On Wednesday, July 12, 2017 at 11:55:17 AM UTC-7, Jeremy Evans wrote:
>
> On Wednesday, July 12, 2017 at 10:36:20 AM UTC-7, [email protected] 
> wrote:
>>
>>
>>
>> On Wednesday, July 12, 2017 at  UTC+3, Jeremy Evans wrote:
>>>
>>> Hi Jeremy, our admins gave me access to our dev servers, but specs fail 
>>>> even for yesterday's master.
>>>>
>>>> I've run test suite on three different servers with similar results, 
>>>> like
>>>>
>>>> 736 runs, 3442 assertions, 21 failures, 214 errors, 28 skips
>>>>
>>>> Not sure if it is issue with my setup.
>>>>  I've made a gist with logs from each run, 
>>>> https://gist.github.com/timon/7a4affab68e4b0b1cd3ec8d330c1b369
>>>>
>>>>
>>> I probably should have mentioned this earlier, but the oracle adapter 
>>> tests require the :autosequence option.  If you weren't using that before, 
>>> could you try rerunning the specs with that option and see if it fixes the 
>>> issue?  If that doesn't fix it, I'm assuming the failures are either due to 
>>> Oracle version or configuration differences.  Here's what I get when 
>>> running the specs using the older Oracle VM I currently test with (server 
>>> version 11020002):
>>>
>>> 759 runs, 10604 assertions, 0 failures, 0 errors, 28 skips
>>>
>>
>> Yes, that did the trick.
>> On Oracle 11 all tests pass.
>>
>> On Oracle 12, after I moved limit method before lock (having them after 
>> lock produced invalid queries), I have two failing tests, which I cannot 
>> figure out:
>> Sequel::DatabaseError: OCIError: ORA-02014: cannot select FOR UPDATE from 
>> view with DISTINCT, GROUP BY, etc.
>>
>> Lines that fail are spec/integration/dataset_test.rb:180 
>> and spec/integration/model_test.rb:235
>>
>
> With the current master branch, all tests not explicitly skipped pass in 
> both Oracle 11 and 12 (yay!).   I'll have to test your branch and see if I 
> can figure out what the issue is.  It's possible we won't be able to use 
> the new offset syntax in certain cases if it causes problems.
>

Oracle doesn't like that offset syntax when used with locking.  This 
appears to work around the problem and make all tests pass:

      def supports_fetch_next_rows?
        server_version >= 12000000 && !@opts[:skip_locked] && !@opts[:lock]
      end

Can you include that fix, rebase and squash your changes into a single 
commit, and then submit it as a pull request to jeremyevans/sequel?

Thanks,
Jeremy 

-- 
You received this message because you are subscribed to the Google Groups 
"sequel-talk" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To post to this group, send email to [email protected].
Visit this group at https://groups.google.com/group/sequel-talk.
For more options, visit https://groups.google.com/d/optout.

Reply via email to