+1 for tests in different time zones. And also tests with different Guava versions. In fact I’m working on it: see https://github.com/julianhyde/calcite-avatica/commits/stage <https://github.com/julianhyde/calcite-avatica/commits/stage> and https://app.travis-ci.com/github/julianhyde/calcite-avatica/builds/238880701 <https://app.travis-ci.com/github/julianhyde/calcite-avatica/builds/238880701>.
-1 for randomized testing as part of CI. CI should be deterministic. So that if it breaks, the person to fix it is clear: it’s generally the last person who changed something. I support randomized testing via scripts. People can run the script for a couple of hours and log any bugs it uncovers. We already have several such scripts in Calcite (e.g. rex fuzz testing). Julian > On Oct 1, 2021, at 3:28 AM, Alessandro Solimando > <[email protected]> wrote: > > Thanks Vladimir for the proposal. > > +1 frome me, there are more and more "axes" for testing (OS, locale, JVM, > timezone, etc.), a randomized approach sounds like a good solution to tame > the combinatorial explosion. > > Best regards, > Alessandro > > On Fri, 1 Oct 2021 at 08:32, Vladimir Sitnikov <[email protected]> > wrote: > >>> Of course it would be fine to just change the timezone for one of more >>> existing tests which are targeting different JVM versions >> >> Alessandro, team, what do you think of >> https://github.com/vlsi/github-actions-random-matrix ? >> >> I'm kind of hesitant to introduce more and more /vlsi/ dependencies even >> though I believe they are the best tools available :) >> random-matix is a 200 line script that generates build matrix by selecting >> random values for the axes (e.g. jvm version, os, guava version, etc). >> >> Vladimir >>
