Additions

A generic error for when a scraper fails
When using -h, any extensions that have print_help_<ext-name>(), will also print help
load_interface a command to load an interface the correct way instead of setting the intereface variable directly.
load_thumbnail_viewer same as load_interface but for thumbnail viewers
load_url_handler same as load_interface but for url handlers
post_scrape() a function that happens after everything is scraped
print_help_<ext-name>() an extension can print it's own help text
YTFZF_CHECK_VARS_EXISTS can now be used instead of check_vars_exists

Addons

buttons - adds extra buttons that when clicked do things such as:
- enable audio only
- enable/disable detach
- ...
ipc - allows getting variables from ipc_file which is located in the session's cache dir, see $HOME/.cache/ytfzf
- example usage:

echo "get is_audio_only" > ipc_file
cat "ipc_file.is_audio_only

^{yeah it's a bit clunky but working with sockets would be worse}

w3m thumbnail preview is now an addon
in interface_text if tput is not found, default to $COLUMNS, then 80 columns instead of defaulting to 80 immediately
info wait uses quick_menu instead of the read command.
handle_actions is much more efficient when handling many selections

curl adds a \r character causing weirdness when using --channel-link (issue: #557)
get_missing_thumbnails always says all thumbnails are missing, causing them to be redownloaded every time the menu is loaded with the -l option enabled.
remove_ansi_escapes removed more than just ansi escapes causing [ to be missing from text.
Using new lines in exported variables or variables set before running ytfzf would cause errors (#552 @HoffsMH).
line 163 unexpected operator when dependancy is not installed.
custom_shortcut_binds not doing anything.
__is_submenu does not save it's state when interface opens.
Writing watch history uses the jq command less, and should be slightly faster, more so if many videos are selected.
ytfzf assumes custom_info_wait_action_$info_wait_action exists.
Makefile creates and deletes a file called ${PROG}.bak (issue: #556)
the wrong invidious instance was used in the multi scraper.

The codebase is shellcheck compliant (with many disable comments)
refactor loading addons, such as interfaces, url_handlers, etc.. so that addons, or the user can just use load_<addon-type> addon-name.

resolving #379 by adding a margin on the side
implementing next_page or the alt-p shortcut in the invidious_search scraper