Problem building Tika with the latest POI (3.7)

[Li Leon <le...@gmail.com>]

Hi,


I have downloaded the latest POI(3.7 containing bug fix) from SVN and have a
successful build, but stuck on how to integrate this into Tika.

I hosted a local repository and put new POI jars[1] into their separate
directories according to modified tika-parsers\pom.xml
(<poi.version>3.7</poi.version>) content. I can see that tika-parsers
referred to the wanted local POI when built, but the tests failed[2] in the
end.

Any ideas on how to update the test or ignore the testing procedure as a
workaround?

[1]:
poi-ooxml-schemas-3.7.jar
poi-3.7.jar
poi-ooxml-3.7.jar
poi-scratchpad-3.7.jar

[2]:
Failed tests:
  testExcel(org.apache.tika.parser.microsoft.ooxml.OOXMLParserTest)
  testWord(org.apache.tika.parser.microsoft.ooxml.OOXMLParserTest)
  testPowerPoint(org.apache.tika.parser.microsoft.ooxml.OOXMLParserTest)



Thanks,

Re: Problem building Tika with the latest POI (3.7)

[Li Leon <le...@gmail.com>]

Sorry if this is confusing.

I built a POI 3.7(has bug fix) myself and was able to build Tika against it
from local repository.
The only problem I have during building of Tika was that tika-parsers test
would fail, so I executed  "mvn install -Dmaven.test.skip=true" and
succeeded in building parser.

But when I tried to "mvn install -Dmaven.test.skip=true" under Tika root dir
only 8M+of tika-app-0.5.jar generated which was different than just "mvn
install" yet they both seemed to function well.

So I just wonder what has been left out from this smaller jar, test cases or
all the external dependent jars? If the latter, how it runs differently
compare to jar generated after "mvn install".
I hope I made myself clear.

Thanks,

2009/12/18 Jukka Zitting <ju...@gmail.com>

> Hi,
>
> On Fri, Dec 18, 2009 at 5:00 AM, Li Leon <le...@gmail.com> wrote:
> > I was able to skip the tika-parsers test and have a successful build by
> > executing "mvn install -Dmaven.test.skip=true" in the root directory. But
> > this produced smaller 8M+ tika-app-0.5.jar.
>
> I don't think I understand what you're trying to do. If you want POI
> 3.6 support (there is no POI 3.7), check the Tika trunk where we've
> already done the upgrade.
>
> BR,
>
> Jukka Zitting
>

Re: Problem building Tika with the latest POI (3.7)

[Jukka Zitting <ju...@gmail.com>]

Hi,

On Fri, Dec 18, 2009 at 5:00 AM, Li Leon <le...@gmail.com> wrote:
> I was able to skip the tika-parsers test and have a successful build by
> executing "mvn install -Dmaven.test.skip=true" in the root directory. But
> this produced smaller 8M+ tika-app-0.5.jar.

I don't think I understand what you're trying to do. If you want POI
3.6 support (there is no POI 3.7), check the Tika trunk where we've
already done the upgrade.

BR,

Jukka Zitting

Re: Problem building Tika with the latest POI (3.7)

[Li Leon <le...@gmail.com>]

I was able to skip the tika-parsers test and have a successful build by
executing "mvn install -Dmaven.test.skip=true" in the root directory. But
this produced smaller 8M+ tika-app-0.5.jar.

So what has been left off from this jar, test cases or all the external
dependent jars? If the latter, how it runs differently compare to jar
generated after "mvn install".


Thanks,

2009/12/17 Li Leon <le...@gmail.com>

> Hi,
>
>
> I have downloaded the latest POI(3.7 containing bug fix) from SVN and have
> a successful build, but stuck on how to integrate this into Tika.
>
> I hosted a local repository and put new POI jars[1] into their separate
> directories according to modified tika-parsers\pom.xml
> (<poi.version>3.7</poi.version>) content. I can see that tika-parsers
> referred to the wanted local POI when built, but the tests failed[2] in the
> end.
>
> Any ideas on how to update the test or ignore the testing procedure as a
> workaround?
>
> [1]:
> poi-ooxml-schemas-3.7.jar
> poi-3.7.jar
> poi-ooxml-3.7.jar
> poi-scratchpad-3.7.jar
>
> [2]:
> Failed tests:
>   testExcel(org.apache.tika.parser.microsoft.ooxml.OOXMLParserTest)
>   testWord(org.apache.tika.parser.microsoft.ooxml.OOXMLParserTest)
>   testPowerPoint(org.apache.tika.parser.microsoft.ooxml.OOXMLParserTest)
>
>
>
> Thanks,
>