Running treetest with USE_GENERIC_TREECTRL set to 1 on Win11, the text contrast of highlighted entries is not good (white on light-blue):
Text color should be black like in native control (USE_GENERIC_TREECTRL is 0):
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you are subscribed to this thread.![]()
Should probably be simple enough to fix, but I wonder why do you prefer to use the generic version under MSW?
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you are subscribed to this thread.![]()
It's much faster when the native one when deleting and re-adding many entries. At least when it's invisible.
Native:
image.png (view on web)Generic:
image.png (view on web)—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you are subscribed to this thread.![]()
We almost certainly should call Freeze()/Thaw() before/after calling TreeView_DeleteAllItems(). But you should try using wxWindowUpdateLocker around the code recreating the items in the control in your code, this should help.
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you are subscribed to this thread.![]()
We use wxWindowUpdateLocker already: https://gitlab.com/kicad/code/kicad/-/blob/0cda13206400db4b3114812b3538a1ab3ba2021b/eeschema/net_navigator.cpp#L340
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you are subscribed to this thread.![]()
OK, sorry, I didn't know this. Unfortunately I don't have any other ideas about improving its speed immediately. You might want to open an issue about this separately because there are advantages in using the native control, notably accessibility (but it also looks more natively, of course).
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you are subscribed to this thread.![]()
Took me more time than planned because I've also refactored the sample to allow testing the generic control more easily (e.g. to compare the performance with it later), but is finally done in the linked PR.
Please test it if you can, it does fix the problem of this issue, but I'd like to double check that it doesn't introduce any new ones. TIA!
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you are subscribed to this thread.![]()
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you are subscribed to this thread.![]()