[Mondrian] checkJavadoc failure

John V. Sichi jsichi at gmail.com
Sat Sep 20 00:05:22 EDT 2008


Julian Hyde wrote:
>> ant test using Derby is failing with the error below.  I 
>> think you asked 
>> Ajit about this before, and I didn't see a response, so I'm going to 
>> disable the test unless there are objections.
> 
> Or fix it, if you can. Ajit is no longer an active contributor to the
> project so is not going to fix this. If we go around disabling tests the
> functionality will rot.

Problem is, he used HashSet/Map all over AggregateFunDef, so the SQL 
that gets generated is non-deterministic, and the test checks a number 
of different dialects, making maintenance painful (if I try to fix it 
with Tree or LinkedHash, megatest will tell me that I broke Teradata or 
something).

Most likely the test really wants to be doing an order-insensitive 
comparison on something upstream of the generated SQL instead.

JVS




More information about the Mondrian mailing list