ISSUE-56: Whether or not schema-informed AT(*) should be permitted to match xsi:nil and xsi:type attributes
Whether or not schema-informed AT(*) should be permitted to match xsi:nil and xsi:type attributes
- State:
- CLOSED
- Product:
- EXI Spec
- Raised by:
- Takuki Kamiya
- Opened on:
- 2010-06-15
- Description:
- Current spec allows xsi:nil and xsi:type attributes to be represented using schema-informed AT(*) productions. This seems to unnecessarily complicate some implementations.
- Related Actions Items:
- No related actions
- Related emails:
- RE: Possible side-effect of test case changes done for ISSUE-55 & ISSUE-56 (from Youenn.Fablet@crf.canon.fr on 2010-06-25)
- Possible side-effect of test case changes done for ISSUE-55 & ISSUE-56 (from tkamiya@us.fujitsu.com on 2010-06-24)
- Agenda for 23 June EXI Telecon (from tkamiya@us.fujitsu.com on 2010-06-22)
- Agenda for 16 June EXI Telecon (from tkamiya@us.fujitsu.com on 2010-06-15)
- ISSUE-56: Whether or not schema-informed AT(*) should be permitted to match xsi:nil and xsi:type attributes [EXI Spec] (from sysbot+tracker@w3.org on 2010-06-15)
Related notes:
Youenn pointed out xsi:nil with invalid boolean values need to match AT(*) [untyped value]. We'll still need to discuss whether the similar is also true for xsi:type.
Takuki Kamiya, 16 Jun 2010, 20:17:06Youenn agrees @xsi:type always match AT(xsi:type) productions in deviation mode.
http://lists.w3.org/Archives/Member/member-exi-wg/2010Jun/0067.html
It needs to be discussed what to do about @xsi:type in "AT(*) + no AT(xsi:type) production" case, which happens only in strict schema-informed grammars.
http://lists.w3.org/Archives/Member/member-exi-wg/2010Jun/0070.html
The WG reached consensus that xsi:type attributes should never match AT(*) or AT(*) [untyped value] event types in schema-informed grammars.
Takuki Kamiya, 23 Jun 2010, 21:11:36Display change log