stastic harvests identifiers: error on capture, results in error in Record

[ghukill]

When running static harvests, you have the ability to include an XPath expression to find a value to use as the record_id for a Record. However, it appears that if that process fails, it is written to the Record's error column with the error that occurred.

This is helpful in many ways, but the error is not necessarily with the document, which that column is usually reserved for, it's only the identifier. This results in skipping those Records for indexing to ElasticSearch.

Either:

• Continue to note error in error field for Record, but have ES attempt to index if the document column contains valid XML
• Do not record in error field, but note somewhere else

[ghukill]

This brings up a further point of how to handle errors after the fact:

• ability to "clear errors" for the Job?
• when using as input for a new Job, allow shunt for success/error Records?

[ghukill]

More tolerance for static harvesting. AmbiguousIdentifier errors still result in success == True for Record, allowing indexing.

Other errors still result in success being False.