Uploaded image for project: 'HttpComponents HttpClient'
  1. HttpComponents HttpClient
  2. HTTPCLIENT-2336

Update PublicSuffixMatcher to use "formal algorithm"

    XMLWordPrintableJSON

Details

    • Improvement
    • Status: Resolved
    • Major
    • Resolution: Fixed
    • None
    • 5.4
    • HttpClient (classic)
    • None

    Description

      We’ve been using the PublicSuffixMatcher component from “httpcomponents-client” to implement comparisons of domains and public suffixes according to the rules defined by the standard at https://github.com/publicsuffix/list/wiki/Format#formal-algorithm and using the Public Suffix List produced by Mozilla at https://publicsuffix.org/list/effective_tld_names.dat.

      We noticed unexpected results where the current behaviour of PublicSuffixMatcher deviates from the formal algorithm. We ported over the unit tests from https://github.com/publicsuffix/list/blob/master/tests/test_psl.txt to determine all the differences as there are a number of tests which fail. This follows on from an earlier discussion on the mailing list at https://lists.apache.org/thread/ylom7gcopxtrcb4zm6q8c9k7fo0jt5km.

      We have made changes to get all the tests to pass and will raise a PR shortly to reference this ticket.

      Attachments

        Activity

          People

            Unassigned Unassigned
            massdosage@gmail.com Mass Dosage
            Votes:
            0 Vote for this issue
            Watchers:
            2 Start watching this issue

            Dates

              Created:
              Updated:
              Resolved: