
While working on getting icc to not choke on Serialization (the Boost test boxes indicated that icc was failing to build with the addition of my changes), I came across a very nasty bug the occurs when using icc with version 4.5 or higher of GNU's standard library - GNU's <iomanip> uses what are either illegal or C++0x semantics which icc doesn't support. Serialization uses <iomanip>, and icc seems to select the latest version of libstd++ installed on a Linux machine as it's default standard library (I have been unable to find an Intel standard library - I'm assuming such a thing doesn't exist).
When using an older version of libstd++, icc + Serialization compiled fine. I removed all uses of IO parameterized manipulators in Serialization (there were only maybe half a dozen cases)*, and got icc to compile Serialization with libstd++ v4.5.
Oh, Intel's compiler isn't supported with gcc-4.5, period. Basically you shouldn't use it with anything except 4.4 or lower - there should be an installation option to control which gcc version gets picked up, but it's such a while since I had to do that I can't remember how it all works :-(
Is there any chance the linux/darwin Intel build bots can be set up to use libstd++ v4.4, if they're not already using it? Intel is apparently aware of this issue, but it won't be fixed until their next major release (http://software.intel.com/en-us/forums/showthread.php?t=74691).
You can always find out what compiler and library versions are in use by going to the Boost.Config test results and clicking on the config_info results for the test runner you're interested in, for example: http://tinyurl.com/2ujk8td indicates that Darwin Intel-11.1 is on top of gcc 4.0.1.
If the build bots aren't using v4.5 of GNU's standard library, then I'm at a loss as to why Serialization is failing to build on them. My only other thought is that the timeout for the build cycle is too low - the failures indicate that the error is a timeout after 300 seconds.
I see what you mean... not very helpful! The best you can do is contact the test runner for more info. As a general note, the Intel-Darwin tests seem to have more than their fair share of unexpected/unexplained failures, I can only assume that Intel's support for Darwin is a lot less mature than for Windows/Linux.
On a more general note, is Intel on Linux/Darwin a "supported" Boost compiler?
Issue with core "supported" compiles are shown on the issues page: http://beta.boost.org/development/tests/trunk/developer/issues.html, there are no Intel-Darwin failures there, just the VC7.1 ones. HTH, John.