Hope this will make people using feed_command less likely: this hides bugs.
Already found at least two:

1. msgpackparse() will show internal error: hash_add() in case of duplicate
   keys, though it will still work correctly. Currently silenced.
2. ttimeoutlen was spelled incorrectly, resulting in option not being set when
   expected. Test was still functioning somehow though. Currently fixed.

Note: typo, ttimeoutlen not set ever.

Mention @hardenedapple

If we `set pastetoggle=abcde`, and manually type it, then `vgetorpeek()`
sees part of the option before it has all been inserted into the
typebuffer.
To signify this it sets `keylen = KEYLEN_PART_KEY`, but the condition
about whether to return the current key from `vgetorpeek()` only checks
for `keylen = KEYLEN_PART_MAP`.
Add a check for `KEYLEN_PART_KEY` to account for the `'pastetoggle'`
option.

- Vim "unix default" of 'noshowcmd' is serving few users. And it's
  inconsistent.
- 'ruler' and 'belloff=all' improve the out-of-the-box experience.
- Continue to use 'noshowcmd' and 'noruler' by default in the functional
  tests to keep them fast.

TODO: Add a "disable slow stuff" command or mapping to address the
use-case of a very slow terminal connection.

This allows executables to be found by :!, system(), and executable() if
they live next to ("sibling" to) nvim.exe. This is what gvim on Windows
does, and also matches the behavior of Win32 SearchPath().

https://github.com/vim/vim/blob/c4a249a736d40ec54794827ef95804c225d0e38f/src/os_win32.c#L354-L370

After #4964 environment variables in the XDG "fallback" table are no
longer expanded.

Fallback to correctly expanded $LOCALAPPDATA, $TEMP. If that fails
(unlikely), fallback to hard-coded paths (e.g. ~/AppData/Local).

Closes #5255

As noted in “:help 'packpath'”, the default value is supposed to be the
same as that for 'runtimepath'.  This was missed in the original port of
the packages functionality from Vim.

Closes #5193