[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: [OASIS Issue Tracker] Commented: (OFFICE-2687) ODF 1.2 Part 3:Unspecified Signature-File Name Allocation
[ http://tools.oasis-open.org/issues/browse/OFFICE-2687?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=22499#action_22499 ]
David LeBlanc commented on OFFICE-2687:
---------------------------------------
Capturing discussion between myself and Michael Brauer from ODF TC list:
David LeBlanc said:
It is a matter of being able to efficiently interoperate. If one implementation puts document signatures in foosigantures.xml and the other puts it in barsignatures.xml, and a third puts it in bazsignatures.xml, then we all end up with code looking for each file name. If we all agree to put document signatures into documentsignatures.xml, then all of us have simpler code.
The alternative is for an implementation to open *signatures* and then try to guess at the purpose of each signature, when we have not defined a way to designate purpose.
I also place a high value on not breaking an existing implementation, and a lot of value in my code interoperating nicely with the existing implementation, should Microsoft decide to implement signatures for ODF.
Michael Brauer replied:
David,
David LeBlanc wrote:
> It is a matter of being able to efficiently interoperate. If one implementation puts document signatures in foosigantures.xml and the other puts it in barsignatures.xml, and a third puts it in bazsignatures.xml, then we all end up with code looking for each file name. If we all agree to put document signatures into documentsignatures.xml, then all of us have simpler code.
You are right, but I have nowhere suggested that we give up the reserved file name "documentsignatures.xml"
>
> The alternative is for an implementation to open *signatures* and then try to guess at the purpose of each signature, when we have not defined a way to designate purpose.
My view on this is:
ODF defines a set of predefined filenames which each contains a kind of signatures. Right now we have only "documentsignatures.xml", but we may add something like "contentsignatures.xml", which would be what you have proposed in another mail. We had "macrosignatures.xml", but there was too little we could say about them that there seems to be consensus to drop them.
For all these signature files, implementations don't have to guess.
There may be anyway other signatures files. Implementations may guess what their purpose is, or they may ignore them. The issue at hands, if I got it correctly, is whether we need a naming scheme for this that somehow includes the vendor name.
> ODF 1.2 Part 3: Unspecified Signature-File Name Allocation
> ----------------------------------------------------------
>
> Key: OFFICE-2687
> URL: http://tools.oasis-open.org/issues/browse/OFFICE-2687
> Project: OASIS Open Document Format for Office Applications (OpenDocument) TC
> Issue Type: Bug
> Components: Packaging, Part 1 (Schema), Part 3 (Packages), Security
> Affects Versions: ODF 1.2 CD 05
> Environment: This issue applies to ODF 1.2 Part 3 CD01 and its revisions. The text discussed is as in OpenDocument-v1.2-part3-cd1-rev04.odt
> Reporter: Dennis Hamilton
> Assignee: Dennis Hamilton
> Fix For: ODF 1.2 CD 06
>
>
> Part 3 Section 2.5 establishes Digital Signatures as stored in one or more files whose "relative pahs" begin with 'META-INF/' and that contain the "term" 'signatures'. I take this to read that digital signature files SHALL have names of the form 'META-INF/*signature*' in the Zip archive, using conventional wild-card notation. The format of these files is established in Section 4.
> Part 3 Section 4 says nothing about the naming of the file, although it refers to <dsig-document-signatures> as a root element and it asserts that the schema for the file is that given in Appendix A. There are no explicit producer conformance requirements. The only strong requirements on consumers concern differences in signature verification depending on whether or not a digital signature file is encrypted or not (in section 4.2).
> Part 3 Section 7.2.1 on Conforming OpenDocument Packages asserts in (PD1.4-PD1.6) that the only permitted files beside META-INF/manifest.xml with META-INF/* name structure are those which have name structure META-INF/*signatures* and which are XML 1.0 documents with root element <dsig:document-signatures>, along with a few additional conditions. (Note that occurrence in /META-INF/manifest.xml is not required, so such files do not have explicit MIME types nor are they subject to encryption when not listed in META-INF/manifest.xml.)
> Part 3 Section 7.2.2 on Conforming OpenDocument Extended Packages does not have any limitation on what can have a META-INF/* form name.
> There is no stipulation that production of META-INF/* and META-INF/*signature* files be implementation defined. There is no stipulation that consumer-permissive limitations on what is accepted as a valid digital signature are implementation defined.
> ALLOCATION OF META-INF/*signatures* NAMES
> The point of the preceding review is to confirm that there is no provision for how META-INF/*signature* name forms are allocated nor whether and how "implementation-defined" appropriation of such a name for a particular purpose is accomplished. There is similarly no basis for an implementation determining whether the occurrence of such a named Zip archive file is one for which the implementation has been designed.
> There is no provision similar to those in other places where ODF 1.2 provides for explicit identification of implementation-defined provisions via namespace bindings.
> At the same time, all occurrences of META-INF/*signature* name forms for XML 1.0 documents having <dsig:digital-signatures> root elements and satisfying a few other conditions are permitted as part of Conforming OpenDocument Packages.
> Likewise, consumers may restrict what is permissible in a digital signature without any requirement that this be implementation-defined and with no provision considering how consumers apply restrictions on what digital signatures there are and which ones are subject to what constraints.
> An interoperable arrangement is required.
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: http://tools.oasis-open.org/issues/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]