crawler: add support to re-crawl zones

[wander]

When a zone has been completely crawled, we should re-crawl it from time to time to find new NSEC3 hashes. Unfortunately we have to send all n queries again to the server but at least we already have the query names and do not need to do BF-range computation.

[wander]

See the attached dump files of vc. today and vc. from some point in the past. Both seem to be complete for the crawler without re-crawling feature.

[wander]

sort by serial

[wander]

(In [462]) write flags (OPTOUT) and types (in text presentation format) for each NSEC into database

closes #28

note: all currently existing NSEC3 records have these fields empty, need to implement re-crawl feature (references #21)

[wander]

(In [1017]) added re-crawl feature, closes #21

design choices:

• when using re-crawl, all current NSEC entries are marked as obsolete, thus duplicating all database entries if nothing has changed
• all obsolete NSEC entries are hashed at *EVERY STARTUP* with non-optimized Python code to identify gap cutters without brute-force (TODO: we should either move this into OpenCL or at least make this a CLI argument)

[wander]

(In [1018]) commit during re-crawl, references #21

[wander]

(In [1019]) withhold named cursor after commit, references #21