Looks like some of the Windows failures have cleared up. We're still waiting for results on Windows 2000 but hopefully Farid's patch for STDCXX-543 will have resoled those as well.
Unfortunately, we continue to be having problems in Intel C++ (both 9.1 and 10.0) builds on Windows XP. It might be important that the failures are isolated to the servers xp64build (XP/EM64T Service Pack 1) and stein (XP/X86 Service Pack 2). The servers xp64build2 (also XP/EM64T Service Pack 1), box or mug (also XP/X86 Service Pack 2), don't seem to exhibit the problem in the latest builds. Farid, please let us know where you stand on this, if you've made any progress since the last time we discussed it and what we can do to help. For reference, AFAIK, the last update on this is here: http://www.nabble.com/Windows-build-failures-tf4379160.html#a12638734 Travis, can you please try to determine the difference between the "good" and the "bad" servers to see if we're missing patches or what else could be causing the failures? If there is a problem with the environment (incorrectly set up compilers, etc.), please let Andrew know so he can start working on getting it corrected. Thanks Martin