x-ole-storage instead of vnd.ms-outlook for *.msg files

[glazzari]

Is there a way to detect *.msg (outlook) files? I'm getting x-ole-storage instead of vnd.ms-outlook, as returned by Tika, for example. This is my code:

mimeTypeDetector.detectMimeType("filename", new ByteArrayInputStream(fileContent));

My msg file starts with D0 CF 11 E0 A1 B1 1A E1.

[adamhooper]

Oof! Kinda.

The vnd.ms-outlook format is non-standard -- so non-standard that we at Overview use it for something else (PST files). We implemented ours in extra-mime-info.xml.

To mimic Tika (which detects PST as vnd.ms-outlook-pst), we'd need to change overview-server and mime-types, in lockstep. It's painful.

I think I prefer a different route: let users supply their own mime.cache with whatever types they want. This repo could track shared-mime-info exactly, so there wouldn't be such rigmarole next time. We'd need a small API change (a new ctor for MimeTypeDetector) and a big HOWTO guide.

@glazzari Are you willing to take on the task of tweaking the API and adding a HOWTO? We'd need a new public MimeTypeDetector(InputStream mimeCache) constructor; the default constructor would become MimeTypeDetector(getClass().getResourceAsStream(MimeCache)) and we'd need a unit test with a non-standard mime.cache and a non-standard file (maybe .msg). The HOWTO should guide users on editing extra-mime-info.xml and re-running ./rebuild-mime-cache.sh. The existing <mime-type> in extra-mime-info.xml should be commented out -- it could serve as an example.