Hi Ben, How about automating this using travis and gerrit, so no commit gets in unless it passes sanity tests? This also simplifies review process as well.
Regards, Pritesh > On Jul 13, 2015, at 5:14 PM, Ben Pfaff <b...@nicira.com> wrote: > > Lately I've starting hearing about commits breaking the Windows kernel > module. If that's happening frequently, then it probably means that > developers are not testing their changes before they post them. It's > easy to get into that habit not testing, because it saves time and > because sometimes one can get away with it. > > Please do at least minimal testing of kernel module changes. To me, > that means that the kernel module should load and at least a "ping" > should go through. When the particular bug or feature that the commit > affects is one that can be easily tested (sometimes this is not true, > for sure), then please do test it in particular. > > It's nice to mention what testing you did in your commit message. And I > think that reviewers should review that part, too: if testing isn't > mentioned in the commit message, then it's worth asking what testing the > developer did. > > (This is usually less of an issue for changes outside the kernel > modules, because "make check" covers most of the rest of the system > pretty well. Of course, in those cases, if you're adding a new feature > or fixing a testable bug, then it's nice to add a new test case to the > testsuite or to adjust an existing test so that it would have found the > problem.) > _______________________________________________ > dev mailing list > dev@openvswitch.org > http://openvswitch.org/mailman/listinfo/dev _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
