From ade2373711d9974f8bf11d5ab7aa5b174d43f20a Mon Sep 17 00:00:00 2001 From: Adam Miller Date: Fri, 4 Mar 2022 02:01:55 +0000 Subject: [PATCH] Fixing referer on request with null hop path --- warcprox/crawl_log.py | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/warcprox/crawl_log.py b/warcprox/crawl_log.py index 5bc26bd..c859159 100644 --- a/warcprox/crawl_log.py +++ b/warcprox/crawl_log.py @@ -68,7 +68,7 @@ class CrawlLogger(object): hop_path = recorded_url.warcprox_meta.get('metadata', {}).get('hop_path', '-') if hop_path is None: hop_path = "-" - hop_path_referer = recorded_url.warcprox_meta.get('metadata', {}).get('hop_path_referer', recorded_url.referer) + hop_path_referer = recorded_url.warcprox_meta.get('metadata', {}).get('hop_path_referer', "-") if hop_path_referer != recorded_url.url.decode('ascii'): if hop_path == "-": hop_path = "B" @@ -80,7 +80,7 @@ class CrawlLogger(object): '% 10s' % content_length, recorded_url.url, hop_path, - recorded_url.referer or hop_path_referer or '-', + recorded_url.referer or hop_path_referer if hop_path != "-" else "-", recorded_url.mimetype if recorded_url.mimetype is not None and recorded_url.mimetype.strip() else '-', '-', '{:%Y%m%d%H%M%S}{:03d}+{:03d}'.format(