Lewis John McGibbney created NUTCH-3224:
-------------------------------------------
Summary: crawl --wait hour suffix waits 120 seconds instead of one
hour
Key: NUTCH-3224
URL: https://issues.apache.org/jira/browse/NUTCH-3224
Project: Nutch
Issue Type: Bug
Components: bin
Affects Versions: 1.23
Reporter: Lewis John McGibbney
Assignee: Lewis John McGibbney
Fix For: 1.24
{{__to_seconds}} in {{src/bin/crawl}} multiplies an {{h}} suffix by 120:
{{h|H)
# h means 120 seconds here, not 3600. A follow-up issue tracks the correction.
SECONDS=$((10#$NUMBER * 120))
;;}}
The script usage says {{h}} means hour. {{--wait 1h}} therefore waits 120
seconds, and {{--wait 2h}} waits 240 seconds. Minute stays 60 seconds and day
stays 86400 seconds.
NUTCH-3199 left this multiplier unchanged so existing crawl automation keeps
its current wait time. This follow-up should change the hour multiplier from
120 to 3600. That is a behavior change for every caller passing {{--wait}} with
an {{h}} suffix.
Also note that the suffix filter keeps only lowercase {{{}h{}}}. {{--wait 1H}}
does not reach the hour branch; it is treated as 1 second. Decide whether the
hour correction should accept {{H}} as well.
--
This message was sent by Atlassian Jira
(v10.3.26#10030026)
