<< ASP.NET 2.0 - Web Deployment Projects
|
Home
|
Ruby on Rails >>
posted @ Thursday, November 10, 2005 10:44 AM
I forgot to ever post a follow up on my
original post, apparently. I did finally figure out what the deal was. Through a process of elimination, with the help of Microsoft tech support, we determined that the
Adobe IFilter for PDFs does not support the BindIFilterFromStream interface. This seemed crazy to me because you would think if the indexing service COM object could find the right filter given a file on the file system, it could certainly do it for the same file delivered as a stream (after all, it's the same data!), but apparently it relies on the IFilter implementation to support this type of call, and therefore can't find the Adobe IFilter. So for now, I am stuck with serializing these out to disk to index them. I'll find out if this is a big deal when I see how many of these things get stuck in the database, and I may have to change the plan.