Panel Control: Panel is displayed after deactivate

[jorix]

This happens when all panel's controls are deactivate and panel is deactivated or destroyed.

The patch includes a test. It work well in: FF36, IE8, Chrome6 & Safari5.

[ahocevar]

Without having tried, but what part of your patch fixes the deactivate behavior? It only changes destroy? Or do the new deactivate tests also pass without the patch applied?

[jorix]

Replying to ahocevar:

Without having tried, but what part of your patch fixes the deactivate behavior?

this.redraw(); this line in desactivate method

[ahocevar]

Good catch. Thanks for the patch.

[ahocevar]

r10732

[ahocevar]

We had to partially revert r10732 with r10834. The deactivate sequence is still fixed, but the destroy sequence should be fixed as part of #2210.

[jorix]

Replying to ahocevar:

We had to partially revert r10732 with r10834. The deactivate sequence is still fixed, but the destroy sequence should be fixed as part of #2210.

Here we have crossed several tickets: r10813 changed the test that failed due to #2850 due to an oversight of mine in #2835 where he said:

  var div = panel.div; 
  panel.destroy();
  t.ok(panel.div.innerHTML == "".

had to say:

  var div = panel.div; 
  panel.destroy();
  t.ok(div.innerHTML == "",

but r10732 put:

  var div = panel.div; 
  panel.destroy();
  t.eq(panel.div, null,

is not the same (this is a test that #2850 is fixed).

But r10790 undo fix #2850, and test changet by r10813 failed.

this looks like a scene from the Marx Brothers.

#2850, is ok (and test ... div.innerHTML == "", ... works with this.deactivate() in panel destroy, with or without #2850).
The original patch from #2864 and mine working properly with the right test, but they are different views of the problem (I thought of something like #2210)
And #2210 is ok.

Uff!

Happy ending? :-))

[jorix]

Replying to jorix:

...
but r10732 put:
...

Errata:

the right is:

...
but r10813 put:
...