[ http://tools.oasis-open.org/issues/browse/OFFICE-3317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=20238#action_20238 ]
Dennis Hamilton commented on OFFICE-3317:
-----------------------------------------
Michale Stahl:
Yes, you're right. Now that I think about it, the ODF 1.0/1.1 consumer would likely treat the xml:id as a foreign attribute and ignore it.
I was concerned about a case where the document is considered by an implementation's XML processor to have more than one ID-valued attribute instance with the same NCName as that value. I agree that it is probably very difficult for an implementation to have a combination of arrangements (between itself and the XML processor it uses) that would create a problem here.
I will look at your file to see how to adjust this issue and any other that needs to be created for the other cases.
> ODF 1.2 Part 1 5.5.2 <text:changed-region> ID attribute inconsisently available
> -------------------------------------------------------------------------------
>
> Key: OFFICE-3317
> URL: http://tools.oasis-open.org/issues/browse/OFFICE-3317
> Project: OASIS Open Document Format for Office Applications (OpenDocument) TC
> Issue Type: Bug
> Components: Part 1 (Schema), Schema and Datatypes, Text
> Affects Versions: ODF 1.2 CD 05
> Environment: This defect applies since at least ODF 1.2 Part 1 CD04. This issue is presented in terms of ODF 1.2 CD05 Part 1.
> Reporter: Dennis Hamilton
> Priority: Blocker
> Fix For: ODF 1.2 CD 06
>
>
> 1. Every <table:changed-region> attribute requires an attribute of type ID in order to be referenced by the IDREF text:change-ide attribute in one or more change marks elements.
> However, with the current RNG Schema, there is no <table:changed-region> attribute with a value of type ID when there is no xml:id attribute. That is because the text:id attribute, which must be present when there is no xml:id attribute, has a value of type NCName, not type ID.
> 2. The definition for text:id and xml:id usage in this case does not correspond to what is required to satisfy the requirements for use of ID in accordance with the [XML-ID]. Those conditions surely have precedence over any supplementary conditions in the text of ODF 1.2. In particular, the text of section 19.811.5 specifies that there must always be an xml:id when there is a text:id, but that contradicts the schema, which makes explidit allowance for a standalone text:id attribute on a <text:changed-region> attribute.
> 3. It is not clear how backward compatibility is served by the peculiar requirement to have xml:id and text:id have the same value when both are present. This presumes that a down-level consumer will ignore the xml:id and not consider there to be a violation of [XML-ID] by the apparent appearance of to ID-valued attributes with the same value (according to the down-level schema plus [XML-ID]).
> [NOTE: There are other places where the same device appears to be attempted with regard to introduction of an xml:id attribute that pre-empts the type of another attribute formerly (and in isolation) having a value of type ID. Similar measures may be appropriate for those cases. This issue does not address those potential cases.]
--
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
