You must remember the window on top does not necessarily, nor always meant to, have the input focus.
I myself often input to a window that is mostly obscured, coping from and modifying text, from the top window. I would hate being forced to have the top window always be the one with the focus. This is also one reason why I use 'follow mouse' without autoraise type of focus. I don't want to have to click in a window forcing it to the top, just to be able to type in it. On Mon, Jun 27, 2016 at 8:45 AM, Aaron Sloman <[email protected]> wrote: > > Another bit of the story has emerged. I don't know why I did not notice > this previously. It seems that Firefox is having difficulty working out > which of its windows is on top! > > I have now noticed that when a Firefox window is apparently rejecting > keyboard input the text sometimes goes into another window, though the > result may be invisible. It is only visible if the window that grabs focus > has a text input box e.g. a search box (opened by Ctrl-F in FF), or if I > had previously clicked in its address bar. > > I have not been able to detect a pattern behind the occurrence of this > 'theft' of focus. I did not notice it previously because with large numbers > of FF windows in a workspace (most of them minimised) I have to do an > explicit search to see which window has accepted the input and sometimes > there's no evidence because the focus was in the main part of the window, > which does nothing visible with text input. (There may have been a text > search triggered, but without leaving any visible record.) > > Sometimes the evidence is destroyed because I try typing in another window > which accepts input, but wasn't previously grabbing the text. > > I haven't found evidence that text is going to a non-current tab, but I've > never tried searching through all the tabs. > > Anyhow, it looks as if FF can tell which window is on top if I move in and > out of a workspace, or if I 'open' a short-cut button for editing, but can > get confused about which is on top if I minimise and maximise windows. > > My most recent experiment with a window (WinA) not accepting text is to > partially cover it wtih another window (WinB) and then click on WinA's > title bar to raise it. That always seems to put focus back in the raised > window as expected, when merely opening the window from a minimised state > does not. > > I have no experience writing window manager code, but I would have thought > that when all windows are rectangular it should be easy to determine > whether a window is on top of everything else or if the mouse is in a > portion of the window that is not covered by anything else. > > I have workarounds I can live with for now. > > Aaron > >
