Am 29.04.2017 um 00:08 hat Eric Blake geschrieben: > On 04/28/2017 03:33 PM, Kevin Wolf wrote: > > From: Max Reitz <[email protected]> > > > > There is no reason for the qemu-nbd server used for tests not to accept > > an arbitrary number of clients. In fact, test 181 will require it to > > accept two clients at the same time (and thus it fails before this > > patch). > > Quick question: > > Is this mention of test 181 the same test that occurs earlier in the > pull request as 8/34? If so, should the series have been reordered for > bisectability? If not, should the commit message be touched up? > > At any rate, bisectability of qemu-iotests is less of a concern than > bisectability of 'make check' success, so I don't think your answer > invalidates the pull request.
Yes, it's the same test 181 and if someone had noticed before I sent the pull request, I might have reordered the patches, but I don't think there's any practical bisectability problem with introducing a qemu-iotests case that fails on NBD. If you bisect, you typically bisect for a specific setup and test case rather than running the full set of qemu-iotests. Kevin
pgpYQHV08L4lE.pgp
Description: PGP signature
