Re: [Aperture-devel] aperture bug on mimetypes.xml - please make patch

[Darren G. <da...@on...>]

Gotcha. The reason was I didn't know it needed the filename to detect
openoffice correctly.
Thanks for pointing that out.

One thing I was pondering though, in terms of file type detection, is
that maybe something like
a mini-expert system or rulebase is needed that can sort out the nuances
between certain file types
since many application types use _other_ types of files as a basis. 

For example the OpenOffice rule would be something like:

"IF it looks like a zip archive and IF it ends with .odt and IF it has
_certain openoffice codes in it_"
   "THEN its an openoffice type, zip subtype"

Anyway, just a thought.

On Tue, 2010-04-13 at 23:48 +0200, Antoni Mylka wrote:

> W dniu 2010-04-13 17:25, Darren Govoni pisze:
> > I don't use a URI.
> >
> > Here is my code from my JSP mime extractor
> >
> > ApertureRuntime aperture = new ApertureRuntime();
> >
> > String filename = request.getParameter("filename");
> >
> > RDFContainer container = aperture.extractFrom(new File(filename));
> >
> > System.out.println("FILENAME: "+filename);
> > String mime = aperture.identifyMimeType(new FileInputStream(filename), "");
> 
> And why not
> 
> String mime = aperture.identifyMimeType(
> 	new FileInputStream(filename), filename);
> 
> ???
> 
> > out.println(mime);
> >
> > filename=anything.odt
> > mime=application/zip
> >
> > Darren
> 
> Antoni