On 31/07/2012 08:20, Rémy Maucherat wrote:
> On 07/31/2012 12:57 AM, Mark Thomas wrote:
>> On 30/07/2012 23:49, Shing Wai Chan wrote:
>>> On 7/30/12 3:41 PM, Mark Thomas wrote:
>>>> On 30/07/2012 23:21, Shing Wai Chan wrote:
>>>>> I am looking at issue http://java.net/jira/browse/SERVLET_SPEC-36
>>>>> ("Clarify relationship of metadata-complete and
>>>>> ServletContainerInitializers")
>>>>>
>>>>> The metadata-complete=true means the completeness of data specified in
>>>>> deployment descriptor.
>>>>> But it does not mean the annotation will be ignored in the given
>>>>> application.
>>>>> For instance, CDI annotations will still be scanned and processed.
>>>>>
>>>>> The ServletContainerInitializer is a service provider specified in
>>>>> META-INF/services, not by annotation. In a
>>>>> ServletContainerInitializer,
>>>>> there may be a @HandlesTypes, which does not have corresponding
>>>>> metadata
>>>>> in deployment descriptor. Since a ServletContainerInitializer and
>>>>> @HandlesTypes work together, it should be interpreted whenever a
>>>>> ServletContainerInitializer is loaded.
>>>>>
>>>>> In section 8.2.4 of Servlet 3.0,
>>>>> "In addition to the ServletContainerInitializer we also have an
>>>>> annotation - HandlesTypes. The annotation will be applied on the
>>>>> implementation of ServletContainerInitializer to express interest in
>>>>> classes that are either annotated with the classes specified in the
>>>>> value or if a class extends / implements one of those classes anywhere
>>>>> in the classes super types."
>>>>>
>>>>> I propose the following clarification in wording:
>>>>> "In addition to the ServletContainerInitializer we also have an
>>>>> annotation - HandlesTypes. The annotation will be applied on the
>>>>> implementation of ServletContainerInitializer*, independent of
>>>>> metadata-complete,* to express interest in classes that are either
>>>>> annotated with the classes specified in the value or if a class
>>>>> extends
>>>>> / implements one of those classes anywhere in the classes super
>>>>> types."
>>>>>
>>>>> Shing Wai Chan
>>>> What about the ordering issue? To quote one of your comments from
>>>> SERVLET_SPEC-36:
>>>> <quote>
>>>> ... excluded jars in absolute-ordering independent of
>>>> metadata-complete=true will not be scanned for
>>>> ServletContainerInitializers.
>>>> </quote>
>>>>
>>>> I take this to mean "...will not be scanned for classes that match
>>>> SCIs"
>>>> which works for me.
>>> The previous quote in issue tracker is actually in p.67 of Servlet 3.0.
>>> Yes, that means the excluded jar (from <absolute-ordering>) will not
>>> scan the ServletContainerInitializer.
>> Excellent.
>>
>>> In 8.2.4, I try to clarify the relation between @HandlesTypes and
>>> metadata-complete, not about the loading of ServletContainerInitializer.
>>> Please let me know whether the spec is clear enough in this case.
>> I think we need to clarify the text in 8.2.2.1.d
>>
>> Currently is states:
>> Excluded jars are not scanned for classes to be handled by any
>> ServletContainerInitializer.
>>
>> I think this needs to be:
>> Irrespective of the setting of metadata-complete, JARs excluded by an
>> <absolute-ordering> element are not scanned for classes to be handled by
>> any ServletContainerInitializer.
>>
>> or something along those lines.
> I don't see how or why HandlesTypes is related to metadata. This is not
> an annotation which defines metadata.
First of all, the spec states that JARs excluded from absolute ordering
are excluding from scanning for classes that match classes defined by
the HandlesTypes annotation.
The query was what happens when metadata-complete is true since one
might expect that fragment ordering is irrelevant in that case.
The clarification is that regardless of metadata-complete (true or
false) if an absolute ordering is present then JARs excluded from the
absolute ordered are not scanned for classes that match HandlesTypes.
Since there is no way to disable the scanning for SCIs, the
clarification is a good thing since it means that you can at least
disable the potentially very expensive scanning of every single class in
the application (and it's full class hierarchy) checking for
HandlesTypes matches. For large applications that don't want the SCIs to
handle any classes this can result in a significant time saving on start up.
Mark