Unfortunately, it is not the case with xmlsec :( I guess I need to fix it :) Aleksey > Hum, I'm pretty sure I was walking that list *backward* precisely > to have user provided handler taking priority, at least at the libxml2 > level: