Summary: ... Ruby : 833 tests, 0 failures, 7 known bugs ... Builder : 101 tests, 173 assertions, 0 failures, 0 errors ... HTML5lib: 542 tests, 1090 assertions, 0 failures, 0 errors ... Mars : 377 tests, 377 assertions, 0 failures, 0 errors !!! OpenID : 859 tests, 6428 assertions, 0 failures, 2 errors ... REXML : 350 tests, 1278 assertions, 0 failures, 0 errors ... RakeCont: 87 tests, 221 assertions, 0 failures, 0 errors ... xmpp4r : 171 tests, 949 assertions, 0 failures, 0 errors Details: http://intertwingly.net/projects/ruby19/ Explanation: A lot of the value of Ruby comes from the rich ecosystem of gems that are available. Some gems will simply work on Ruby 1.9. Others will require updates. I've made a number of fixes to get these gems working. When possible, I've gotten those fixes into the source code repository for those projects, in the hopes that these changes will make it into the next release of their associated gems. Fixes that are awaiting incorporation can be found in the patches directory in the URI mentioned above. I'm not running all the tests for all of the packages listed above. In particular, only a subset of the html5 and xmppr4 tests are run due to dependencies. My purpose on publishing this page is to increase the efficiency of core developer and gem developer communications. Knowing that a given test passed yesterday, and fails today, can improve problem resolution. The OpenID issue identified above is an example. For this to work effectively, it helps if prior runs pass. While there can be no guarantee that the combination will pass all tests at any given point in time, if people are responsive to issues, the problem determination times can be made dramatically shorter. At the moment, I'm not running any official tinderbox software package or product. This is simply a few modest shell scripts. But they have been tailored to provide value add, like the filtering of the outputs through addr2line. Initially, I'm only running this twice a day. Over time, I can look into increasing the frequency, making the feedback loops quicker. If I can a few products a weeks to this list, in a matter of a few months we can have a substantial amount of regression testing and reporting one. - Sam Ruby