Y, that was helpful. The guessing code starts with the HPSF info if that exists and then backs off to the former font-based hackery.
All tests pass, including the file that Dominik identified. The file from https://bz.apache.org/bugzilla/show_bug.cgi?id=60952 still yields junk -- there is no HPSF in that doc. So there's clearly still room to figure out how to map the correct encoding to a given section. But I think we're good for now. Thank you, again, Andi! r1791002 -----Original Message----- From: Allison, Timothy B. [mailto:talli...@mitre.org] Sent: Tuesday, April 11, 2017 6:45 AM To: POI Developers List <dev@poi.apache.org> Subject: Re: POI 3.16 Final? Got it. Will take a look shortly. ________________________________ From: Andreas Beeker <andreas.bee...@gmx.de> Sent: Tuesday, April 11, 2017 3:12:47 AM To: POI Developers List Subject: RE: POI 3.16 Final? > For clarification, when you say property sets...do you mean the Document > Properties from the tablestream? No, I meant the HPSF document-/summary information, which codepage looked better than some asian codepage on the 4th font object of one of the failing files. [1] [1] 2nd entry in http://people.apache.org/~centic/poi_regression/reports/ --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@poi.apache.org For additional commands, e-mail: dev-h...@poi.apache.org