I am having a very strange error. I am running open office on my CentOS server and i'm trying to convert user-uploaded documents to html. Generally, the process works fine for smaller documents. however, once the document size reaches anything larger than 1.5MB our files have trouble being fed through the open office headless conversion process.
We have tried using
$ /usr/bin/libreoffice --headless --convert-to html example.doc -outdir ../
and
$ unoconv -f html -o /var/www/vhosts/shmim.shmimshmam.com/tmp/convertdocs/example.html bible.doc
the first creates a rogue process on our server that hogs system resources until we manually kill it.
the second throws an error in our error log and never makes it to the conversion process. that error is the one i posted in the title, but here it is again, this time without the file path removed.
unoconv: UnoException during import phase: The document 'file:///var/www/vhosts/shmim.shmimshmam.com/tmp/convertdocs/08336587c498af3d3292654657061828.doc' could not be opened.
when i try to open the file that has been uploaded to our server (through an html form), it seems to be corrupted. This is strange to me, as the upload process is very simple and the uploaded file is not tampered with at all. Observing the two files in binary form, there does not seem to be any differences in their makeup. However, when i convert both the original file and the uploaded file to an md5 hash for comparison, they are different.
This is all very puzzling to me, and I am at a point now where i want to burn my underwear and stop wearing deodorant and call my mom and tell her she is a hussie. Please help keep me from destroying my personal life!
