The qmin system test was printing spurious output. On investigation,
the test case turned out to be both broken and ineffective: its
expectations were wrong, and it was printing the output because its
wrong expectations were not met, and those failed expectations were
not causing a test failure. All of this has been corrected.