Alisdair Meredith wrote: > I am currently doing a search for other places where borland v 0x0561 is > assumed, as I don't think the latest patch fixed any issues that would > affect boost and it would be a shame to have to choose between boost and > the patch.
OK, borland 0x561 is assumed in quite a few spirit files (spirit.dev list informed), quite a lot of the MPL (around 20 files) and type_traits/add_reference.hpp. Unfortunately we are still certifying that patch here, and the PC it is on does not have easy access to Boost CVS, I cannot test in a reasonable frame for 1_30 (which I assume means testing today at the latest.) However, there are no issues addressed in that patch that would resolve boost workarounds that I am aware of, and suggest the safest solution might be to simply update the version check to 0x564 in all those files. At the worst we provide an extra unneeded workaround. The alternative is that these libraries, or libraries that rely on them, are quite likely to fail with the latest borland compiler. Of course, it is very easy to ask for sweeping changes right on the line when you aren't the release manager! If there is time I would very much like to push for the change though. On related note for AFTER the release, it would be useful if anyone on the list with access to the Kylix compiler could see how many of these tests should in fact be for 0x570. It is noticable that Kylix fails 50% more tests than the Windows compiler, but I do not have a Linux box to test with (yet!) -- AlisdairM _______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
