I can confirm this bug locally.
Looking at the D-Bus log, I see:
method call sender=:1.41 -> dest=org.ayatana.bamf serial=480 path=/org/ayatana/bamf/window60817505; interface=org.ayatana.bamf.window; member=Xprop string "_GTK_UNIQUE_BUS_NAME" method return sender=:1.30 -> dest=:1.41 reply_serial=480 string ""
...
method call sender=:1.41 -> dest=org.ayatana.bamf serial=461 path=/org/ayatana/bamf/window60817505; interface=org.ayatana.bamf.window; member=Xprop string "_GTK_UNIQUE_BUS_NAME" method return sender=:1.30 -> dest=:1.41 reply_serial=461 string ""
method call sender=:1.38 -> dest=org.ayatana.bamf serial=524 path=/org/ayatana/bamf/window60817505; interface=org.ayatana.bamf.window; member=Xprop string "_GTK_UNIQUE_BUS_NAME" method return sender=:1.30 -> dest=:1.38 reply_serial=524 string ""
in essence: bamf is reporting that the _GTK_UNIQUE_BUS_NAME property is not set on the window when it is first presented.
This means that either
a) libreoffice is not setting the property before mapping the window (libreoffice bug) b) bamf is telling lies (bamf bug)
Pretty sure we can rule out the chances of this being in any other component...
I can confirm this bug locally.
Looking at the D-Bus log, I see:
method call sender=:1.41 -> dest=org. ayatana. bamf serial=480 path=/org/ ayatana/ bamf/window6081 7505; interface= org.ayatana. bamf.window; member=Xprop BUS_NAME"
string "_GTK_UNIQUE_
method return sender=:1.30 -> dest=:1.41 reply_serial=480
string ""
...
method call sender=:1.41 -> dest=org. ayatana. bamf serial=461 path=/org/ ayatana/ bamf/window6081 7505; interface= org.ayatana. bamf.window; member=Xprop BUS_NAME"
string "_GTK_UNIQUE_
method return sender=:1.30 -> dest=:1.41 reply_serial=461
string ""
...
method call sender=:1.38 -> dest=org. ayatana. bamf serial=524 path=/org/ ayatana/ bamf/window6081 7505; interface= org.ayatana. bamf.window; member=Xprop BUS_NAME"
string "_GTK_UNIQUE_
method return sender=:1.30 -> dest=:1.38 reply_serial=524
string ""
in essence: bamf is reporting that the _GTK_UNIQUE_ BUS_NAME property is not set on the window when it is first presented.
This means that either
a) libreoffice is not setting the property before mapping the window (libreoffice bug)
b) bamf is telling lies (bamf bug)
Pretty sure we can rule out the chances of this being in any other component...