This one has come up a few times, both on our lists and elsewhere, eg: https://jira.atlassian.com/browse/CONF-18721 http://apache-poi.1045710.n5.nabble.com/Found-a-TextHeaderAtom-not-followed-by-a-TextBytesAtom-or-TextCharsAtom-Followed-by-3999-td2289102.html http://blog.gmane.org/gmane.comp.jakarta.poi.devel/month=20100901 It seems there are a few common causes of this warning, where the file isn't quite as we expect, but it's close enough that we can still work with it. We should handle those cases, so that the warning remains only for cases which really need flagging up
Hopefully fixed in r1577803. We now skip over "empty" TextHeaderAtoms, where there's no records before the next TextHeaderAtom If there's a TextRulerAtom / MasterTextPropAtom / TextSpecInfoAtom before the Text Chars / Bytes, we skip over those