In org.archive.crawler.basic.Frontier.next(), a
CandidateURI with a true forceFetch() will be sent to
emitCuri even if another URI of the same class is in
progress.
However, emitCuri() calls noteInProgress() which
asserts that no other URI of the same class is in progress.
Rather than an out-of-band forceFetch, we need a
facility that guarantees the "forced" URI is the next
one to be fetched, within the normal constraints.
I think this could be accomplised by adding a stack to
each class object... then each class would have both a
queue and a stack -- adding to the queue means
"eventually" and pushing to the stack means "before
everything else"... the stack would be exhausted before
the queue is considered.
This facility could be useful for fetching related
(embedded) items soon after where they originate, as
well -- but more investigation is necessary.
Gordon Mohr
None
None
Public
|
Date: 2007-03-14 00:08
|
|
Date: 2004-03-25 22:14 Logged In: YES |
| Field | Old Value | Date | By |
|---|---|---|---|
| status_id | Open | 2004-03-25 22:14 | kristinn_sig |
| resolution_id | None | 2004-03-25 22:14 | kristinn_sig |
| assigned_to | nobody | 2004-03-25 22:14 | kristinn_sig |
| close_date | - | 2004-03-25 22:14 | kristinn_sig |
Copyright © 2010 Geeknet, Inc. All rights reserved. Terms of Use