Created attachment 27328 [details] git-format-patch svn-id trunk@1150704 The Sheet::findTextRun method made a number of incorrect assumptions about the structure of atoms following a TextHeaderAtom, resulting in a number of error messages being spewed to stderr for PowerPoint files which did not match its expectations. Some of the cases I have seen which POI is incorrectly handling include: * A TextHeaderAtom immediately following a TextHeaderAtom * TextRulerAtoms and MasterTextPropAtoms between a TextHeaderAtom and the corresponding TextCharAtom / TextBytesAtom
The patch looks OK but I'm reluctant to apply it without a unit test. Can you upload a junit test that covers the cases that POI currently does not handle? Yegor
I think I've handled those cases when I've rewritten the text handling. If this still throws exceptions for your files, please attach them and I'll try to figure out what's wrong.