At the moment, the only testing infrastructure issue affecting regression results that I'm aware of is intel-linux-9.0 results in Sandia's results. This is not a blocker, since the same toolset is present in Martin's results, and is all green. Other failures look more like problems with libraries, compilers and their interaction, and that's something only library authors can solve. We talked about setting two-week period for addressing remaining failures. Two weeks from now is March 1. I would appreciate if library authors look at failures in their libraries and try to have them fixed by that date. On March 1, I'll mark all failures that are present as expected. It would be reasonable to declare complete code freeze at the same date. When changing anything in code, please post a patch to the list, so that everybody knows what changes are being done. Also, please refrain from committing anything to RC branch that is neither fix for a reported failure, nor a doc-only change. Thanks, Volodya