On 07/20/11 09:26, Stefan Bodewig wrote:
> Hi,
> 
> I'm trying to document the attribbutes of the new license matcher I
> added to the Antlib yesterday, basically they are the same as those of
> described in
> <http://ci.apache.org/projects/rat/apache-rat-plugin/examples/custom-license.html>
> 
> I wanted to show the impact of the attributes, i.e. where does which
> value show up inside the report.  So I dug into pieces of RAT's code
> that I've never really visited before.
> 
> So far I have
> 
> * notes
> 
>   Ends up in a metadata attribute RAT_URL_HEADER_SAMPLE and becomes the
>   text content of the <header-sample/> element.
> 
>   Invisible inside the plain text report as header samples are only
>   emitted for unknown licenses.
> 
> * licenseFamilyName
> 
>   Ends up in a metadata attribute RAT_URL_LICENSE_FAMILY_NAME and
>   becomes the name attribute of a <license-family/> element.
> 
>   This one is checked against the list of approved license names and
>   results in <license-approval name='false'/> (or true, depending on the
>   approved licenses, of course).
> 
> * licenseFamilyCategory
> 
>   Ends up in a metadata attribute RAT_URL_HEADER_CATEGORY and becomes
>   the name attribute of the <header-type/> element.
> 
>   This is the only visual impact any of the attributes has on a plain
>   text report.  It is shown next to the file name in the list of scanned
>   files.
> 
>   BUT it also ends up in a metadata attribute
>   RAT_URL_LICENSE_FAMILY_CATEGORY and becomes the name attribute of a
>   (second) <license-family/> Element.
> 
> So we end up with two <license-family/> elements, one with the "short
> name" and one with the longer one.  Is this intentional?  The plain text
> report ignores those elements completely.

IIRC the original intent was to support license reasoning but I'd be
very happy to see the reporting simplified by removing them

Robert

Reply via email to