support "flat" and "deep" type detection of a given document
A "flat" detection means specifying the document format by using
the URL and some configuration data only. That will perform but produce
may invalid results if e.g., the extension of the document is wrong.
A "deep" detection means looking into the document stream to be right
which format it supports. Of course that includes a "flat" detection before.
The combination of both ones should produce stable results everytime.
make a "deep" detection or optional a "flat" detection by using a
MediaDescriptor
Instead of XTypeDetection::queryTypeByURL() this function
use a MediaDescriptor to specify the document for detection.
Such descriptor hold different informations about the document.
He will be passed to any part of made detection process and every part
can change it to actualize it.
The property MediaDescriptor::URL should be set on
this descriptor as minimum. It specifies the location of the document.
If this parameter is missing another one is required: MediaDescriptor::InputStream.
This can be usefull to prevent operaton against multiple opening of the stream
and perform the operation. If this stream isn't already included the detection
will open it (if allowed!) and add it to the descriptor so it will be available for
all following parts.
A combination of both parameters can be usefull to perform the operation
and make results more stable; but only one of them is required.
Of course its possible to specify more document properties (e.g.
MediaDescriptor::ReadOnly).
As an additional feature it's possible to suppress "deep" detection by using argument
AllowDeep.
Parameter Descriptor
means the MediaDescriptor which specify the ressource for detection
Parameter AllowDeep
if it is set to true a "deep" detection will be follow a "flat" detection
if it is set to false a "flat" detection will be made only
Returns
an internal type name which represent the detected format
or an empty value for unknown ones