[ 
https://issues.apache.org/jira/browse/UIMA-1302?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Marshall Schor closed UIMA-1302.
--------------------------------

       Resolution: Fixed
    Fix Version/s: 2.3

Should be fixed by fixes in linked issue.  If not, please re-open

>  PerformanceReports for PEAR not being generated in a CPE
> ---------------------------------------------------------
>
>                 Key: UIMA-1302
>                 URL: https://issues.apache.org/jira/browse/UIMA-1302
>             Project: UIMA
>          Issue Type: Bug
>          Components: Collection Processing
>    Affects Versions: 2.2.2
>            Reporter: Kevin Cunnane
>            Priority: Minor
>             Fix For: 2.3
>
>
> When an installed PEAR is included as an AE in a Collection Processing 
> Engine, the performance report generated by the CPE doesn't include any 
> figures for the PEAR. The only work-around is to install the pear and then 
> reference the AE descriptor directly, however this ruins the entire point of 
> using the PEAR - it automatically sets the classpath, datapath etc for you. I 
> discussed this issue with Thilo Goetz, who said I should raise a bug against 
> it. Please see correspondence below about this issue with Thilo.
> *My Message:*
> Hi Thilo,
> I work for the LanguageWare team in Dublin, and am having an issue getting 
> performance numbers for PEARs used in a Collection Processing Engine. 
> Essentially, when you refer to the PEAR using the pearSpecifier file, then 
> the performance of the PEAR is not included in the Performance Report. 
> However, if you run the PEAR by referring to its main descriptor directly 
> (and setting up the classpath etc manually) you will correctly get the 
> performance of the components in the PEAR included in the performance report.
> Am I doing something wrong, or is there a way to work around this? For 
> example, is this just an issue when using PEARs in a CPE, or can I get the 
> information some other way? I'm trying to set up an automated process for 
> testing the performance of a PEAR against a corpus, and would like to avoid 
> having to write code / scripts for this
> *His Responses:*
> Hi Kevin,
> that issue somehow seems familiar.  I've looked into those PerformanceReports 
> because of another bug I found, and I think I noticed this, too 
> (http://www.mail-archive.com/uima-dev%40incubator.apache.org/msg06565.html).  
> So I think this is a UIMA issue, due to the special nature of pear 
> descriptors (they are not regular AE descriptors).  So the only workaround I 
> can think of at this time is to not use pears, or install the pear and then 
> reference the AE descriptor directly, if that's possible.
> *Related Posts*
> I found this post http://markmail.org/message/doknwh5iecuhtai3 that seems to 
> refer to the exact same issue

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.

Reply via email to