Ivan, Thanks for the patches. I will find time to review them.
P.S. We submit patches via Bugzilla. Please use it next time. Brief instructions for contributors are here: http://jakarta.apache.org/poi/getinvolved/index.html I opened a new bug and attached your patches to it: http://issues.apache.org/bugzilla/show_bug.cgi?id=42474 Could you attach sample ppts that were causing problems? One for Notes - Slides linkage problem and another for NPE in RichTextRun.isBold()? Regards, Yegor > Forgot the attachments... > On 20.05.2007 11:58, Ivan Todoroski wrote: >> 1. Incorrect matching of notes to slides >> >> This is a long standing problem apparently, I noticed a TODO comment >> about it. Anyway, after a bit of trial and error the problem seems to be >> that for some PPT files the SlidePersistAtom from the note's >> SlideAtomSet sometimes has a wrong slideIdentifier, so matching it to >> the associated slide is impossible. But, if you look at the NotesAtom >> from the actual Notes record (retrieved via getCoreRecordForSAS()), it >> has the correct slideId, so that's what I use in my patch. >> >> I don't know yet why >> SlideAtomSet.getSlidePersistAtom().getSlideIdentifier() is different >> from Notes.getNotesAtom().getSlideId(), maybe that's the real bug and my >> patch is simply working around it? Someone more knowledgeable could shed >> some light whether these two values are even supposed to be equal, in >> the mean time this patch works perfectly for all my test cases which >> exhibited incorrect note->slide association. >> >> 2. NPE in RichTextRun.isBold() when the RichTextRun comes from a Notes >> model object >> >> The Notes object wasn't doing setSheet(this) on the text runs it >> returned, causing a NPE inside RichTextRun.isCharFlagsTextPropVal() when >> any style accessor is called, because it tries to get the style info >> from the master sheet if not present. However, doing setSheet() in Notes >> was not enough, it merely moved the NPE a couple of lines down because >> NotesMaster objects are not yet implemented in HSLF, so I had to add a >> few null checks in RichTextRun too. >> >> >> P.S. Not subscribed, please CC. >> --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] Mailing List: http://jakarta.apache.org/site/mail2.html#poi The Apache Jakarta POI Project: http://jakarta.apache.org/poi/
