Joshua Reich <[EMAIL PROTECTED]> writes: > Just a general question - I submitted a patch for contrib/cube (adding a > new function & converting everything from V0 to V1), what is the process > from here onwards? Do we have an active maintainer of this code?
It sounds like you've just acquired that position ;-) > How is it reviewed? Same as everything else, pretty much: patches go to pgsql-patches and are (supposed to be) reviewed before being committed. If it's in the nature of a new feature rather than a simple bug fix or code cleanup, you might want to first start a discussion on pgsql-hackers --- if anyone has a better idea about how to do things, it's better to find it out before you start coding instead of after you finish. > I noticed that one of the regression tests now fail, as the original > regression test expects a certain ordering from a query that doesn't > request ordered results. Should I fix the test case so that ordering > will be assured? Should I add test cases for most of the functions that > currently aren't being tested? Because we're relying so heavily on the buildfarm these days, failing regression tests are quite unacceptable. Adding an ORDER BY might be the best solution, or maybe you should just change the expected output --- do you understand exactly why the results changed? As for adding more tests, you can within reason --- don't make the running time enormous. regards, tom lane ---------------------------(end of broadcast)--------------------------- TIP 9: In versions below 8.0, the planner will ignore your desire to choose an index scan if your joining column's datatypes do not match