These are *regression* tests. You can't just make them pass by changing them.
Let's see some rationale!! Why should the output change? And if it
does, then there better be some commentary about WHY that is. I see no
comments about why this is allowed to change. No explanation. No
The tests should produce the exact same result. That is why they are
there. To ensure we haven't buggered something up.
Any time the output is supposed to be different now, then we need a
full explanation on why that has happened. We may need to write an
errata. We may need to update documentation. Or more likely, we need
to fix some bugs.
Why are all these tests changing? This doesn't seem right.
On Tue, Aug 3, 2010 at 09:56, <philip_at_apache.org> wrote:
> Author: philip
> Date: Tue Aug 3 13:56:58 2010
> New Revision: 981885
> URL: http://svn.apache.org/viewvc?rev=981885&view=rev
> Make resolved_tests.py 1 pass in single-db.
> * subversion/tests/cmdline/switc_tests.py
> (resolved_on_wc_root): Tweak expectations for single-db.
> Modified: subversion/trunk/subversion/tests/cmdline/resolved_tests.py
> URL: http://svn.apache.org/viewvc/subversion/trunk/subversion/tests/cmdline/resolved_tests.py?rev=981885&r1=981884&r2=981885&view=diff
> --- subversion/trunk/subversion/tests/cmdline/resolved_tests.py (original)
> +++ subversion/trunk/subversion/tests/cmdline/resolved_tests.py Tue Aug 3 13:56:58 2010
> @@ -112,6 +112,8 @@ def resolved_on_wc_root(sbox):
> 'A/B/E/alpha', 'A/B/E/beta',
> + if svntest.main.wc_is_singledb(sbox.wc_dir):
> + expected_disk.remove('A/B/E', 'A/B/F', 'A/B')
> expected_status = svntest.actions.get_virginal_state(wc, 2)
> expected_status.tweak('iota', 'A/B', 'A/D/gamma',
Received on 2010-08-03 19:44:41 CEST