# HG changeset patch # User Zhengang Li <lizg(a)cn.ibm.com&gt; # Date 1207727823 25200 # Node ID f58194a3271fe20b19894c9f3472164316e1ee4c # Parent b5857217c64a081450cd21219035084eac455d85 [TEST] redirect logging to /dev/null instead of stderr The original default basicConfig() added a stderr Streamhandler. The log is duplicated to both the stderr and the file log 'vsmtest.log'. Outputting all the log to the screen makes the test less easier to read. Signed-off-by: Zhengang Li <lizg(a)cn.ibm.com&gt;

Kaitlin Rupert wrote: > zli(a)linux.vnet.ibm.com wrote: >> # HG changeset patch >> # User Zhengang Li <lizg(a)cn.ibm.com&gt; >> # Date 1207727823 25200 >> # Node ID f58194a3271fe20b19894c9f3472164316e1ee4c >> # Parent b5857217c64a081450cd21219035084eac455d85 >> [TEST] redirect logging to /dev/null instead of stderr >> >> The original default basicConfig() added a stderr Streamhandler. >> The log is duplicated to both the stderr and the file log 'vsmtest.log'. >> Outputting all the log to the screen makes the test less easier to read. >> >> Signed-off-by: Zhengang Li <lizg(a)cn.ibm.com&gt; > > Thanks Zhengang for this patch - I've applied this. I'd noticed the > output of the suite had changed, but wasn't sure what the cause was. > > Agreed - the output is much easier to read with this patch applied. > However, I could see why duplicating the log to stderr might be useful > in some cases. > > Does anyone see value in this? We could add an option that allows to > user to enable logging to stderr. Although, if we decide to do this, > some of the tests will need to be cleaned up. Some tests are quite > verbose, which might need to be pared down. > > Thoughts? I was going for '-v' for 'verbose'. But we already have a '-v' for 'virt'. How about '-d' for 'debug output'? And I'd like this '-d' disabled by default. Ideas?