[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: Fwd: [PATCH] Fix a ruby test failure

From: Joe Swatosh <joe.swatosh_at_gmail.com>
Date: Fri, 19 Dec 2008 22:51:40 -0800

On Fri, Dec 19, 2008 at 6:37 AM, Hyrum K. Wright
<hyrum_wright_at_mail.utexas.edu> wrote:
> Joe Swatosh wrote:
>> On Thu, Dec 18, 2008 at 10:23 PM, Joe Swatosh <joe.swatosh_at_gmail.com> wrote:
>>> On Thu, Dec 18, 2008 at 10:15 PM, Hyrum K. Wright
>>> <hyrum_wright_at_mail.utexas.edu> wrote:
>>>> Joe Swatosh wrote:
>>>>> To the list this time. Doh.
>>>>>
>>>>>
>>>>> ---------- Forwarded message ----------
>>>>> From: Joe Swatosh <joe.swatosh_at_gmail.com>
>>>>> Date: Thu, Dec 18, 2008 at 9:06 PM
>>>>> Subject: Re: [PATCH] Fix a ruby test failure
>>>>> To: "Hyrum K. Wright" <hyrum_wright_at_mail.utexas.edu>
>>>>>
>>>>>
>>>>> Hi Hyrum,
>>>>>
>>>>> On Thu, Dec 18, 2008 at 12:40 PM, Hyrum K. Wright
>>>>> <hyrum_wright_at_mail.utexas.edu> wrote:
>>>>>> Kou, Joe,
>>>>>> This fixes a ruby test failure on trunk, some variation thereof fixes the
>>>>>> failure on 1.5.x. My ruby-test-swig-fu is pretty weak, so I'm posting here
>>>>>> before committing. If either of you could look at it, that'd be great.
>>>>>>
>>>>>> -Hyrum
>>>>>>
>>>>>> [[[
>>>>>> * subversion/bindings/swig/ruby/test/test_wc.rb
>>>>>> (test_relocate): Fix a ruby test expectation.
>>>>>> ]]]
>>>>>>
>>>>>> Index: subversion/bindings/swig/ruby/test/test_wc.rb
>>>>>> ===================================================================
>>>>>> --- subversion/bindings/swig/ruby/test/test_wc.rb (revision 34821)
>>>>>> +++ subversion/bindings/swig/ruby/test/test_wc.rb (working copy)
>>>>>> @@ -995,10 +995,14 @@
>>>>>> assert_equal([
>>>>>> [@fs.uuid, dir2_uri, nil],
>>>>>> [@fs.uuid, dir2_uri, dir2_uri],
>>>>>> - [nil, "#{dir2_uri}/#{dir1}", dir2_uri],
>>>>>> - [nil, "#{dir2_uri}/#{dir1}/#{file1}", dir2_uri],
>>>>>> - [nil, "#{dir2_uri}/#{dir2}", dir2_uri],
>>>>>> - [nil, "#{dir2_uri}/#{dir2}/#{file2}", dir2_uri],
>>>>>> + [@fs.uuid, "#{dir2_uri}/#{dir1}", nil],
>>>>>> + [@fs.uuid, "#{dir2_uri}/#{dir1}", dir2_uri],
>>>>>> + [@fs.uuid, "#{dir2_uri}/#{dir1}/#{file1}", nil],
>>>>>> + [@fs.uuid, "#{dir2_uri}/#{dir1}/#{file1}", dir2_uri],
>>>>>> + [@fs.uuid, "#{dir2_uri}/#{dir2}", nil],
>>>>>> + [@fs.uuid, "#{dir2_uri}/#{dir2}", dir2_uri],
>>>>>> + [@fs.uuid, "#{dir2_uri}/#{dir2}/#{file2}", nil],
>>>>>> + [@fs.uuid, "#{dir2_uri}/#{dir2}/#{file2}", dir2_uri],
>>>>>> ],
>>>>>> values)
>>>>>> assert(dir2_uri, access.entry(@wc_path).url)
>>>>>>
>>>>>>
>>>>> I've been trying to figure out when that test started failing and why
>>>>> the validate notification is now called so many more times than it
>>>>> used to be. I can only work on this for a few hours a week, so I
>>>>> don't know when I'll have answers to those questions. Personally, I
>>>>> prefer to include more info about why the test expectations need to
>>>>> change in the commit, but defer to your best judgement wrt committing
>>>>> this patch.
>>>> And my best judgment says to let somebody who knows something about the Ruby
>>>> bindings make the call. :) This was just an attempt to get the test suite
>>>> passing again, especially given the imminent release of 1.5.5. But if you or
>>>> Kou thinks the problem is more than skin deep, I'm happy to defer to your best
>>>> judgment.
>>>>
>>> Honestly, at this point I don't know that fixing the expectations
>>> isn't the best answer. All I know is the problem started on trunk
>>> somewhere between r33776 and r33810, and I'm done for tonight.
>>>
>>> G'nite
>>> --
>>> Joe
>>>
>>
>> Couldn't sleep. It was
>>
>> ------------------------------------------------------------------------
>> r33786 | rhuijben | 2008-10-20 11:49:51 -0700 (Mon, 20 Oct 2008) | 9 lines
>>
>> Fix info_tests 3: "info on new dir with mkdir". Copy the repository uuid
>> from the parent entry into the entries for new directories.
>>
>> * subversion/libsvn_wc/adm_ops.c
>> (svn_wc_add3): Copy uuid from the parent directory entry to a new
>> subdirectory if available.
>> * subversion/tests/cmdline/info_tests.py
>> (test_list): Remove XFail from info_on_mkdir.
>>
>> ------------------------------------------------------------------------
>>
>> Which to me explains why the UUIDs started showing up as the first
>> element of the inner arrays, but not why we get two notifications
>> about some of the elements: the first with a nil root_url, the second
>> with a root_url. Hmm, looks like that was the expected behaviour for
>> the first element in the original expectation.
>>
>> I think fixing the expectation is the reasonable thing to do. Ta for
>> your patience.
>
> Joe,
> Thanks for looking at this! I'll go ahead and commit the fix to the test
> expectation and nominate something similar for the 1.5.x branch. Sorry to keep
> you up!
>
> -Hyrum
>

This was only a tiny part of it :-) and a welcome distraction in the
middle of the night. Job stress is the real culprit :-( Gotta love
year end issues.

I'm glad I could help reduce one of the 1.5.5 issues.

Thanks for herdin' them cats for releases.

--
Joe
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=988048
Received on 2008-12-20 07:52:00 CET

This is an archived mail posted to the Subversion Dev mailing list.