I'm running a Linux server and Tikiwiki 5.1 and trying to use search indexing to search within uploaded files.
This is working fine for other mime types but not MS Word documents. E.g.
"application/pdf" works using "pdftotext %1 -" as it's file handler
"application/vnd.ms-excel" works with "strings %1".
I've tried using strings and catdoc, but neither are populating the the search_data field in the tiki_files table. I've tried reindexing and re-uploading the file. The file_type field in the db correctly identifies the mime type.
Both strings and catdoc work successfully from the command line so I'm inclined to write a script to run the commands from cron and update the data into the database. But this is obviously a bit of a bodge so I'd rather have it done natively.
I've chatted in IRC but was unable to get any solution to the problem.