Corrupt search with Umlaute

XMLWordPrintable

    • Type: Bug
    • Resolution: Duplicate
    • Priority: Medium
    • None
    • Affects Version/s: 2.7.1
    • Component/s: Search - Core
    • Environment:

      Confluence standalone 2.7.1 on Linux with MySQL 5 as Database, newest JDBC MySQL Driver (5.1),
      Storage-Enginge for MySQL: InnoDB.
      Java Version 1.6.0_04.
      Database set to UTF-8.

      Howto see this bug:

      • Create a new page with any title, text containing German Umlaute. (my example: unternehmensänderung)
      • Save the Page
      • Wait until page is indexed
      • Attach a file named unternehmensänderung.zip to this page

      Now search for:

      • unternehmensänderung (finds both, okay)
      • unternehmens* (finds both, okay)
      • unternehmensä* (finds only the attached file !)

      Why can confluence search attachments for the umlaut combined with the wildcard, but not the page itself?

      I think confluence is a really good wiki, but a good search enginge is really important for the acceptance in a company.

            Assignee:
            Chris Broadfoot [Atlassian]
            Reporter:
            Johannes Wissen
            Votes:
            2 Vote for this issue
            Watchers:
            5 Start watching this issue

              Created:
              Updated:
              Resolved: