Change `ignore_missing_stub` default to `false`?

[Avasam]

I've been wondering about this.

Possible reasons:

1. More and more stubs are "stubtest" complete (not counting pyright and mypy)
   1.1. It should be relatively easy to "minimally complete" missing ones using stugben. Especially after the reduction of stubtest false-positives by Stubtest: Improve heuristics for determining whether global-namespace names are imported mypy#14270
   1.2. Eventually most if not all stubs will have ignore_missing_stub = false, which could be the only entry under [tool.stubtest]
2. It's easier to search for ignore_missing_stub = true than to search for a missing line.
3. None would mean False in the metadata parser. (feels more intuitive?)
4. Remove partial classes/modules from CONTRIBUTING? #8955

   I don't want to encourage new incomplete stubs

Cons:

• Have to update the metadata parser
• A commit that flips (adds/removes) the value in all metadata.toml
• Maybe wait until there's even less (if not all) "stubtest incomplete" stubs left.

[AlexWaygood]

I agree that we should change the default here. @hauntsaninja proposed doing this a while back (can't find the reference immediately), and I opposed it then. But I think I've been proven wrong since.