This patch was originally posted on this Arch Linux forum thread: https://bbs.archlinux.org/viewtopic.php?id=176489 It was created against the combined wmii+ixp release found here: http://dl.suckless.org/wmii/wmii+ixp-3.9.2.tbz I'm having a hard time remembering the details of the bug this patch fixes now, but here is the e-mail I sent to Kris Maglione and Arch's wmii package maintainer if you're interested: > The problem comes down to the focusout() function in cmd/wmii/event.c. > It tests for the condition that the event's mode is NotifyGrab and > there is a KeyPress event later in the queue, and if both of these > fail then it proceeds to handle the FocusOut event, eventually > clearing disp.focus to c_magic in focusout_event() > (cmd/wmii/client.c). Then, when the KeyPress event finally does > arrive, client_focus() is unable to update the previously-focused > client's border with normcolors because disp.focus was already > cleared. > > My fix simply tweaks focusout()'s logic so that the else branch never > executes if ev->mode == NotifyGrab, regardless of whether there's a > KeyPress event in the queue yet. I don't know whether it should be > considered a bug in X that we received a FocusOut-NotifyGrab event > before the KeyPress event is available, but this change seems to work. Corey
…le for building on systems without LinoType's Palatino font.
…ation. Closes issue #228.
… politely on sigterm/sigint.