[hibernate-dev] Processing mapping information followup

[Steve Ebersole]

What I am discussing here has no effect on this concern.

In the new terminology, what we are discussing is the process for 
handling "metadata sources" (o.h.metamodel.source).  What you describe 
is really a parallel source (o.h.metamodel.source.jdbc???).  So it is 
going to be completely up to the developer of that code how the binding 
of source information to metamodel works.

We will strive to make sure nothing requires the domain classes to be 
present.  However, please not that in developing this new code I made a 
distinction.  Specifically, I believe this principal does *not* extend 
to types, identifier generators, etc.  Entity classes, component 
classes.. anything that is actually part of the domain model though we 
will strive to make not required to build this metadmodel.  You 
specifically asked about "model classes and types"; just wanted to 
highlight the difference in my mind.

However, keep in mind that this is a completely new piece of code.  I 
would be shocked if we get it completely right out of the gate.  So just 
let us know where we missed stuff.


On 06/14/2011 05:40 AM, Max Rydahl Andersen wrote:
> Any idea how much this would affect tools ability to reflect on the metamodel without requiring the actual
> model classes and types to be available ?
>
> At what time will the classes/types be *required* to be available?
>
> i.e. for H3 as long as we filled out all the type info in the xml hibernate did not require
> access to types or model classes until buildMappings() occurred.
>
> Allowing us to iterate the mapped classes and tables for code generation.
>
> /max
>
> On Jun 13, 2011, at 23:29, Steve Ebersole wrote:
>
>> Wanted to start a follow up discussion to the conversation we had at the
>> IRC meeting on 6/13 with regards to changing the way we process mapping
>> information to follow dependencies in the various types of information.
>>
>> What we do currently (in the metamodel code) is essentially the same as
>> the legacy code.  That is to say we process through a mapping source in
>> full and then move on to the next one, saving off "second passes"
>> whenever we encounter information on which we have to wait.  The second
>> passes in the new metamodel code are isolated I guess (so far) to
>> org.hibernate.metamodel.source.internal.EntityReferenceResolver
>>
>> What I proposed instead is quite different, more like what I did for the
>> rest of the MetadataImpl constructor code in terms of defining
>> dependencies and processing stuff in an order that makes sure to avoid
>> the need for second passes.
>>
>> Here is the initial swab at the levels of information:
>> 1) entity - basic entity information such as name, hierarchy
>> 2) identifiers - normal identifiers could almost be handled in the first
>> level.  key-many-to-one still needs some form of delayed queuing.
>> 3) secondary tables
>> 4) attributes
>> 5) associations
>>
>> The dependencies flow downward; (2) depends on (1); (3) depends on (2); etc.
>>
>> I liked handling (3) secondary tables separately because they basically
>> expand the possible set of columns available for (4) and (5).
>>
>> --
>> Steve Ebersole<steve at hibernate.org>
>> http://hibernate.org
>> _______________________________________________
>> hibernate-dev mailing list
>> hibernate-dev at lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/hibernate-dev
>
> /max
> http://about.me/maxandersen
>
>
>

-- 
Steve Ebersole <steve at hibernate.org>
http://hibernate.org