The benefit of this is that users only pay for what they use. If e.g.
only `vim.lsp.buf_get_clients()` is called then they don't need to load
all modules under `vim.lsp` which could lead to significant startuptime
saving.

Also `vim.lsp.module` is a bit nicer to user compared to
`require("vim.lsp.module")`.

This isn't used for some nested modules such as `filetype` as it breaks
tests with error messages such as "attempt to index field 'detect'".
It's not entirely certain the reason for this, but it is likely it is
due to filetype being precompiled which would imply deferred loading
isn't needed for performance reasons.

refactor(lsp): move glob parsing to vim.glob

Moving the logic for using vim.lpeg to create a match pattern from a
glob into `vim.glob`. There are several places in the LSP spec that
use globs, and it's very useful to have glob matching as a
generally-available utility.

vim.iter is now both a function and a module (similar to vim.version).

test: replace lfs with luv

luv already pretty much does everything lfs does, so this duplication
of dependencies isn't needed.

- version.cmp(): assert valid version
- add test for loading vim.version (the other tests use shared.lua in
  the test runner)
- reduce test scopes, reword test descriptions