[CONNECTORS-983] Rethink how document version string is used in IRepositoryConnector for communicating "always retry" and "delete document"

• Type: Improvement
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: ManifoldCF 2.0
• Fix Version/s: ManifoldCF 1.7
• Component/s: Framework crawler agent
• Assignee: Karl Wright
• Reporter: Karl Wright
• Created: 24/Jun/14 04:48
• Updated: 27/Aug/14 23:02
• Resolved: 27/Aug/14 23:02

Description

Right now, all IPipelineConnector implementations require specification information to be packed into a version string. This is enforced by the fact that the methods that use this information do not receive a Specification object with which to work directly – which guarantees consistency between version string and what is indexed.

IRepositoryConnector, on the other hand, still passes Specification objects to all methods. This is in part because version strings in IRepositoryConnector have an extended meaning: a blank value means that the document must be refetched on the very next crawl attempt. It would be better, perhaps, to separate these concerns and come up with a common way of dealing with version strings.