[PDFBOX-2867] Correct use of Float.NaN

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 1.8.9, 2.0.0
• Fix Version/s: 1.8.10, 2.0.0
• Component/s: Text extraction
• Assignee: Tilman Hausherr
• Reporter: Tilman Hausherr
• Created: 09/Jul/15 18:03
• Updated: 23/Jul/15 06:35
• Resolved: 09/Jul/15 18:08

Description

replace

wordSpacing == Float.NaN

with

Float.isNaN(wordSpacing)

because wordSpacing == Float.NaN is always false.

A test by Tim Allison showed 9 differences in TIKA (but "None of the differences were actual words. "), and no difference in a manual test with ExtractText.