Tim Allison created PDFBOX-4370:
-----------------------------------
Summary: Jempbox's ResourceEvent crazily slow to initialize
Key: PDFBOX-4370
URL: https://issues.apache.org/jira/browse/PDFBOX-4370
Project: PDFBox
Issue Type: Task
Components: JempBox
Affects Versions: 1.8.16
Reporter: Tim Allison
Attachments: slow.zip
In our new batch of regression files on Tika, one of the new PDFs caused a
timeout. This is not an infinite loop, but it does take several minutes. This
may not be fixable.
Admittedly, the XMP is large, and there are quite a few events.
This is the code that triggers the problem.
{noformat}
XMPMetadata xmp = XMPMetadata.load(is);
XMPSchemaMediaManagement mmSchema = xmp.getMediaManagementSchema();
mmSchema.getHistory();
{noformat}
The slow part _seems_ to be setting the attribute namespace when creating a new
ResourceEvent. When I comment out the following in ResourceEvent's
initializer, the processing time is quite fast (1 second).
{noformat}
parent.setAttributeNS(
XMPSchema.NS_NAMESPACE,
"xmlns:stEvt",
NAMESPACE );
{noformat}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
