Two change sets - first the "boring" part: Fix compiler warnings, etc.
http://cr.opensolaris.org/~gwr/cthon-nfs-tests-fixwarn/webrev/ Now the more interesting stuff: Some time ago, I observed that the Connectathon NFS tests would often terminate early after any test error. That may be helpful for regression testing, but in most cases it will give better coverage if the tests try to keep running, and just report whatever failures happened. These changes add a "-k" flag (for "keep going") to the various runtests scripts, which when set, makes the tests keep going where possible. Test scripts now show a summary of failures at completion, so failures are obvious even when using the -k flag. There are also minor changes to some of the tests to make them more tolerant of unsupported functionality, also with the goal of letting tests continue as much as possible. (This allows using the tests on smbfs mounts, etc.) Here is the "keepgoing" change set: http://cr.opensolaris.org/~gwr/cthon-nfs-tests-keepgoing/webrev/ In both cases, I uploaded the whole workspace, so if you want to try these out, you can just "wget -r" from the URL above with the /webrev/ part removed. Then try: cd ...keepgoing make ./runtests -f -k -a /test/mount/location So far I've tried this version on: Solaris NFS, Solaris smbfs, and a Linux (Ubuntu 9.4) cifs mount. Feedback appreciated. Thanks, Gordon -- This message posted from opensolaris.org
