Helps when the "Content-Type" header is not configured in the "Section" indexer.conf command, and thus is not stored in the database. When generating excerpts using Cached Copy, search.cgi tries first to fetch Content-Type information from the database. If it's not found, search.cgi tries to detect Content-Type automatically using some well-know substrings like
<HTMLor
<!DOCTYPE HTMLin the beginnig of the document, unpacked from Cached Copy. If automatic guesser fails to detect the document type, it chooses "text/plain" by default. You can change the default value to "text/html" or "text/xml". But usually, a better idea (if you use Cached Copies) is to activate storing Content-Type using a proper "Section" command in indexer.conf.