I have a Bdc connector and associated security trimmer that worked fine with SharePoint 2013. On SharePoint 2016, the Bdc connector still works and the content is crawled. But when a user runs a search and my security trimmer is consulted, the
documentCrawlUrls value passed to the
CheckAccess function is a list of empty strings, where it used to be a list of my Bdc urls. Since I try to parse this as a
Uri it fails, but if I remove that code the search results do correctly show my content.
My understanding is that these urls come from my
INamingContainer implementation, but nothing there has changed and all calls to
GetDisplayUri return valid Uris.
Does anyone have any ideas for how to resolve this?