Well, I figured out the problem, and it's not present in the originial minimize.c, but instead in the Quinn patched minimize.c.  A } was moved below that else segment, thus moving which if statement it is attached to.  In Quinn's version that else statement attaches to the second evaluation instead (!mw->adjust).  So there was nothing wrong with the original, sorry for any confusion.  I'll post something about this on the <a href="http://compiz.net">compiz.net</a> messageboard. <div>On 7/10/06, David Reveman <<a href="mailto:davidr@novell.com">davidr@novell.com</a>> wrote: <blockquote class="gmail_quote" style="PADDING-LEFT: 1ex; MARGIN: 0px 0px 0px 0.8ex; BORDER-LEFT: #ccc 1px solid">On Mon, 2006-07-10 at 09:43 -0700, John Hess wrote: > Sorry, unfortunately it doesn't fix it.  The problem is that the > section you added it to doesn't execute if w->invisible is true, and > so if a window is on another viewport then the window never gets it's > scaling fixed. hm, it does execute if w->invisible is true as it executes when the following check fails: if (!w->invisible         &&    (ms->wMask & w->type) &&    minGetWindowIconGeometry (w, &mw->icon)) The only difference I see in your patch is that if w->invisible is true the moveWindow call that should position the window at it's correct position is not called and I fail to see why this would fix anything. I suspect that the problem is somewhere else. I don't see any artifacts when unminimizing window that are not on the current viewport. Does it happen to all applications for you? -David </blockquote></div> -- .: Animacide :.