Anders Melander

I think @AngusJohnson can do that once he's ready to commit the file. For now it's more important to leave the repository in a consistent state.

> maybe separated *.lfm & *.lpr can be eliminated. That would be excellent. Thanks for the info.

The `>` vs `>=` doesn't matter. The result is the same. The missing check for an empty rect however is unfortunate since this alone means the behavior is different from...

I'm not sure I understand you. Do you mean that both with and without this PR the repaint optimizer doesn't work? Then what is the point of the PR? >...

> In my opinion, the repaint optimizer is fundamentally broken. I will create a new issue for that where I explain the defect. See #202

> Once more, in other words, when repaint optimizer ready to draw with partial areas (when issue #202 solved), without this PR, LCL GTK wouldn't retrieve the benefit at all....

> Is it good to just close it now? Yes I think so. The repaint logic has been fixed (in the **bugfix/repaint_optimizer_broken** branch) and once `IUpdateRectSupport` has been implemented for...

I _have_ looked at it but I just don't have time to comment on it right now. I'll try to find time this week-end. P.S. Don't @mention people unless there's...

First of all your code seems well written and is nicely formatted. It would have been interesting to have it benchmarked against the existing Graphics32 ellipse functions instead of GDI...

Re: benchmark: My bad. I misread your first comment. 1. I personally have no desire to be backward compatible with versions before XE but until we announce that the older...