Greetings Martin I haven't had time to investigate this issue, but I suspect it may be fallout from the changes that were made to separate the processing logic within the static_export.pl script. The reordered logic takes longer to retrieve the index data from the database, and it's possible that the index data is updated during the retrieval process, resulting in the observed misbehavior. Table locking probably isn't option, as it likely would result in a stall for the rest of the automated test system.
--Andrew Black Martin Sebor wrote: > Andrew, I suspect recent changes to the static_export.pl script > are causing duplicate rows for some builds. For example, we have > two rows for a 15S build with VisualAge C++ 6.0 on AIX 5.2. There > are a number of other duplicates as well, including a few results > for Intel C++ 9.1 that appear in triplicate, and one for EDG eccp > is there four times. > > Thanks > Martin
