
4 Jan
2011
4 Jan
'11
9:52 a.m.
Further experiments show that the issue is present in boost 1.42, but fixed in boost 1.45. Sorry for the noise :)
It's an Intel bug in any case - or at least a result of using the Intel compiler with an unsupported GCC version - just including <cmath> is often enough to trigger it :-( I can't remember the issue number, but I reported it to Intel, and it's been fixed in their most recent compiler releases (as well as a workaround in our code). HTH, John.