The GitHub Actions job "Java CI with Maven" on stormcrawler.git/issue-2082-host-canonicalisation has failed. Run started by GitHub user abhinav-phi (triggered by rzo1).
Head commit for run: 18f394a86ae5ac669ae898b2ccb11061a7dbdefe / abhinav-phi <[email protected]> Never-throwing percent-decoder, one canonical host for every queue mode (#2082) Review feedback on #2122: - URLDecoder.decode throws on a malformed escape and maps + to a space, both reachable from crawled content which the security model treats as hostile. getCanonicalHost now uses a hand-rolled percent-decoder that leaves + alone and keeps malformed escapes as literal characters; it never throws - the javadoc no longer claims to return 'the host okhttp connects to': okhttp keeps a trailing dot, this method normalises it away. It returns the form used to key politeness queues and the robots.txt cache - FetcherBolt and SimpleFetcherBolt canonicalise the host once and use it in all three queue modes: byIP and byDomain still resolved and split the raw escaped host, so http://%65xample.org/ kept its own queue in those modes - IDN hosts are not collapsed yet (unicode vs punycode spellings); recorded as #2145 Report URL: https://github.com/apache/stormcrawler/actions/runs/34152857105 With regards, GitHub Actions via GitBox
