Hi,It seems that I'm wrong. After reading the spec better it seems is needed to c14n the xmlsignedinfo. Perhaps there is a way of avoiding the c14n & reimport because it only make one test case fail. If anyone has any idea...
My patch don't handle well this test case. It seems that it take on account that the signed info is going to be c14n, reparsed & reimported. But this is not alway the case. The SignedInfo is not c14n and reimported if the c14n method is "safe". As stated in the second paragraph of this mail
http://lists.w3.org/Archives/Public/w3c-ietf-xmldsig/2001OctDec/0054.html. And also in the REC http://www.w3.org/TR/xmldsig-core/#sec-CanonicalizationMethod-NOTE, it saids clearly that the above behavior is not always but only for arbitrary c14n methods.
What do you think is the good behavior? For me it is weird to have a test case that relays in this kind of unstandard behavior. And the parse and imports is a very wasteful process that need to be only done with insecure c14n. But if you think that the test is correct I can correct my patch and send it back again.
Regards
Regards
