Re: --XFAIL--

Pierre Joye wrote:

On Sun, May 25, 2008 at 6:46 PM, Hannes Magnusson <hannes.magnusson@gmail.com> wrote:

On Sun, May 25, 2008 at 6:21 PM, zoe <zoe.slattery@googlemail.com> wrote:

    
In general the idea is to be able to mark a set of tests that you don't want

to worry about right now (for whatever reason) so that it makes it slightly easier to pick out unexpected failures.

And for this to work correctly, people have to remember to remove the XFAIL section when the bug is fixed, otherwise once the fix breaks the test will incorrectly end up again in the "expected failure" section.

    I wonder if a check like this is also required?

if (test contains XFAIL) $$ (test passes) { warn: The test that you expected to fail is passing }

Yes, that's the main idea behind this feature :-) Another nice addition (based on Hannes suggestion) would be to allow SKIPIF to detect a die("xfail") (it works with "skip" now) so one can add a conditional xfail. A good example is a test working with a given version of a library and fails with older versions (libxml has a couple of such cases as well as gd). Might be a little more work I think :-)

Thanks for your work! Cheers,