As I replied privately to the reporter, I've looked at this issue: I've fixed a few things in the ODT generation code, and now the wiki:syntax export passes the validator cited here. These changes are only in the git repo for now. (use "git clone http://aurelien.bompard.org/projects/dokuwiki/odt.git")
However, odt2txt still fails to recognize it as a valid ODT file. The strange thing is : if I unzip the odt file, and re-zip it without changing anything, odt2txt handles it just fine ! Therefore, I'm not sure the bug is in my ODT generation code. It's probably the "zipping" part, which is done in PHP by Dokuwiki, and is perfectly usable with unzip or OpenOffice. Maybe the "kunzip" library bundled with odt2txt lacks something to handle this uncommon archive. I'm no C expert, so that's where my investigation stops. Thanks for the report. By the way, the odfpy library (http://opendocumentfellowship.com/development/projects/odfpy) has some nice command line tools to convert ODT to other formats (like XHTML). I don't think it's packaged in Debian yet, but you may find it interesting. Aurélien -- http://aurelien.bompard.org ~~~~ Jabber : [EMAIL PROTECTED] "Computer Science is no more about computers than astronomy is about telescopes." -- E. W. Dijkstra -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]