Create a bugzilla issue anyway and mark it as an enhancement request with a low priority. I could work on it when I haven't anything else to do, i.e. probably never. :-)
Best regards
Rainer Klute
Rainer Klute IT-Consulting
Dipl.-Inform.
Rainer Klute E-Mail: [EMAIL PROTECTED]
Körner Grund 24 Phone: +49 231 5349423
D-44143 Dortmund Mobile: +49 172 2324824
OpenPGP fingerprint: E4E4386515EE0BED5C162FBB5343461584B5A42E
Beskow, Christer schrieb:
> Beskow, Christer schrieb:
>
>>> I'm using POIFS to read and write CustomProperties to the
>>> DocumentSummaryInformation.
>>> Some word documents only contain a section with formatId
>>> D5CDD505-2E9C-101B-9397-08002B2CF9AE, which is the second id in
>>> SectionIDMap.DOCUMENT_SUMMARY_INFORMATION_ID.
>>> When I try to create an instance of DocumentSummaryInformation , method
>>> isDocumentSummaryInformation in PropertySet returns false, as it only
>>> compares the formatId of the section with the first id. This results in a
>>> org.apache.poi.hpsf.UnexpectedPropertySetTypeException: Not a
>>> org.apache.poi.hpsf.DocumentSummaryInformation.
>>>
>>>
>>> Is it safe to alter the code so that isDocumentSummaryInformation() also
>>> returns true if the id is D5CDD505-2E9C-101B-9397-08002B2CF9AE instead of
>>> D5CDD502-2E9C-101B-9397-08002B2CF9AE?
>>>
>>>
>> Hm, good question! Regarding to the official documentation from
>> Microsoft at
>> <http://msdn.microsoft.com/en-us/library/aa379840(VS.85).aspx> and
>> especially the remarks about 'The DocumentSummaryInformation and
>> UserDefined property set', I'd say that the second section cannot exist
>> without the first section. Therefore my first answer is: No, it is not
>> safe to alter the code as you described.
>>
>> On the other hand, it might be a good idea to make HPSF more
>> fault-tolerant so that it accepts a DocumentSummaryInformation property
>> set that only consists of the second section and inserts an empty first
>> section while reading the property set.
>>
>> Please file an extension request under
>> <https://issues.apache.org/bugzilla/> and provide a sample document with
>> only the second section!
>>
>
> Today I found out that all the office documents with the behaviour as
> described above in our system (A) originates from another
> system (B). I reviewed their source code and found out that they add a
> section with id D5CDD505-2E9C-101B-9397-08002B2CF9AE to the document in a way
> that for some documents creates the behaviour as described above.
>
> So, this is imho no bug in POI, it's a bug in the application where the
> document has been altered/created.
> Do you still want me to create an issue in bugzilla?
>
> I will ask the project team for system (B) to fix their implementation.
>
> FYI: Word automatically repairs the document and add section 0 to the
> document if it doesn't exist.
>
> thanks for your help
> regards
> Christer
signature.asc
Description: OpenPGP digital signature
