5 Aug
2010
5 Aug
'10
6:09 p.m.
On Tue, Aug 3, 2010 at 12:30 PM, Vladimir Prus <vladimir@codesourcery.com> wrote:
Actually, as far as Boost.Build tests are concerned, there's no particular emergency. Historically, those tests tend to care a lot about testing environment, so one better look only on gcc and msvc test results. Those reveal that 'print' test is failing everywhere -- which is something I've missed, but is not important.
Expected failures tend to hide real failures, and make it hard to tell at a glace if a library has problems that need to be addressed. That's why we have the markup mechanism. So please either fix the failures or mark them up as expected. Thanks, --Beman