You are correct, and this is indeed a bug. I'll create a ticket and check in a fix.
Thanks! Karl On Tue, Dec 17, 2013 at 5:38 PM, Markus Schuch <[email protected]> wrote: > Hi, > > i think there might be a bug in the SolrConnector regarding the generation > of the output specification string: > > During several recrawls i saw that many documents were refetched even if > the version string of the documents was unchanged. > I debugged into the WorkerThread and learned that the output connection > also can trigger recrawling by returning a version String with the > getOutputSpecification Method. > > The problem is that the SolrConnector somtimes delivers a different output > specification string. I configured a document length limit in the output > connector which is packed into the output specification string. Somtimes > the specification String does not include the length limit parameter and > looks just like "0+0+". > > I think the reason is that the getOutputSpecification method relies on > member variables which are filled when getSession() is called. But it is > not garanteed that getSession() was called when getOutputSpecification is > invoked resulting in different states of the member variables. > > Is this a known issue? (i use manifoldcf 1.1) > > Thanks, > Markus > > >
