Vasili Gulevich
Vasili Gulevich
Lazy Tree reveal is 340 times faster on the prototype branch for 100000 elements in the tree. The reval operation execution time is linear over the tree size. Other operations...
@laeubi @SyntevoAlex could you create a temporary branch for my changes, so that I could target my PRs without disrupting the master? In particular, I'd like to start with a...
I want performance test results to be a part of the multi-stage review process. I was not aware I could have my own actions. I was under impression they use...
@laeubi merging first PR with failing tests in the master branch will fail build for everyone.
I need to merge the first stage of change so that second stage can be reviewed independently with clear source delta and easily comparable test results. As the process is...
> Glancing at `gtk_tree_store_get_value()` I do not observe linear complexity. > It seems to derefence iterator and obtain the Tree's "row" in O(1). https://github.com/GNOME/gtk/blob/3.24.38/gtk/gtktreestore.c#L658 > Can you please clarify? This...
> Please use Java profiler on your test on original SWT code to identify what exactly is slow. So far I have an impression that you're acting on false theory...
# Workaround On SWT 3.129, GTK 3.24.41, using X11, screenshots taken with this method are always the same as the first taken screenshot. In other words, the state of screen...
Test failures are caused by https://github.com/eclipse-platform/eclipse.platform.swt/issues/1564
Converting to draft until an alternative of renderer reconfiguration instead of recreation is investigated.