Another week, another quick summary of topics of interest (to me) as we close out the Feast of Saint Patrick. As before, if you feel you have something that you feel needs more eyeballs, please reply and add it. Don't be shy.
A few significant mailing list threads: - We've talked a lot about testing at various points in the past. There's generally an understanding about what gets tested before committing, but here's a proposal for codifying test (new?) standards for commits ( https://lists.apache.org/thread.html/0854341ae3ab41ceed2ae8a03f2486cf2325e4fca6fd800bf4297dd4@%3Cdev.cassandra.apache.org%3E ) . At the risk of editorializing, formalizing this seems like a great idea. - This dovetails into an earlier thread about testing in general - https://lists.apache.org/thread.html/0854341ae3ab41ceed2ae8a03f2486cf2325e4fca6fd800bf4297dd4@%3Cdev.cassandra.apache.org%3E . I'm not sure what the action item is here - I suspect it's coming to an agreement that 1) committers/reviewers need verifiable test results, and 2) they need to be shared in a common format (even if it's not a common CI system, since the historical de facto common CI system isn't necessarily available for everyone anymore). - There's a proposal to kill the wiki. Would anyone notice? Short of an objection, maybe we just transition this to a VOTE thread: https://lists.apache.org/thread.html/1614ec4e1d6ee47666e4c8d2328a6f5e9cabc62d163854d33ebbdb4f@%3Cdev.cassandra.apache.org%3E Relevant to the test theme of the week on the mailing list, here are some test-related JIRAs. If you feel like making Cassandra better this weekend while you're off the clock, writing (or fixing) some unit tests never hurt: - https://issues.apache.org/jira/browse/CASSANDRA-9012 - (at least) 26 individual areas we know aren't well tested right now. If you feel like writing tests, these may be a great starting point - https://issues.apache.org/jira/browse/CASSANDRA-12944 has been updated with an example of testing PendingRangeCalculatorService. I suspect Stefan would really really appreciate some feedback on this proposal. - More generally, https://issues.apache.org/jira/browse/CASSANDRA-9012/?jql=project%20%3D%20CASSANDRA%20AND%20component%20%3D%20Testing%20AND%20resolution%20%3D%20Unresolved%20ORDER%20BY%20due%20ASC%2C%20priority%20DESC%2C%20created%20ASC&selectedTab=com.atlassian.jira.jira-projects-plugin:issues-panel is a link to all of the "Testing" labelled JIRAs still unresolved. Some JIRAs that are Patch-Available and no Reviewer (there are an unfortunate number of these, but here's 5 that aren't mine): - https://issues.apache.org/jira/browse/CASSANDRA-13197 ( +=/-= shortcut syntax bugs/inconsistencies ) - https://issues.apache.org/jira/browse/CASSANDRA-13262 (cqlsh bug when querying the same column multiple times) - https://issues.apache.org/jira/browse/CASSANDRA-12870 (Optimization to the pending range calculator) - https://issues.apache.org/jira/browse/CASSANDRA-12696 (changing log level by component aliases instead of class names) - https://issues.apache.org/jira/browse/CASSANDRA-13307 (cqlsh incorrectly hard codes protocol version so it won't downgrade when talking to old c*) Finally: we've had a handful of tickets for ppc64 / ppc64le / aarch64. Personally, I think this is great - these aren't officially supported, but I don't think anyone objects to committing patches if we have reason to believe it'll help make Cassandra run on those platforms. However, given lack of test hardware available to the project, if you happen to own (or sell, or create) this hardware and you're submitting a patch to make it work, AT THE VERY MINIMUM, you need to help the committer out and show us the state of unit tests - that means run "ant test > test.log" when submitting a patch, and upload the log so the reviewer/committer has an idea of what's going on (and ideally, get all of your patches together in a single branch with one JIRA, so we can do a single review/commit, don't open a JIRA for each change you need to make). Enjoy your weekend, - Jeff
