Menu
▾
▴
Re: [Exist-development] [Exist-open] Help needed with a test case
|
From: Stanislav <st...@gm...> - 2018-10-12 11:49:57
|
Attached you will find a zip file containing the results (as printed on
stdin and stderr) of running the tests (by `./build.sh test` )
with OpenJDK10 (or to be more precise OpenJDK 64-Bit Server VM 18.3 (build
10.0.2+13, mixed mode))
(Ok, you won't find it attached, as the exist-open mail list does not allow
posting mails containing zip attachments)
In the test log, there are a lot of warnings of the type:
[junit] WARNING: An illegal reflective access operation has occurred
Specifically, the test that seemed to fail in AppVeyor reports:
[junit] Running org.exist.fluent.QueryServiceTest
[junit] Tests run: 19, Failures: 0, Errors: 0, Skipped: 0, Time
elapsed: 6.905 sec
In particular, this means I can't reproduce what's bothering AppVeyor in
its OpenJDK10 build tests.
Can somebody else reproduce that problem?
I will push a dummy commit to start another IC test cycle.
Kindly,
/Stanislav
On Wed, Oct 10, 2018 at 9:47 AM Stanislav <st...@gm...> wrote:
> I need to install OpenJdk10.
> Later today, perhaps
>
> On Wed, Oct 10, 2018, 6:38 AM Adam Retter <ad...@ex...> wrote:
>
>> It looks to me from the AppVeyor log that the test [junit] Running
>> org.exist.fluent.QueryServiceTest
>>
>> Is failing with an error. Can you test with OpenJdk10 locally on Linux?
>>
>>
>> On Wed, 10 Oct 2018, 08:30 Stanislav, <st...@gm...> wrote:
>>
>>> Mystery of all mysteries:
>>>
>>> All the integration tests pass now with the single exception of the
>>> Java10 OpenJDK amd64 build which failed.
>>> The message is not too informative, it only states:
>>> BUILD FAILED
>>> 3610
>>> <https://ci.appveyor.com/project/AdamRetter/exist/builds/19381496/job/gtukpcxvq0wfftrs#L3610>/home/appveyor/projects/exist/build/scripts/junit.xml:329:
>>> The following error occurred while executing this line:
>>> 3611
>>> <https://ci.appveyor.com/project/AdamRetter/exist/builds/19381496/job/gtukpcxvq0wfftrs#L3611>/home/appveyor/projects/exist/build/scripts/extensions-build.xml:68:
>>> The following error occurred while executing this line:
>>> 3612
>>> <https://ci.appveyor.com/project/AdamRetter/exist/builds/19381496/job/gtukpcxvq0wfftrs#L3612>/home/appveyor/projects/exist/build/scripts/extensions-common.xml:339:
>>> Test org.exist.fluent.QueryServiceTest failed
>>> 3613
>>> <https://ci.appveyor.com/project/AdamRetter/exist/builds/19381496/job/gtukpcxvq0wfftrs#L3613>
>>> 3614
>>> <https://ci.appveyor.com/project/AdamRetter/exist/builds/19381496/job/gtukpcxvq0wfftrs#L3614>Total
>>> time: 22 minutes 40 seconds
>>> 3615
>>> <https://ci.appveyor.com/project/AdamRetter/exist/builds/19381496/job/gtukpcxvq0wfftrs#L3615>Command
>>> exited with code 1
>>> 3616
>>> <https://ci.appveyor.com/project/AdamRetter/exist/builds/19381496/job/gtukpcxvq0wfftrs#L3616>Build
>>> failed
>>>
>>>
>>> On Tue, Oct 9, 2018 at 10:25 PM Stanislav <st...@gm...> wrote:
>>>
>>>> Hi Adam,
>>>>
>>>> Thanks for your reply, but I have figured out what you have said myself.
>>>> And nevertheless, the test was failing.
>>>> So, after much ado about nothing (read this 'debugging my and also
>>>> older tests')
>>>> I've figured out (by the grace of God or some other deity that is
>>>> inclined to compassionately aid developers lost in legacy code)
>>>> that the problematic line:
>>>>
>>>> checkIndex(docs, broker, new QName[]{new QName("title")}, "Buick", 1);
>>>>
>>>> miraculously becomes un-problematic if only I convert "Buick" to all small letters - "buick"
>>>>
>>>> 2.5 hours lost. Case closed. Working TC submitted.
>>>>
>>>> :)
>>>>
>>>>
>>>> On Mon, Oct 8, 2018 at 10:45 PM Adam Retter <ad...@ex...> wrote:
>>>>
>>>>> Hi Stanislav,
>>>>>
>>>>> I didn't implement the Lucene index in eXist-db, that would have been
>>>>> Wolfgang Meier.
>>>>>
>>>>> However to try and help you, I just studied that code. It looks to me
>>>>> that the LuceneIndexTest#checkIndex function scans over the entire index,
>>>>> capturing each occurence of one or more provided QNames in the `qn`
>>>>> parameter and then returns the total number of occurrences.
>>>>>
>>>>> Hope that is helpful?
>>>>>
>>>>> On Mon, 8 Oct 2018 at 16:51, Stanislav <st...@gm...> wrote:
>>>>>
>>>>>> Ok,
>>>>>> may be this is asking for too much.
>>>>>> How about something simpler then -- could someone please add
>>>>>> meaningful documentation to the
>>>>>>
>>>>>> private Occurrences[] checkIndex(final DocumentSet docs, final DBBroker broker, final QName[] qn, final String term, final int expected)
>>>>>>
>>>>>> method in extensions/indexes/lucene/test/src/org/exist/indexing/lucene/LuceneIndexTest.java ?
>>>>>>
>>>>>> /St.
>>>>>>
>>>>>>
>>>>>> On Fri, Oct 5, 2018 at 5:22 PM Stanislav <st...@gm...> wrote:
>>>>>>
>>>>>>> Ok,
>>>>>>>
>>>>>>> I've crafted a basic test case and it does not work (not sure if it
>>>>>>> is the test that is wrong or the code tested that is wrong)
>>>>>>> Could someone knowledgeable in LuceneIndexTest(s) take a look at the
>>>>>>> test that I just added and tell me why it fails:
>>>>>>>
>>>>>>>
>>>>>>> https://github.com/eXist-db/exist/pull/2169/commits/39896b258b34571e4e264c46e4759b6ed14292ce
>>>>>>>
>>>>>>> /St.
>>>>>>>
>>>>>>> On Thu, Oct 4, 2018 at 6:43 PM Stanislav <st...@gm...> wrote:
>>>>>>>
>>>>>>>> Exactly - as Dom has explained it.
>>>>>>>> (Excuse me for not making myself clearer)
>>>>>>>>
>>>>>>>> /Stanislav
>>>>>>>>
>>>>>>>> On Thu, 4 Oct 2018 at 12:43, Dominic Latham <
>>>>>>>> dom...@gm...> wrote:
>>>>>>>>
>>>>>>>>> Dear Adam,
>>>>>>>>>
>>>>>>>>> Thanks for your help.
>>>>>>>>>
>>>>>>>>> The comments in the TEI express the tests
>>>>>>>>> e.g.
>>>>>>>>> <!-- this should get indexed -->
>>>>>>>>> <!-- this should not get indexed -- attribute name ns does not
>>>>>>>>> match -->
>>>>>>>>>
>>>>>>>>> I think the question is how to express that as a test in exist
>>>>>>>>> i.e. that a particular text node has been indexed.
>>>>>>>>> Does that make sense?
>>>>>>>>>
>>>>>>>>> With best wishes,
>>>>>>>>>
>>>>>>>>> Dominic
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> On Wed, 3 Oct 2018 at 09:43, Adam Retter <ad...@ex...>
>>>>>>>>> wrote:
>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> Erm... you don't say what you are trying to test exactly?
>>>>>>>>>>
>>>>>>>>>> On Wed, 3 Oct 2018 at 12:59, Stanislav <st...@gm...> wrote:
>>>>>>>>>>
>>>>>>>>>>> Hi guys,
>>>>>>>>>>>
>>>>>>>>>>> A test case is required for the code that implements:
>>>>>>>>>>> "Allow more elaborate XPath expressions in the Lucene index spec
>>>>>>>>>>> in collection.xconf" (#2169)
>>>>>>>>>>>
>>>>>>>>>>> and I have trouble understanding how existing indexing tests do
>>>>>>>>>>> work (and I am a bit tired of reverse-engineering)
>>>>>>>>>>> It would be great if someone familiar with the matter could give
>>>>>>>>>>> me a hand here:
>>>>>>>>>>>
>>>>>>>>>>> *The collection.xconf should be like this:*
>>>>>>>>>>> <collection xmlns="http://exist-db.org/collection-config/1.0">
>>>>>>>>>>> <index xmlns:tei="http://www.tei-c.org/ns/1.0">
>>>>>>>>>>> <!-- Lucene indexes -->
>>>>>>>>>>> <lucene diacritics="no">
>>>>>>>>>>> <analyzer
>>>>>>>>>>> class="org.apache.lucene.analysis.standard.StandardAnalyzer"/>
>>>>>>>>>>> <text field="title"
>>>>>>>>>>> match="//tei:title[@xml:lang='Sa-Ltn']" boost="2.0"/>
>>>>>>>>>>> </lucene>
>>>>>>>>>>> </index>
>>>>>>>>>>> </collection>
>>>>>>>>>>>
>>>>>>>>>>> *While the document being indexed should be like this:*
>>>>>>>>>>> <?xml-model href="../schema/1.0/tei.rng" schematypens="
>>>>>>>>>>> http://relaxng.org/ns/structure/1.0"?>
>>>>>>>>>>> <TEI xmlns="http://www.tei-c.org/ns/1.0">
>>>>>>>>>>> <teiHeader>
>>>>>>>>>>> <title type="longTitle" xml:lang="Sa-Ltn">Buick</title>
>>>>>>>>>>> <!-- this should get indexed -->
>>>>>>>>>>> <title type="longTitle"
>>>>>>>>>>> lang="Sa-Ltn">Cadillac</title> <!-- this should not get indexed --
>>>>>>>>>>> attribute name ns does not match -->
>>>>>>>>>>> <title type="longTitle" xml:lang="En"> Ford</title>
>>>>>>>>>>> <!-- this should not get indexed -- attribute value does not match -->
>>>>>>>>>>> <title type="longTitle">
>>>>>>>>>>> Dodge </title> <!-- this should not get indexed -- attribute is entirely
>>>>>>>>>>> missing -->
>>>>>>>>>>> </teiHeader>
>>>>>>>>>>> </TEI>
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> Kind regards,
>>>>>>>>>>> /St.
>>>>>>>>>>>
>>>>>>>>>>> _______________________________________________
>>>>>>>>>>> Exist-development mailing list
>>>>>>>>>>> Exi...@li...
>>>>>>>>>>> https://lists.sourceforge.net/lists/listinfo/exist-development
>>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> --
>>>>>>>>>> Adam Retter
>>>>>>>>>>
>>>>>>>>>> eXist Core Developer
>>>>>>>>>> { United Kingdom / United States }
>>>>>>>>>> ad...@ex...
>>>>>>>>>>
>>>>>>>>> _______________________________________________
>>>>>>>>>> Exist-open mailing list
>>>>>>>>>> Exi...@li...
>>>>>>>>>> https://lists.sourceforge.net/lists/listinfo/exist-open
>>>>>>>>>>
>>>>>>>>> _______________________________________________
>>>>>> Exist-development mailing list
>>>>>> Exi...@li...
>>>>>> https://lists.sourceforge.net/lists/listinfo/exist-development
>>>>>>
>>>>>
>>>>>
>>>>> --
>>>>> Adam Retter
>>>>>
>>>>> eXist Core Developer
>>>>> { United Kingdom / United States }
>>>>> ad...@ex...
>>>>>
>>>> _______________________________________________
>>> Exist-development mailing list
>>> Exi...@li...
>>> https://lists.sourceforge.net/lists/listinfo/exist-development
>>>
>>
|
