the current views in the perspective and is not hidden (however this works on Windows and Linux!). In additon you can't resize the view (it's liek a free floating layer with the default dimensions of the JOGL perspective).

If I then open in this perspective the NEWT view or switch to the JOGL perspective the NEWT canvas gets reparented in it's view and is resizable again!

This is a strange behaviour only occuring on MacOSX (tested with Mojave).

Any thoughts what could cause this behaviour?

Hi Marcel,

just a general note on our bug reporting.

We need a way to better document these complicated issues.

Textual, we might need to write something like the following:

+++

JogAmp version:

Platforms working: (a) ,l;kdl (b) didih
Platforms exposing the issue: (1) skls (2) kjioudh

Expected behavior as occurring on (a) (b): bla bla.
1) open the window with a mouse click
2) do this
3) do that
4) RESULT OK

Erroneous behavior as occurring on (1) (2): blu bli
1) open the window with a mouse click
2) do this
3) do that
4) ERROR

+++

And sometimes, maybe we even need pictures later on.

After having so much time spend with the above detailed
description, adding a unit test should be feasible at this point,
IF this doesn't need like the whole Eclipse framework to be imported
to our unit tests ;-)

+++

Then this surely needs to be added to our bugzilla,
please ask for an account via personal email to me.

I had to disable 'self creating account' features in
bugzilla and wiki, because some spam bots polluted
the whole ecosystem sadly.

Thank you

~Sven

On 4/10/19 11:04 AM, Marcel [via jogamp] wrote:

Hello Sven,

yes, sorry for the hastily typed description. I will gladly do that.

E-Mail was sent.

On 4/11/19 9:02 AM, Marcel [via jogamp] wrote:
> Hello Sven,
>
> yes, sorry for the hastily typed description. I will gladly do that.

Please don't see it as a critic to you specifically.
Just a thought which occurred, how we could create
a nice template to describe these complicated issues.

Here is a similar difficult issue, Phil produced
a great report combined with debug logs:
https://jogamp.org/bugzilla//show_bug.cgi?id=1311

Here Christian reported one issue with the resolution
explained https://jogamp.org/bugzilla//show_bug.cgi?id=1358
He might have also just created a patch, but left it
open for me to choose :)

It surely is a science to properly communicate issues,
so the fix can be most efficient.

The balance of who spends time on which side,
reporting and/or fixing.
In the end it is a collaborative effort of course.

Once reporters were able to produce a unit test
w/ a pull request or git patch, we might even see
them producing the solution/fix ...

But I am just mumbling loud.
Happy that I have seen so many high quality bug reports
lately, which allowed me to produce a fix rather fast.
And yes, pull requests also have been sent.

In a way - on a relative small scale,
the open source model has been proven: Patches have been made.

To all of you: Thank you.

Cheers, Sven

Hello Sven I created the bug report:

https://jogamp.org/bugzilla/show_bug.cgi?id=1378