I was trying to build AAF components and it failed during unit testing.

I stumble upon this weird looking function (fromHex and fromHexNo0x) and I could not resist to test it.

It had wrong output for odd-count of hex digits, it somewhat works only for even digit count.

I made a new implementation which IMHO behaves (read: works) better. But maybe the original version was made so strange for reason which escapes me - so maybe this is not a bug...

I will add my source code to this ticket (HelloAAF.java) with both implementations and simple usage:

osp@tuxbox:~/Projects/wtf-aaf$ javac HelloAAF.java && java HelloAAF aaf
{{AAF: 10 160 }}
Me: 10 175

As you can see the original version converted 'af' as 160 which si clearly wrong...

Off-topic: Frankly, I really dislike when someone is reimplementing stuff which was done so many times before and most likely done better. Especially rolling one's own authentication/security stuff is very bad thing to do...there are bunch of proven libraries and schemes and these things are not to take lightly - I bet that any average pentester would found many holes in the whole AAF.

Please, ONAP people, stop reinventing wheels...