[PDFBOX-456] PDFTextStripperByArea never finds any text (pageNo check in PDFTextStripper always returns false)

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 0.8.0-incubator
• Fix Version/s: 0.8.0-incubator
• Component/s: Text extraction
• Environment: 0.8.0-incubator as well as checkout from SVN (rev#767932).; Not affected: lastest sf.net release (0.7.3)
• Assignee: Unassigned
• Reporter: Hannes Erven
• Created: 24/Apr/09 11:19
• Updated: 21/Oct/09 09:52
• Resolved: 29/Apr/09 07:56

Description

PDFTextStripperByArea does not return any text from pages.

This is due to a check in PDFTextStripper#processPage() (first line) that compares the currentPageNo number (initially 0) against the startPage (initially 1). Since PDFTextStripperByArea does not set startPage and/or currentPage, this comparison always gives false and no text is extracted.

A possible fix is to include the following code in PDFTextStripperByArea#extractRegions right before the call to processPage():
setStartPage(0)
setEndPage(0)

Since I'm not very familiar with the inner PDFbox workings, this might be more of a hack than a solid fix.

The issue was introduced in PDFTextStripper 1.70 (old SF.net CSV), where the currentPage++ was removed from just before the check in processPage().