Already fixed

well gerrit has the patch in it seems so... let's call this resolved? at leats it has already been solved upstream a while back. :)

Pleace, look D4514

I did it: https://review.tizen.org/gerrit/#/c/106910/
@raster if you have any time, just review if I did everything right, and there are no leaks or anything critical inside
that's just backporting, but well... better recheck it as much as possible to make sure there won't be any unforeseen consequences

oh yeah.... the data ref. i did change something here... i think i dropped it entirely so no data is reffed or unreffed as it didnt make sense to me here to do it as this is during evas render and nothing should cause the object to be deleted... sure threads in the bg might continue but they would be working on non-eo data in parallel. so thats your issue. i refactored the need to ref and unref out of evas render there...

AAAAAAND a little update before I go home.
Added some bits of stuff I forgot to add, so here you go a really working patch both on -O0 and -O2

So it turned out as a working fix, with some BUT...

Fix to include only used resources.

oh yeah... i changed the phase1 code a lot... i got some performance improvements for :lots of objects, very few changing" overhead... ;]

So, fresh updates on rendering fun!!!

Thank you for your patch. It uncover one more issue. Could you please repeat the same flow with fast clicks on Cookies folder?