You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@pdfbox.apache.org by Tim Allison <ta...@apache.org> on 2022/09/07 21:18:32 UTC

timeout on Jempbox xmp media management schema's getHistory

All,
  This issue is ringing a bell.  I'm sorry if there's an open issue or
you/we've decided long ago that this is not an issue.
   One of the timeouts in the most recent run was caused by Jempbox's
handling of the history in the media management schema.  There are 32000
elements in the history. :(
   On the Tika side, we limit history elements to 1024.  However, Jempbox
still has to load the full list, and on this xmp, it takes a long, long
time (I was never patient enough to let it finish).
   If I do enough subclassing and limit getEventSequenceList to 1024, all
is good. [0]
   Is it worth spending time to fix the underlying performance issue in
Jempboxx, or is this type of hack on the Tika side the best option?

       Best,

             Tim

[0] int length = items.getLength();
length = 1024;
for(int i = 0; i < length; ++i) {

    Element li = (Element)items.item(i);
    retval.add(new ResourceEvent(li));
}

Re: timeout on Jempbox xmp media management schema's getHistory

Posted by Maruan Sahyoun <sa...@fileaffairs.de>.
What about opening a ticket and attaching the XMP in question 

BR
Maruan 

> Am 07.09.2022 um 23:19 schrieb Tim Allison <ta...@apache.org>:
> 
> All,
>  This issue is ringing a bell.  I'm sorry if there's an open issue or
> you/we've decided long ago that this is not an issue.
>   One of the timeouts in the most recent run was caused by Jempbox's
> handling of the history in the media management schema.  There are 32000
> elements in the history. :(
>   On the Tika side, we limit history elements to 1024.  However, Jempbox
> still has to load the full list, and on this xmp, it takes a long, long
> time (I was never patient enough to let it finish).
>   If I do enough subclassing and limit getEventSequenceList to 1024, all
> is good. [0]
>   Is it worth spending time to fix the underlying performance issue in
> Jempboxx, or is this type of hack on the Tika side the best option?
> 
>       Best,
> 
>             Tim
> 
> [0] int length = items.getLength();
> length = 1024;
> for(int i = 0; i < length; ++i) {
> 
>    Element li = (Element)items.item(i);
>    retval.add(new ResourceEvent(li));
> }

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@pdfbox.apache.org
For additional commands, e-mail: dev-help@pdfbox.apache.org