Re: What is "XPASS"?

Greg Ward wrote:
> I just tried to merge recent upstream work (ie. svntest + cvs2svn
> trunk) with my cvs2hg repository.  Everything went fine except now one
> test case ("generate a manpage for cvs2hg") fails with "XPASS".  Huh?
> That's a new one; is that a recent addition to svntest?
> 
> The really odd thing is that:
>  * cvs2hg's man-page-generating code is nearly identical to that of
> cvs2svn and cvs2git
>    (the 3 scripts are nearly identical)
>  * the behaviour is the same (write ~11000-14000 bytes to stdout,
> nothing to stderr, exit status 0)
>  * the test functions are identical
> 
> So, any clue what XPASS is, and how to track down why a test would
> fail with XPASS?

There are four possible test results:

PASS - worked

FAIL - expected to work, but didn't

XFAIL - known not to work and indeed did not work.  This is not
considered a failure of the test suite, as XFail is used to mark tests
that have been written for features that are not yet implemented.

XPASS - expected not to work, but mysteriously worked.  This is
considered a "failure" of the test suite because it indicates that human
attention is needed to mark the test as "should pass".

I marked the cvs2hg_manpage test as XFail in r4957 of the SVN trunk
because it wasn't working.  In the commit that fixes the test, you
should also remove the "XFail()" marker on the test in the test_list
array at the bottom of run-tests.py.

Michael

------------------------------------------------------
http://cvs2svn.tigris.org/ds/viewMessage.do?dsForumId=1667&dsMessageId=2430309

To unsubscribe from this discussion, e-mail: [[email protected]].