Hi Bas,
This is going to be a quick answer as I'm out of the office till
Wednesday or Thursday.
I think that you are right that the location attribute may be required
to start with a "/". I'm not sure this restriction is necessary, it may
just be the way I had imagined things would happen. I'll have to take a
look when I return.
The crawlableDatasetImpl element allows any XML to be included as child
elements. The intent of that was to allow CrawlableDataset
implementations to be handed some configuration information. You could
try handing your implementation some base part of the URL you want to
scan rather than putting the whole thing in the location. Not sure this
really make sense in this case, I had more been thinking
CrawlableDataset implementations might need more complicated
configuration information. But I think treating some base part of the
URL as config information could work. Sorry, that adds a bit of
complexity (mainly parsing some XML) to CrawlableDatasetDods
When I return, I'll take a look at the restrictions on the location
attribute and see if there are good reasons for them.
Ethan
Bas Retsios wrote:
Hello Ethan,
I have managed to download all missing .jar libraries and build the
thredds.war file. There are some compile errors, but ant is still able
to build the .war file, and I believe that the major part of Thredds
is not affected.
Now I am trying to make sure my new class CrawlableDatasetDods is
called. The following is a datasetScan that I have created for it.
<datasetScan name="DodsTest" path="testDods"
location="http://reason.gsfc.nasa.gov/opendap-bin/nph-dods/FTP_DATA/Giovanni/OPS/TOMS/EP/";
ID="mytest" addDatasetSize="true" addLatest="true">
<crawlableDatasetImpl
className="thredds.crawlabledataset.CrawlableDatasetDods" />
<metadata inherited="true">
<serviceName>mytestservice</serviceName>
</metadata>
</datasetScan>
Unfortunately this does not work. After inspecting log-files, and
creating additional log entries, I have concluded that the function
thredds.catalog.parser.jdom.InvCatalogFactory10.readDatasetScanNew()
works as I expect. However, the function
InvDatasetScan.buildCatalogBuilder() is never called, whichever link I
try. I was expecting that this function would be called when I go to
the following link:
http://my_thredds_server:8080/thredds/catalog/testDods/catalog.html
But the only thing I get is a HTTP Status 404 - message, with
description "The requested resource () is not available".
Note that my location starts with "http:". I have seen code that only
accepts paths starting with '/', so I'm a bit suspicious whether I
have understood the location field properly.
Any ideas?
Bas.
--
Ethan R. Davis Telephone: (303) 497-8155
Software Engineer Fax: (303) 497-8690
UCAR Unidata Program Center E-mail: edavis@xxxxxxxx
P.O. Box 3000
Boulder, CO 80307-3000 http://www.unidata.ucar.edu/
---------------------------------------------------------------------------
