- From: Takeshi Imamura <IMAMU@jp.ibm.com>
- Date: Mon, 11 Mar 2002 17:20:10 +0900
- To: reagle@w3.org
- Cc: merlin <merlin@baltimore.ie>, "Hiroshi Maruyama" <MARUYAMA@jp.ibm.com>, xml-encryption@w3.org
>> >If an xenc:EncryptedData element node being decrypted is the first node >> > in X, the value of its Type attribute MUST NOT be xenc;Content. This >> > prevents an ill-formed XML document with element content appearing at >> > the start of the document. If the xenc:EncryptedData is not the first >> > node in X, the value MUST be xenc;Element or xenc;Content. This >> > prevents binary data from appearing out of place in an XML document. >> >> I think this paragraph looks good, although "with element content >> appearing" should perhaps be "with invalid content appearing"? > >I struggled with that, perhaps I should reuse the "ill-formed" again. >Binary data can appear if it's in CDATA I think, so I don't want to make it >seem like that could never happen. (And invalid content seems to presume >validation...) The first restriction is a restriction for ensuring that a node-set is always single-rooted rather than preventing an ill-formed XML document or an XML document with invalid content. >> Should we generalize to allow a single non-XML EncryptedData >> to appear anywhere in the excepted input, rather than requiring >> that non-XML EncryptedData be the root node? >> >> ... If the xenc:EncryptedData is not the first node in X, and its >> type is neither &xenc;Element nor &xenc;Content, then it MUST >> be the only xenc:EncryptedData in X not referenced by an Except >> element. This prevents mixed decryption of XML and non-XML data, >> and restricts the decryption transform to a single piece of >> binary data at a time. >> >> I'm not terribly pushed on this, it might just make some uses >> easier; for example, I can reference an external XML document >> containing one piece of encrypted binary data that is not the >> root, without using an XPath transform to select the encrypted >> data element; somewhat like the base-64 transform ignoring XML >> data. > >I don't feel very strongly but I would prefer not. I support Merlin. Actually his suggestion is what I intended in the original text by "an xenc:EncryptedData element node being decrypted". That is, EncryptedData element nodes referenced by Except elements can appear anywhere in a node-set and should be ignored when checking if restrictions on the Type attribute are satisfied. This is not only the case for non-XML EncryptedData element but the case for XML EncryptedData element. Thanks, Takeshi IMAMURA Tokyo Research Laboratory IBM Research imamu@jp.ibm.com
Received on Monday, 11 March 2002 03:20:20 UTC