Re: CVS XMP Import: Test result

[Larry Ewing <lewing novell com>]

On Fri, 2006-09-15 at 22:31 +1000, Bengt Thuree wrote:
> Hi Guys
> 
> I imported a set of 2250 or so pictures, and compared the by CVS
> imported tags with my old patch.
> The following is my findings: 
> 
> 1) Missing the following field:
>  >>http://ns.iview-multimedia.com/mediapro/1.0/Location<<

I didn't realize this existed in addition to the Iptc3Xmp Location, is
the IView schema documented anywhere?


> 2) Missing various comment/description fields
>  >>http://ns.adobe.com/exif/1.0/UserComment<<
>  >>http://ns.adobe.com/photoshop/1.0/Headline<<
>  >>http://ns.adobe.com/photoshop/1.0/Caption<<
>  >>http://purl.org/dc/elements/1.1/description<<
>  >>http://purl.org/dc/elements/1.1/title<<

See the comment in the source for these and the ones below.  In general
I'm not happy with moving data from One field to another in the Xmp just
so it shows up as a tag.  We either need to define tags as capable of
having predicates other than dc:subject or make this sort of combination
a manual operation.  I left the tag (dc:subject) merging in for now so
that people could see how it works and dicuss it.

> 3) Would be good to prepare for the following tags as well:
> --> Copyrights related
>  >>http://purl.org/dc/elements/1.1/creator<<
>  >>http://purl.org/dc/elements/1.1/rights<<
> 
> --> Rating related
>  >>http://ns.adobe.com/photoshop/1.0/Urgency<<
>  >>http://ns.adobe.com/xap/1.0/Rating<<
> 
> 4) Need to handle the following
> For the comment and text fields, I have encountered a lot of the
> following "empty" strings...
>       <ns2:UserComment>
>         <rdf:Alt>
>           <rdf:li
> xml:lang="x-default">&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;&#x0;</rdf:li>
> 

I removed the striping code in the original patch because as best as I
can tell the fact that it is making it that far is a bug.  Can you open
a bug and attach the file in question, I need to see which part of the
metadata is generating these.

> 5) I get the following om the console when I run the import.
> ** Message: Bad encoding for
> '/home/bengt/test-pictures/2001/03/20010317_S\xe4len_013.JPG'
> Consider using MONO_EXTERNAL_ENCODINGS
>  

The same for this please.  I assume it is using JIS which depends on
having the mono i18n 

> 6) Not importing the following fields from IPTC
>   Contact                  = Record.Application | 118,
>   ContentLocationName      = Record.Application | 27,
> 

I'll look into these.

Have I ever mentioned we really need unit tests for the image loaders?

> 7) UserComment is not de-coded properly (from Hebrew)
>     <rdf:Description>
>       <ns3:UserComment>
>         <rdf:Alt>
>           <rdf:li xml:lang="x-default">\u05D7\u05EA\u05D5\u05E0
> \u05D4</rdf:li>
>         </rdf:Alt>
>       </ns3:UserComment>
>     </rdf:Description>

The hebrew you sent me before worked here I think.  Go ahread an open a
bug for this too I guess.

> I also encountered two exceptions during import which I raised two bugs
> for.
> 
> Should I raise a bug for the above, or will this be worked upon so I can
> retest soon?

Definitely raise bugs for the things I mentioned just to make sure they
don't slip through the cracks.

--Larry