[TIKA-2523] Regression in ppt parsing -- "typeface can't be null or empty"

• Type: Bug
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Assignee: Unassigned
• Reporter: Tim Allison
• Created: 08/Dec/17 22:00
• Updated: 11/Dec/17 13:34

Description

We noticed a regression in ppt parsing in POI 3.17 while running the large scale regression tests in prep for the release of Tika 1.17. There are about 200 new exceptions, but it looks like there is only one cause.

Stacktrace:

org.apache.poi.hslf.exceptions.HSLFException: Couldn't instantiate the class for type with id 1000 on class class org.apache.poi.hslf.record.Document : java.lang.reflect.InvocationTargetException
Cause was : org.apache.poi.hslf.exceptions.HSLFException: Couldn't instantiate the class for type with id 1010 on class class org.apache.poi.hslf.record.Environment : java.lang.reflect.InvocationTargetException
Cause was : org.apache.poi.hslf.exceptions.HSLFException: Couldn't instantiate the class for type with id 2005 on class class org.apache.poi.hslf.record.FontCollection : java.lang.reflect.InvocationTargetException
Cause was : java.lang.IllegalArgumentException: typeface can't be null nor empty
	at org.apache.poi.hslf.record.Record.createRecordForType(Record.java:186)
	at org.apache.poi.hslf.record.Record.buildRecordAtOffset(Record.java:104)
	at org.apache.poi.hslf.usermodel.HSLFSlideShowImpl.read(HSLFSlideShowImpl.java:279)
	at org.apache.poi.hslf.usermodel.HSLFSlideShowImpl.buildRecords(HSLFSlideShowImpl.java:260)
	at org.apache.poi.hslf.usermodel.HSLFSlideShowImpl.<init>(HSLFSlideShowImpl.java:166)
	at org.apache.poi.hslf.usermodel.HSLFSlideShow.<init>(HSLFSlideShow.java:181)
	at org.apache.tika.parser.microsoft.HSLFExtractor.parse(HSLFExtractor.java:78)

Issue Links

• relates to: TIKA-2525 Upgrade to POI 3.17.1 when available