I have the same problem for a long time. The next time I try to make the bjam V2 work on my setup I would like to do the following: In my case its wide character support (also spirit support). Here is what I plan to do. a) make a small compile only test which passes/fails depending on the existence of wide character support. b) make all tests which are relevant to wide char support dependent on the the test in a) above. c) My test matrix display program leaves blanks in cases where there are no test results - which is just fine by me. i have the same issues regarding support for spirt and dlls which are optional. In fact, I'm even wondering if I should not bother testing which depends on other tests passing. For example. Variant serialization depends upon a working variant implemenation. So the the serialzation tests for variant can be skipped if the more basic tests of the variant library fail. Of course for this to work in a smooth way - the variant serialization should be moved from the serialization library to the variant library where it belongs. But that's another problem. Robert Ramey Peter Dimov wrote:

Does our testing infrastructure allow a test to return a "skipped/not applicable" status, that is, to inform the test script that the current test requires optional features that are not supported by the current compiler?

I'm thinking specifically of move - rvalue reference - support.

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost