Accept uppercase E exponent after a leading-zero mantissa

[eeshsaxena]

tomlkit rejects the float 0E2, which is valid TOML. Both e and E are allowed exponent markers, and tomlkit already parses the lowercase form:

import tomlkit

tomlkit.parse("x = 0e2")   # ok, 0.0
tomlkit.parse("x = 0E2")   # tomlkit.exceptions.InvalidNumberError

The signed forms (+0E2, -0E2, 0E+2) fail the same way, and tomllib from the standard library accepts all of them. The valid/float/exponent-upper.toml case in the toml-test suite trips over this on its +0E2 line.

The cause is the "no leading zeros" guard in _parse_number: a mantissa beginning with 0 is exempted only when it is followed by ., o, x, b, or a lowercase e, so 0E... falls through to the leading-zero rejection. Adding 0E to the exempted prefixes fixes it. I covered the plain and signed forms in test_parser.py.