Forbid negative uvarint values

[jtraglia]

When reading over code, I noticed that it was possible to provide a 10-byte
uvarint (variable length integer) instead of the 9-byte maximum. The functions
did not properly catch it.

Also, fix what I believe is a typo in the ByteArray version. It checked the
wrong variable (result vs index) which would almost always result in a null
value. Might as well throw an exception there like the other implementation.

Fixes #225.

[ajsutton]

Sorry for the delay on this one - I haven't forgotten it, just not an area I'm particularly familiar with so harder to review.

[jtraglia]

No worries, I totally understand. There's no rush.

[ajsutton]

Ok, dug into the specs and this looks right to me - sorry it took so long. I've just added an additional test to confirm that we can successfully encode and decode Long.MAX_VALUE to ensure that the maximum length of 9 bytes is decoded.

[jtraglia]

@ajsutton Just so you know, testRoundTripMaximumValue appears to be a duplicate of testEncodeDecodeMaxLong:

jvm-libp2p/src/test/kotlin/io/libp2p/etc/types/UvarintTest.kt

Lines 28 to 33 in 5c3943d

	@Test
	fun testEncodeDecodeMaxLong() {
	val buf = Unpooled.buffer()
	buf.writeUvarint(Long.MAX_VALUE)
	assertEquals(Long.MAX_VALUE, buf.readUvarint())
	}

jvm-libp2p/src/test/kotlin/io/libp2p/etc/types/UvarintTest.kt

Lines 61 to 66 in 5c3943d

	@Test
	fun testRoundTripMaximumValue() {
	val buf = Unpooled.buffer()
	buf.writeUvarint(Long.MAX_VALUE)
	assertEquals(Long.MAX_VALUE, buf.readUvarint())
	}

[ajsutton]

Oh darn I didn't spot that. I'll have to clean that up. Thanks.