[mapguide-internals] Re: PATCH: GwsQueryEngine chooses incorrect join algorithm for joins involving SDF/SHP sources

[Jackie Ng]

It seems my submission broke one of the existing unit tests:
TestCase_JoinFeaturesChainedInner1ToMany

>From what I've debugged so far, the left query check when it is
eGwsNestedLoopSortedBlock or eGwsSortMerge fails to properly set
left-side-supports-ordering to true because one of the properties
(Autogenerated_SDF_ID) from the test FeatureSource is being checked against
a qualified version of that property (Join1.Autogenerated_SDF_ID). The
resulting failure, means that eGwsBatchSortedBlock becomes the selected
algorithm, which breaks the unit test because batch sorted block doesn't
support multiple joined iterators to it.

I've attached another patch for review, which strips all scope from the
property names being compared. This makes the unit test pass again, but I
would like confirmation of whether this is actually correct.

- Jackie

--
View this message in context: http://osgeo-org.1803224.n2.nabble.com/PATCH-GwsQueryEngine-chooses-incorrect-join-algorithm-for-joins-involving-SDF-SHP-sources-tp7093987p7099922.html
Sent from the MapGuide Internals mailing list archive at Nabble.com.