From: Douglas Gregor (gregod_at_[hidden])
Date: 2002-01-23 23:15:36
On Wednesday 23 January 2002 09:53 pm, you wrote:
> One fairly common way to handle this sort of this is to have a list of
> tests that are expected to fail. This list is subtracted from the list of
> normal (real) failures for most display/report purposes. Developers and
> maintainers can add and remove tests from the expected fail list. I guess
> this is similar to option 3), but a bit simplified because the list of
> expected failures is created by hand instead of deduced from the config
> system or otherwise.
I guess my fear of this approach is because I can foresee this as the easy
way out. It's tough to get library X working on this compiler, so I'll just
expect everything to fail on this compiler and ignore it. While this seems
fair to the developer (in an ideal world, we shouldn't have to deal with
these broken compilers), it's horrible for our users.
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk