Struggling with TCK20

[Craig Russell <Cr...@Sun.COM>]

Javadogs,

I've made a bunch of changes to the tck20 to try to get JPOX to run, 
but have a few issues.

1. I can run maven runtck.jdorisingle on a number of tests that when 
run under maven runtck.jdori fail after a while. Is there a difference 
in the maven goal that might cause this difference?

2. I don't really like all of the <column name= length=/> that are due 
to our defining the schema instead of letting JPOX define the schema. 
Should we just let JPOX define the schema and be done with it?

3. There are a few failures due to the enhancer not enhancing any of 
the persistence-aware classes. I can't seem to tell the JPOX enhancer 
to enhance them.

4. I've only been working on the application identity, so haven't done 
much with the schema2 or the datastoreidentity .jdo files. They are 
included just for reference.

5. I haven't checked in the attached files because I'm not really sure 
that they work.

Craig

Re: Struggling with TCK20

[Michelle Caisse <Mi...@Sun.COM>]

Hi, Craig,

A belated response to your question 1:  There are cases where test A 
leaves something in the database and tests B, C, D... attempt to clean 
it up and fail.  If tests B, C, or D are run alone, they pass.  There 
are two issues here, the stuff left behind and the failure to clean it 
up.  The latter is the subject of the recent thread "pmf not closed".

-- Michelle

Craig Russell wrote:

> Javadogs,
>
> I've made a bunch of changes to the tck20 to try to get JPOX to run, 
> but have a few issues.
>
> 1. I can run maven runtck.jdorisingle on a number of tests that when 
> run under maven runtck.jdori fail after a while. Is there a difference 
> in the maven goal that might cause this difference?
>
...

RE: Struggling with TCK20

[er...@jpox.org]

Craig and All,

>I've made a bunch of changes to the tck20 to try to get JPOX to run, 
>but have a few issues.

>1. I can run maven runtck.jdorisingle on a number of tests that when 
>run under maven runtck.jdori fail after a while. Is there a difference 
>in the maven goal that might cause this difference?

I tried the TCK using MSSQL, and besides of some failures/errors caused
by JPOX, mainly related to the company model, everything else runs
smoothly.

There is one issue running with Derby due to their limited number of
constraints (36). JPOX can create all the required constrains for a
table, but this is disabled by default, unless you enable
org.jpox.autoCreateSchema

For the the OutOfMemoryError, I guess it can be solved if run Derby in
another VM.

>2. I don't really like all of the <column name= length=/> that are due 
>to our defining the schema instead of letting JPOX define the schema. 
>Should we just let JPOX define the schema and be done with it?

For this issue, set org.jpox.validateTables=false.