When the slideshow state is toggled, the info line needs to be redrawn as it
may use the %a format specifier (which shows whether the slideshow is running
or paused). feh_reload_image does this, but also resets all pan and zoom
options and reloads the image from disk. winwidget_render_image only redraws
info, caption and similar lines in this case, which is more efficient and does
not incur unwanted side-effects.

This is in line with other software. save_filelist now defaults to "L".

Quoting glibc rand(3):

The  versions  of rand() and srand() in the Linux C Library use the same random
number generator as random(3) and srandom(3), so the lower-order bits should be
as  random  as  the  higher-order bits.   However,  on  older  rand()
implementations, and on current implementations on different systems, the
lower-order bits are much less random than the higher-order bits.  Do not use
this function in applications intended to be portable when good randomness is
needed.  (Use random(3) instead.)

Closes #218

Closes #326

When feh loses its controlling terminal at runtime, e.g. due to backgrounding /
disowning, it will no longer issue a warning on each terminal keystroke.

Turns out that it is undefined behaviour to pass a value to isctype functions
which does not fit inside a char. Closes #312

Closes #303

like %o and %z in slideshow actions (I would like to use this to zoom in, pan,
and then use an action to crop the window to zoomed in view).

Fixed(?) Makefile document build issue for README.md

Many image collections are organized by directory, so it is nice to have
jump-to-adjacent-directory navigation.

e.g. Given the following file hierarchy:

    .
    ├── A
    │   ├── 1.jpg
    │   ├── 2.jpg
    │   └── C
    │       ├── 1.jpg
    │       ├── 2.jpg
    │       └── 3.jpg
    └── B
        ├── 1.jpg
        ├── 2.jpg
        └── 3.jpg

`feh --recursive` creates the following filelist:

    A/1.jpg <---- current_file
    A/2.jpg
    A/C/1.jpg
    A/C/2.jpg
    A/C/3.jpg
    B/1.jpg
    B/2.jpg
    B/3.jpg

If we press [next_dir], we move the current_file pointer to:

    A/1.jpg
    A/2.jpg
    A/C/1.jpg <-- current_file
    A/C/2.jpg
    A/C/3.jpg
    B/1.jpg
    B/2.jpg
    B/3.jpg

Pressing [next_dir] again moves the pointer to:

    A/1.jpg
    A/2.jpg
    A/C/1.jpg
    A/C/2.jpg
    A/C/3.jpg
    B/1.jpg <---- current_file
    B/2.jpg
    B/3.jpg

[next_dir] now moves the pointer back to the top of the list:

    A/1.jpg <---- current_file
    A/2.jpg
    A/C/1.jpg
    A/C/2.jpg
    A/C/3.jpg
    B/1.jpg
    B/2.jpg
    B/3.jpg

Pressing [prev_dir] from here moves backwards to the first image of the
previous directory:

    A/1.jpg
    A/2.jpg
    A/C/1.jpg
    A/C/2.jpg
    A/C/3.jpg
    B/1.jpg <---- current_file
    B/2.jpg
    B/3.jpg

When starting from an position that is not the first image of a
directory, [prev_dir] moves the pointer to the first image of the
current directory.

These actions combine well with `--sort dirname` since all regular files
in a directory will be sorted before any subdirectories, avoiding a
filelist like the following:

    A/1.jpg
    A/SUBDIR/2.jpg
    A/SUBDIR/3.jpg
    A/4.jpg

With `--sort dirname` that filelist becomes:

    A/1.jpg
    A/4.jpg
    A/SUBDIR/2.jpg
    A/SUBDIR/3.jpg

previously, a button/key definition with an invalid action name would assign
the specified key to the most recent valid action. E.g. "zoom_in 4\ninvalid 5"
wuold assign button 5 to zoom_in.

this actually uses the old "force geometry" workaround to get the
scale down behaviour (which makes the code a lot cleaner) by
first setting the geometry after the window is created (to avoid
creating a 0x0 sized window, which X does not like).

this introduces a bug in feh_wm_set_bg, which will be fixed in the next
commit (only applies when run without HOME)

Patch by sdaau on IRC. Thanks!

This reverts commit b257e011.

Conflicts:

	ChangeLog