Replace Pacman() syscall logic with libalpm

[Torxed]

PR Description:

This PR will enable us to move away from running pacman -XYZ via system calls, and instead rely on libalpm directly, via pyalpm.

This has the added benefit that we can do package search efficiently, meaning we can replace:

archinstall/archinstall/lib/packages/packages.py

Lines 45 to 59 in 26cf046

	def package_search(package: str) -> PackageSearch:
	"""
	Finds a specific package via the package database.
	It makes a simple web-request, which might be a bit slow.
	"""
	# TODO UPSTREAM: Implement bulk search, either support name=X&name=Y or split on space (%20 or ' ')
	# TODO: utilize pacman cache first, upstream second.
	response = _make_request(BASE_URL_PKG_SEARCH, {'name': package})
	if response.code != 200:
	raise PackageError(f"Could not locate package: [{response.code}] {response}")
	data = response.read().decode('UTF-8')
	json_data = json.loads(data)
	return PackageSearch.from_json(json_data)

Allowing us to do local searches, and let pacman deal with databases and searching.
This simplifies custom mirrors and allowing us to search against them too.

Some thoughts

I decided to use a context manager approach, as it allows us to choose between interacting against system-wide pacman data - but during context creation also decide if we want a temporary pacman data structure.

That way we can do:

with Pacman(temporary=True, servers=['http://ftp.lysator.liu.se/pub/archlinux/$repo/os/$arch', 'http://ftp.acc.umu.se/mirror/archlinux/$repo/os/$arch']) as pacman:
	print(pacman.search('nano'))

To not collide with system-wide pacman, and perform searches and syncs independently and irregards if existing pacman is executing.

With this we can remove:

archinstall/archinstall/lib/pacman/__init__.py

Lines 32 to 43 in 26cf046

	pacman_db_lock = Path('/var/lib/pacman/db.lck')
	if pacman_db_lock.exists():
	warn(_('Pacman is already running, waiting maximum 10 minutes for it to terminate.'))
	started = time.time()
	while pacman_db_lock.exists():
	time.sleep(0.25)
	if time.time() - started > (60 * 10):
	error(_('Pre-existing pacman lock never exited. Please clean up any existing pacman sessions before using archinstall.'))
	exit(1)

And if we want to interact with pacman system wide, we simply just do:

with Pacman() as pacman:
	pacman.sync()
	pacman.install("vim")

I'm sure we could achieve the same thing with using only __init__ and some if/else.
But for some reason this felt like a logical thing to do, especially since all pacman operations are "transaction based", meaning we need to create a lock file while doing operations - and then remove the lock file once we're done.

And if libalpm/pacman crashes during an operation, we need a way to do easy cleanup and remove said lock file.

[Torxed]

Some additional commits came in, will revert those/this PR and clean that up.