Hi Alessio,
there are two reasons for this. First of all my boss asked me to provide a way to do such manual verification, so in case of contestations we can give more argumentations than "JBoss told me it was ok"; second, it's my personal interest to understand a bit more how these things work (I often need to re-implement things to ensure I've understood well).
It's strange that I need to transform the xml text in two different ways to calculate the correct digests: for a successful reference validation i need to put the referenced (canonicalized) element on a single line (no spaces/line-breaks between tags), while for a successful signature validation i need to remove only the leading spaces at the beginning of each line.
Anyway I'm confused about xml canonicalization. I thought it was a way to give a single and unambiguous string representation of an xml, focalizing on its semantic. Instead, xml canonicalization preserves indentation of text and I can't understand the reason, I think I'm not getting its real purpose.
Thank you in advance for any clarification :)