[mu rfe] Support indexing PDF attachments

mu search is great, only thing that Is missing is searching the attachments like pdf. For this I fallback to gmail search.

For PDFs - text can be extracted using ghostscript.

Don't expect mu to evolve into a general document indexer. But perhaps it could be done, the multi-core indexing (1.6.x) would help here.

I've added this to IDEAS.org, for longer-term research projects, closing here. Thanks!