gantt 5.5.2, memory leak again on moving tasks

[nickolay]

What happens during "reschedule" operation is that during the initial "reschedule" extra memory is allocated, however further "reschedules" do not do that (all extra allocated memory is released). Please see the video in the github ticket.

To investigate what is allocated during the initial "reschedule", please reduce the amount of data (halving it would be ideal), because taking a heap snapshot after "reschedule" just crashes the browser.

[pincherhgz]

if you alter between reschedule and reschedule forward you see that on every action it increases that memory. the only thing what is done at each of these (alternating!) is that alle the tasks/operations under the object that is rescheduled are set first to inactive.
Saki already analyzed that and we did not find anything wrong in the code, see the entry point of reschedule in the screenshot

[nickolay]

Same thing with alternating between "reschedule" and "rescheduleForward" - initial "rescheduleForward" allocates some extra memory (same as initial "reschedule"), further "rescheduleForward" stays at the same memory usage. See the video in the github ticket.

Any other steps to reproduce the memory leak?

[pincherhgz]

could it be possible that this issue was created using the Sencha framework, when I try it now it is exactly as you describe it. I'll go for more checks and feedback

[nickolay]

Ok, waiting for your feedback.

[pincherhgz]

checked with 5.5.5. seems to be ok, thanks

[mats]

Thanks for confirming!