It was inconsistent with the rest of the documentation and even with
itself, subsections weren't...

Add :col[der] and :cnew[er] commands to menu mode to traverse history of
navigation menus.

Add ...

Combined with automatic stashing on menu switches this allows using
show_*_menu() functions to s...

It should have already been set on menu initialization.

No need to leave menu mode to open the history.

This is done for the symmetry when you open :chistory and see that there
are newer menus, but :c...

If the menu was stashed before or wasn't but is stashable.

This means that nothing special need...

Changes location of menu data while managing its initialized flag.

These commands:
- :keepsel
- :plugin
- :plugins
- :stop

This will make switching back to menus easier and only complicates
things anyway.

It's not used.

It replaces menu of the menu mode. Previously modmenu_set_data() was
used directly by menu impl...

State should be preserved, no need to drop link to current menu data or
view.

As opposed to doing this on initializing one how it was done before.
Previous approach is not ju...

ts=8 results in sw=8 if one uses sw=0, which is inconvenient.

If it's different from the current directory of an active view.

This avoid ugly back and forth mode changes.

This leaves `m` for `menu_data_t`.

For convenience, this doesn't make anything new possible, just makes for
a quicker workflow.

:c...

It was inconsistent with the rest of the documentation and even with
itself, subsections weren't...

When they are used for line continuation in command examples.

They can be used to traverse history of navigation menus.

It displays command-line history of menus.

And to 'sessionoptions' option as well.

It manages storing of command-line history of menus.

Add command-line history to menu mode.

Add "mchistory" value to 'vifminfo' and 'sessionoptions'...

Try to wait for file system changes to be visible, they seem to be
postponed on Windows (at leas...

It didn't do anything because callback is called as if we're in a normal
mode (menu commands his...

It displays command-line history of menus.

Always empty for now.

And to 'sessionoptions' option as well.

It manages storing of command-line history of menus.

Just to make sure that new command-line submode won't be added without
considering this function.

Currently contains only one function.

This puts an end to using menus_reset_data() for internal purposes of
freeing the data.

The dif...

Allow wild menu popup to take up up to 40% of the screen.

In one place it's not used.

It was used without check for NULL in compare.c before (see
b1bd924f4a97c143581e255c5278a3d505b9...

Previously upper limit was 10, which wasn't convenient for a tall
window. Now it's 40% of the s...

Don't even try to clear graphics when menu-like mode is active. The
graphics should have been h...

Its second half was a copy of the first one. Apparently, I meant to use
SK_BY_ITEMS and reorder...

Sets up sorting with up to 2 sorting keys.

The memset() way is not very readable and is less co...

That's the default sorting and both `view_t::sort` and `view_t::sort_g`
are initialized with SK_...

The graphics should have been hidden on entering the mode.

This way we can check current mode to know whether any graphics can be
on the screen or not.

Make register's content being listed in normal rather then reversed
order in :registers/:display...

The order was reversed probably to list the last added file first, but
since then files in regis...

To reproduce:
1. :windo set nonu
2. :set number
3. <c-w><c-x>
4. :set number
5. Nothing hap...

To make tests more readable and drop repetitive code pattern and a
comment.

Don't draw right padding on a truncated rightmost column of a transposed
ls-like view.

Fix line...

Not sure what's it about, but dropping it doesn't seem to break
displaying anything. Let's remo...

Add a generic comment that should help to understand how things are
drawn.

Keep it simple and manage which padding is displayed outside of the
function.

It applies only to left or right miller column and seem to be a
compensation for the separator c...

ui_view_displays_columns() must be used to take care of options that
have higher priority than '...

Instead of doing it on the call site, which was confusing. Also explain
why this is needed in a...

The function will strip padding itself.

ui_view_available_width() is now ui_view_main_area().

And ui_view_main_padded() is `ui_view_ava...

Before this change, padding was included only for ls-like view which was
inconsistent.

I'm pretty sure that padding of both sides should be included in the
value.

Otherwise it looks like the column is drawn in full.

Include padding from both sides in widths of columns.

Use of -1UL in several places suggest that signed values are at least as
good.

Currently offset is passed as -1 in several places despite the argument
being `size_t`. Negativ...

`len` is actually `prefix_len`, not full length of printed output.

Easily visible after `highlight LineNr ctermbg=red` and especially so
after `set millerview` (al...

When using it in column_line_print(), it looks very confusing as total
width is supposed to incl...

Somewhat more readable code in several places.

The code matches exactly.

Build AppImage with enabled gpm support.

Build AppImage with ncurses 6.4.

Fix *.xcf in sample ...

There were issues in the past, see 15122a576fa2b6af730416c74896173ea664423f:

Build ncurses ...

Thanks to aleksejrs.

Add Midi and more graphics to sample vifmrc.

Add *.unknown_video to vifmrcs and Default-256.

U...

Thanks to aleksejrs.

Closes #891 on GitHub.

Thanks to aleksejrs.