Just to recap:
- .odt file is written using libreoffice on windows 7 (libreoffice version 4.2.1.1, 64bit)
- File cannot be opened on linux - hangs using 100% cpu (libreoffice version 4.1.6.2-4.1.x86_64 on opensuse 12.3)
- File can be read on windows.
- I notice in content.xml, biblio entry style names are huge - up to 30 thousand chars! Presumably this is the problem. I see them 7K long in another .odt generated on windows also, which can be read on linux.
I don’t know that the biblio thing is the root cause, but it seems pretty unusual!
Here’s a snippet from content.xml…
<text:index-entry-span>[</text:index-entry-span>
<text:index-entry-bibliography text:bibliography-data-field="identifier" text:style-name="_30__20__5f_30_5f__5f_20_5f__5f_5f <<<snip>>> f_5f_5f_5f_5f_5f_Strong_5f_5f_5f_5 <<<snip>>> _5f_5f_5f_Emphasis"/>
Enormous numbers of 5f are prevalent.
Looking in a linux generated content.xml, I see
<text:index-entry-span>[</text:index-entry-span>
<text:index-entry-bibliography text:bibliography-data-field="identifier" text:style-name="Strong_20_Emphasis"/>
which looks much more normal.
Any ideas? Is this a known defect?