Apache OpenOffice (AOO) Bugzilla – Issue 43230
Suppport import/export OLE Objects from/to wordml using xslt
Last modified: 2011-03-26 15:02:26 UTC
After the fix for i36885 had been integrated we get near to importing MS 2003 XML docuement with OLE Objects. But currently the BASE64 encoded OLE objects are not compatible between WordProcessingML and Open Office XML. To get the OLE Objects you can either save a doc file(with OLE objects) to xml the get them at /w:docOleData/w:binData or save a doc file to html and you will find an OLE objects storage files at <yours_files>/oledata.mso . You can base64 decoder here. http://www.fourmilab.ch/webtools/base64/ Here you can get an MDI MFC Docfile Viewer : http://support.microsoft.com/default.aspx?scid=kb;en-us;139545 . You can use it to browse OLE Structured Storage Files (DocFiles).
There are manys wordml issues which depends on this.
Hi windly, thanks for using and supporting OpenOffice.org... reassigned to mru mci -> mru: Hi mru, I think this is something for you...
windly -> mci Just for your information, I am your colleague in Beijing China. I am currently focusing on Mozilla Addressbook integration and WordML filter. :)
Reassigned to JSI.
-> Gavin: That would be IMO your part or am I wrong?
AFAIU OOo 2.0 would be great otherwise set it to 2.0.1
cc' ing Jack.
Sorry. Maybe my comments are not cleanly enough. I mean the oledata exported by wordml are not in general oledata format. It same that Microsoft compact the streams in it. So it can be render now. We need the framework or the flat filters ti handle this. Wind->jsi Could you please assign it to the right developer?
glu->mib: It seems that OLE bin-data in WordML is a little tripped, compared with OLE binary data supported in OOo XML format. So, is it possible to extend support to this WordML-style OLE binary data? Or if possible, who's properly in charge of it?
Wind, can you please attach a bug document.
Created attachment 22974 [details] bugdoc
There are four files in the bug doc. They are two wordml files and two corresponding OOo flat files. In each wordml files, there is a base64binary named oledata.mso. It contains all ole objects in this file. It's in MS ole docfile format. You decoded them to binary and use stg.exe to view the file structs.
The stream is compressed using zlib. The first 4 byte are a unsigned long. It is the uncompressed length. The reset bytes of the stream is zlib compressed data. After uncompress I get a well formated OLE Structured Storage Files (DocFiles). It works in both OOo and MS Word. So to support both import from and export to wordml in xslt filters, I suggest we deal with oledata in the same way MS Word does. It is: 1. There is only one ole data in the final saved file. 2. It includes multiple streams, each one has a unique name and is a compressed data with 4 byte unsigned long uncompressed length prior to it. 3. Each ole object in the file has property to point out which stream belongs to it.
I had created a xalan extension for this
fixed in CWS wordmlfilter02
reopen to reassign
assign to qa
seen good in cws wordmlfilter02 -> set to verified
seen good in src680_m91 (with workarround) -> closing