OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.

 


Help: OASIS Mailing Lists Help | MarkMail Help

sca-assembly message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]


Subject: NEW ISSUE: Bindings unnecessarily supports filters


Target: sca-assembly-1.2 WD 01 (PDF)

Title: Bindings unnecessarily supports "filters" element

Description:

In section 7 (& 7.5) of the specification, the definition "binding", updated to support eventing, allows for a new "filters" element.  Since the filters can (and will) occur on the containing "consumer", "producer", or "channel", it is at best redundant to identify filters here.  At worst, the interaction of filters specified on the binding, coupled with filters defined on the consumer might lead to subtle and surprising results.

As discussed on the mailing list [1], there might be reason to do this to accommodate existing messaging systems.  However, SCA already enables extensibility, and without concrete use-case scenarios showing otherwise, it is probably safe to assume that working with alternate/existing messaging systems will require more than specialization around filters.  The specification should leave any such specialization to extensions, as it does elsewhere.

As a final note, unless the support for filters within bindings is anticipated for any existing binding (or soon to be defined binding that supports eventing), the existence of such functionality in the specification will be untestable.

Proposal:

Remove the filters element from the definition of bindings and binding.sca.

(A detail related to the above proposal.  If we do not remove the "filters" element from bindings & bindings.sca, there's an alternate issue that arises [2].  The existing text for binding.sca says, of the filters child element, that "This subelement can only be used for bindings specified on channels, consumers, and producers."  This text does not appear on the definition of the base type of bindings - lines 3822 - 3824 - and almost certainly should.  Either an issue a separate issue should be raised to address that, or that question should be addressed in closing this issue.)

[1] http://lists.oasis-open.org/archives/sca-assembly/201007/msg00053.html
[2] http://lists.oasis-open.org/archives/sca-assembly/201007/msg00048.html



[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]