On 14.12.2012 22:12, Branko Čibej wrote:
> On 14.12.2012 21:28, Ben Reser wrote:
>> The odd part there is that it's actually the first --non-interactive
>> checkout that's failing in that test. 
> Sigh. I keep forgetting that I configure with
>
>     --disable-plaintext-password-storage.
>
> The test depends on storing the credentials from one command to the
> next, and since --non-interactive implies it can't even ask about
> storing the password in the keychain, it eventually fails.
>
> The test should probably be split in two parts, one that does and one
> that doesn't depend on that configure option, so that we can
> conditioinally skip the failing part.

For now, I fixed this in r1422171 by adding a conditioinal @XFail
decorator to the test. The actual predicate implementation is not
perfect, but it'll do until something better comes along.


-- 
Branko Čibej
Director of Subversion | WANdisco | www.wandisco.com

Reply via email to