Weird strings that give errors when parsing with lenient_parse.to_float

Just a couple of the potentially interesting strings that I found with the property tests.

- this one is interesting:
- original: `"+2_2_6_275_1.\r\n\u{000B}\n\n "`, shrunk: `"9.\t"`, error: `UnknownCharacter(13, "\r\n")`
- Python gives: `2262751.0` for original and `9.0` for shrunk

These ones probably should be treated as expected for now:

- numbers that go out of bounds (expected to crash)
- original: `"\r \r\r\t-0_09.0_0_1E+7_7_0_4_6\t\u{000B}"`, shrunk: `".9E999"`
- non-finite (expected crash really)
- original: `"\u{000B}\f\r\n\t\r\r+iNFiNITy\n\u{000B}\u{000B}\f\r\n\t\r\r"`, shrunk: `"iNFiNITy"`
- (I imagine `nan` will give the similar)

---

I will hold off on putting more here for now. The subsequent ones that are unexpected as I find them, I will put into example test cases to anchor the tests.