Even though i3 cannot know the width/height of some workspaces as
long as they are not initialized (say you used workspace 1 and 3,
but not workspace 2), some applications require this information.
In this case, it was Firefox which intersects the available workareas
(see mozilla/gfx/src/gtk/nsScreenGtk.cpp) and did not position some
windows correctly when being confronted with zero-width/height
workspaces.
Please note that rdesktop’s -g workarea option will not work on
64-bit systems at the moment because of a bug in rdesktop (see the
rdesktop-devel mailing list).
Starting from this commit, a borderless window will always be
borderless if it is the only window in a container. For example,
you can have Firefox borderless in a tabbed container and as soon
as the download manager or a viewer gets opened, the container
will be rendered like a normal tabbed container.
This solves the user-interface dilemma of borderless/1-px-border
windows inside stacked/tabbed containers, at least for this special
case. Thanks to Merovius for this suggestion.
This little hack runs make recursively to generate include/loglevels.h
before running any other target but skip an explicit dependency on
loglevels.h in each rule. Therefore, you do not need to rebuild
every source file when compiling.
Using shell commands, a bitmask is generated for each file. Additionally,
a C header containing an array of loglevels and their files is created in
include/loglevels.h.
When having 8 windows in a container which has 766 px available,
you ended up losing 0,75 px per window which would quickly sum up.
Now, the rest space (6 px in this example) is distributed in units
of one pixel to as many windows as possible.
This bug could happen if you have floating and tiling windows (for
example Firefox in tiling mode and its Open dialog in autmatically
floating mode) and you opened a new tiling window while in fullscreen.
i3 would then place the window below the floating windows, but
floating clients are above fullscreen windows. Thus, the client
would be placed above the fullscreen window.
This prevents errors in rounding leading to an unoccupied space of
-1 which in turn leads to infinity when calculating the new size
of a container after resizing.
Minimum width/height was not consistent with the limit for grabbing
and resizing a window at its border.
If one of both was violated (width < min_width for example), none
of them were updated.
Michael Stapelberg