Matt, what's the contents of your define_index block for your model?  
It seems like Sphinx can access the results, but ActiveRecord can't...

Also, in your test, you're using some custom methods: search,  
delta_index_search, etc. Can you put those in a gist or pastie so I  
can see how you're searching for the expected results?

-- 
Pat

On 22/05/2009, at 8:51 PM, matt wrote:

>
> Thanks for the reply, but this isn't the solution this time.  My path
> is actually the same on both systems (installed sphinx from tarball on
> mac), and is in the PATH on both, so both OSX/Linux are able to find
> the binaries correctly.  All my other tests pass, and doing a full
> index gives the right behavior (empty response set), but doing a delta
> index on OSX gives me a single nil response for the search instead of
> no responses.  Does anyone know what th right behavior is?  Should I
> be getting an empty response set, or should I be getting a respnse set
> containing nil?
>
> Matt
>
> On May 20, 5:26 pm, Thuva Tharma <[email protected]> wrote:
>> I had a similar problem with delta indexing. I develop on Mac OS X  
>> and
>> deploy on Ubuntu. The problem was that the path to sphinx binaries
>> were different in both systems.
>>
>> In Mac OS X, the path is /opt/local/bin (I used Macports to install
>> Sphinx). In Ubuntu, the path is /usr/local/bin.
>>
>> Setting the bin_path appropriately in config/sphinx.yml solved the
>> issue. Here is how my sphinx.yml file looks like:
>>
>> development:
>>   bin_path: /opt/local/bin
>>
>> production:
>>   bin_path: /usr/local/bin
>>
>> I hope that would fix your problem.
>>
>> -- Thuva Tharma
>>
>> On May 20, 12:35 pm, matt <[email protected]> wrote:
>>
>>> I have TS all setup and working correctly.  I also have a suite of
>>> integration tests that test my assumptions about how sphinx  
>>> works.  On
>>> OS X, these integration tests succeed, but on our CI server (ubuntu
>>> linux), they fail.  The difference being in how delta indexes get
>>> treated.  Basically, the my test (see below) is verifying that  
>>> changes
>>> don't show up in the search until after a delta index is run.  On  
>>> OSX,
>>> the behavior seems to be that post destroy _and_ delta, the result  
>>> set
>>> has one item that is nil, while on linux, the result set has no
>>> items.  The linux variant seems to be the correct one, and thats  
>>> what
>>> we deploy to, so I'll change my test to match, but I'd like to be  
>>> able
>>> to have successful tests on OSX as well.
>>
>>> I'm not sure how to debug this further. Searchd itself is logging as
>>> below on OSX when I run the test.  Note that it is returning a count
>>> of (1) for the second to last request.  On linux this value is 0:
>>
>>> [Wed May 20 12:30:47.584 2009] 0.025 sec [all/5/rel 0 (0,20)] [*]
>>> deltapicture
>>> [Wed May 20 12:30:47.725 2009] 0.000 sec [all/5/rel 0 (0,20)] [*]
>>> deltapicture
>>> [Wed May 20 12:30:49.482 2009] 0.023 sec [all/5/rel 1 (0,20)] [*]
>>> deltapicture
>>> [Wed May 20 12:30:49.560 2009] 0.000 sec [all/5/rel 1 (0,20)] [*]
>>> deltapicture
>>> [Wed May 20 12:30:51.235 2009] 0.000 sec [all/5/rel 1 (0,20)] [*]
>>> deltapicture
>>> [Wed May 20 12:30:52.977 2009] 0.001 sec [all/5/rel 0 (0,20)] [*]
>>> deltapicture
>>
>>>   def test_delta_indexing
>>>     title = "deltapicture"
>>>     assert_equal 0, search(Picture, title).size
>>>     pic = create_picture(:title => title)
>>>     assert_equal 0, search(Picture, title).size
>>
>>>     self.class.delta_index_search
>>
>>>     results = search(Picture, title)
>>>     assert_equal 1, results.size
>>>     assert_equal pic.id, results.first.id
>>
>>>     pic.destroy
>>>     results = search(Picture, title)
>>>     assert_equal 1, results.size
>>>     assert_nil results.first
>>
>>>     self.class.delta_index_search
>>
>>>     results = search(Picture, title)
>>>     assert_equal 1, results.size
>>>     assert_nil results.first
>>
>>>     self.class.index_search
>>
>>>     assert_equal 0, search(Picture, title).size
>>>   end
> >


--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups 
"Thinking Sphinx" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to 
[email protected]
For more options, visit this group at 
http://groups.google.com/group/thinking-sphinx?hl=en
-~----------~----~----~----~------~----~------~--~---

Reply via email to