[JIRA] Commented: (MDA-48) create stage to strip all evidence of a given namespace

Ian Young (JIRA) noreply at shibboleth.net
Wed Sep 14 17:12:26 BST 2011


    [ https://issues.shibboleth.net/jira/browse/MDA-48?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13566#comment-13566 ] 

Ian Young commented on MDA-48:
------------------------------

I have a working version of this in the UKf project.  I will hang on to it there until we get closer to 0.7 so that I can beat up on it in our production environment.

> create stage to strip all evidence of a given namespace
> -------------------------------------------------------
>
>                 Key: MDA-48
>                 URL: https://issues.shibboleth.net/jira/browse/MDA-48
>             Project: Metadata Aggregator
>          Issue Type: New Feature
>          Components: DOM Metadata
>    Affects Versions: 0.6
>            Reporter: Ian Young
>            Assignee: Ian Young
>             Fix For: 0.7
>
>
> Need a stage to remove all evidence of a given namespace from a DOMMetadataItem.
> * remove all attributes in that namespace
> * remove all elements in that namespace
> * remove all namespace prefix definitions for that namespace
> If the namespace in question is used on the document element, that would have to result in an error being attached and the DOMMetadataItem being left otherwise unchanged, as we can't have a DOMMetadataItem without any associated element.
> The use cases for this are:
> * ingress filtering from a mdx partner who uses specific local namespaces you have no interest in
> * egress filtering when producing an export aggregate where you don't want to have your local namespaces visible because they aren't relevant to an mdx partner.
> There is a case to be made for another stage which strips all evidence of all namespaces not in a provided whitelist.  Although the mechanics would be similar, I think that it should be a separate stage if required at all, as the API would be quite different and I think the natural tendency would be to want some kind of status logged for each *unknown* namespace in that case: in this case, the namespace is expected but not wanted so no indication is required.
> I am assigning this one to myself because I am planning to prototype one of these in my UK federation extensions project.

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira


More information about the commits mailing list