Hi !!

I recently updated to your newer version and wanted to ask 3 things. :)

1) Was there a reason you removed the global checking to stop a cursor being set over and over and constantly allocating a surface.

2)I will test this myself but thought I'd ask too, is it possible to just you sprite_get_texture, in the function window_set_cursor_sprite when ( os_browser != browser_not_a_browser ). That way I dont have to allocate a surface as an intermediate for window_set_cursor_surface_part. Like I said probably gonna atleast test myself :D

3) in window_set_cursor_surface_part, you have 2 lines

r = window_set_cursor_impl(hwnd, d, w, 0, 0, w, h, cx - l, cy - l);

r = window_set_cursor_impl(hwnd, b, sw, 0, 0, w, h, cx - l, cy - l);

both using cy - l, I was just curious if this was supposed to be cy - t

Anyway, TY for the tool again, I been using it in my project for a while tyty

1) I checked the revision history and uploaded versions, and there had not been any additional conditions (perhaps you modified your local copy?) - since there is no quick way to compare sprite state in GM, forcing caching would cause side effects if source content is modified. If you need to change cursors very often, you might instead want to rearrange scripts to pump out a cursor buffer and then take such buffer.

2) There is a source canvas somewhere inside the the HTML5 texture object, but it is a few layers deep. I usually avoid implementing extensive hacks in smaller extensions.

3) That would seem correct, I'll fix that for the next release

Ah ty for the replies YAL :]