Possible Over-Iteration in PrettyBytes

[VaslD]

I was looking for a not %02x way of converting binary to hex String, and I stumbled upon the implementation in PrettyBytes.swift.

swift-crypto/Sources/Crypto/Util/PrettyBytes.swift

Lines 45 to 51 in 9cc89f0

	self.regions.forEach { (_) in
	for i in self {
	hexChars[Int(offset * 2)] = itoh((i >> 4) & 0xF)
	hexChars[Int(offset * 2 + 1)] = itoh(i & 0xF)
	offset += 1
	}
	}

I'm not 100% sure about the Sequence conformance of non-contiguous DataProtocol, but it seemed to me that the self on line 46 should be the ignored parameter $0 (or maybe eliminate the region logic since looping over self most likely would have taken non-contiguous memory regions into account?), otherwise the code reads:

For each segment of the data buffer, loop over the entirety of data buffer and convert each byte to two hex chars.

Unless non-contiguous DataProtocol has non-stable indices and iterators, this code will over/re-iterate the data buffer.

[Lukasa]

Yup, it sure does! Would you like to submit a patch to fix it? As a practical matter I think we can ignore the regions here, and just use the sequence conformance directly.

[VaslD]

PR done. Don't know how to construct a multi-region Data to test it yet, but it shouldn't get worse. :-)