Hello devs,
I'm integrating BundleRepository with a ResolveContext and I found a
problem while trying to get some capabilities from repository when using
R5 spec findProviders().
I have discovered that findProviders's result is not returning a Content
capability only one Identity capability containing attributes from both
Identity and Content.
I did a debug and found that BR's SpecXMLPullParser class is the
responsible for merging those attributes into one identity capability.
Anyone knows what is the reason for that ? this behavior is not
following the spec, right ?
thanks,
Cristiano
Below is some methods taken from SpecXMLPullParser that shows the
current behavior:
private static Capability parseCapability(XmlPullParser reader,
ResourceImpl resource, URI baseUri) throws Exception
{
String namespace = reader.getAttributeValue(null, NAMESPACE);
if (IdentityNamespace.IDENTITY_NAMESPACE.equals(namespace))
{
*parseIdentityNamespace(reader, resource);**
** return null;*
}
if (ContentNamespace.CONTENT_NAMESPACE.equals(namespace))
{
if (resource.getURI() == null)
{
* parseContentNamespace(reader, resource, baseUri);**
** return null;*
}
// if the URI is already set, this is a second
osgi.content capability.
// The first content capability, which is the main one, is
stored in the Resource.
// Subsequent content capabilities are stored are ordinary
capabilities.
}
private static Resource parseResource(XmlPullParser reader, URI
baseUri) throws Exception
{
ResourceImpl resource = new ResourceImpl();
try
{
int event;
while ((event = reader.nextTag()) == XmlPullParser.START_TAG)
{
String element = reader.getName();
if (CAPABILITY.equals(element))
{
Capability capability = parseCapability(reader,
resource, baseUri);
*if (capability != null)*
resource.addCapability(capability);
}
else if (REQUIREMENT.equals(element))
{
Requirement requirement = parseRequirement(reader);
if (requirement != null) {
resource.addRequire(requirement);
}
}
else
{
PullParser.ignoreTag(reader);
}
}
PullParser.sanityCheckEndElement(reader, event, RESOURCE);
return resource;
}
catch (Exception e)
{
throw new Exception("Error while parsing resource " +
resource.getId() + " at line " + reader.getLineNumber() + " and column
" + reader.getColumnNumber(), e);
}
}
private static void parseIdentityNamespace(XmlPullParser reader,
ResourceImpl resource) throws Exception
{
Map<String, Object> attributes = new HashMap<String, Object>();
parseAttributesDirectives(reader, attributes, new
HashMap<String, String>(), CAPABILITY);
// TODO need to cater for the singleton directive...
for (Map.Entry<String, Object> entry : attributes.entrySet())
{
*if (IdentityNamespace.IDENTITY_NAMESPACE.equals(entry.getKey()))**
** resource.put(Resource.SYMBOLIC_NAME, entry.getValue());*
else
resource.put(entry.getKey(), entry.getValue());
}
}
