From 6f21ede979168b1aef48eaaa529ce9e0e0aad1be Mon Sep 17 00:00:00 2001 From: Noah Levitt Date: Sun, 26 Oct 2014 19:17:23 -0700 Subject: [PATCH] seriously, fix extraInfo handling in AMQPCrawlLogFeed --- .../org/archive/modules/postprocessor/AMQPCrawlLogFeed.java | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/contrib/src/main/java/org/archive/modules/postprocessor/AMQPCrawlLogFeed.java b/contrib/src/main/java/org/archive/modules/postprocessor/AMQPCrawlLogFeed.java index fadb64fe..0df3c592 100644 --- a/contrib/src/main/java/org/archive/modules/postprocessor/AMQPCrawlLogFeed.java +++ b/contrib/src/main/java/org/archive/modules/postprocessor/AMQPCrawlLogFeed.java @@ -145,8 +145,10 @@ public class AMQPCrawlLogFeed extends AMQPProducerProcessor implements Lifecycle if (ei == null) { ei = new JSONObject(); } - jo.put("warc_filename", checkForNull(ei.opt("warcFilename"))); - jo.put("warc_offset", checkForNull(ei.opt("warcFileOffset"))); + // copy so we can remove unrolled fields + ei = new JSONObject(curi.getExtraInfo().toString()); + jo.put("warc_filename", checkForNull(ei.remove("warcFilename"))); + jo.put("warc_offset", checkForNull(ei.remove("warcFileOffset"))); jo.put("extra_info", ei); String str = jo.toString();