Util: rename/adjust hexStringToBytes to prefixedHexStringToBytes

[holgerd77]

We have made bad experiences with being loose on hex string prefixes - also on the input side - and allow for lazy input and addition or removal, getting more strict here was therefore a substantial part of one of former breaking release rounds.

The current hexStringToBytes() method does a lazy if (hex.startsWith('0x')) { hex = hex.slice(2) } conversion here.

This leads to the problematic situation that it is just not deterministically decidable if a 0x string is either a prefix or part of the string itself, in case both is allowed.

So let's say one has the - let's assume - prefixed hex string:

0x0x1234

The following would then be the - unprefixed hex string:

0x1234

So it is just not decidable if 0x is part of the string or it is a prefix and it is highly recommendable to request from users to decide on the format before-hand. The prefixed version is preferred here since once can always throw if a non-prefixed version is provided, and so do an always-working format check.

This is not only a theoretical example. People can put arbitrary byte data on some fields (e.g. extraData) or might even do this on purpose 😋 to annoy developers and cause things to break.

So, long story short: we should definitely update here before the breaking releases and rename to prefixedHexStringToBytes(), remove the 0x stripping and throw for all non-0x-strings passed.

I will already add/reference the updated name to the release notes.