On Fri, Apr 6, 2012 at 8:17 PM, Greg Stein <gstein_at_gmail.com> wrote:
> On Fri, Apr 6, 2012 at 17:36, Hyrum K Wright <hyrum.wright_at_wandisco.com> wrote:
>> On Fri, Apr 6, 2012 at 4:31 PM, Greg Stein <gstein_at_gmail.com> wrote:
>>> On Fri, Apr 6, 2012 at 16:53, Hyrum K Wright <hyrum.wright_at_wandisco.com> wrote:
>>>>...
>>>> Just so you know, I would have expected this section to crash Ev2 in
>>>> copy test 67, since there are several directories being added, but the
>>>> children array is (currently) empty. I wonder if there is a problem
>>>> with the internal Ev2 state checking.
>>>
>>> I'm suspicious that the checks are not working because of the "pass
>>> URLs as relpath" that you're doing.
>>>
>>> I should add some canonical assertions for all the relpath params, and
>>> get you to fix that bug :-)
>>
>> Go for it! You're welcome to use the branch, but it'd probably be
>> best for these checks to end up on trunk, either originally, or via a
>> cherrypick merge from the branch.
>
> r1310655 on trunk. With shims activated: 339 failures now :-)
Do the failures happen on trunk or the branch? (I suspect the branch
as well as the trunk, knowing you. :P )
> I only added relpath assertions. I did not write the code to check for
> improper parent/child addition order. I'll do that in a second commit.
>
>> I plan on fixing the various ### comments in copy.c on the branch, as
>> those are issues I've identified with the current code. It's
>> unfortunate that we don't have tests which test these cases. :(
>
> Sounds like some new tests are going to come along :-P
Thanks for volunteering. :)
-Hyrum
--
uberSVN: Apache Subversion Made Easy
http://www.uberSVN.com/
Received on 2012-04-07 05:30:23 CEST