On Fri, Nov 8, 2013 at 3:03 PM, eberhard speer jr. <[email protected]> wrote: > ...Over the weekend I'll update the 'master' user-agent file here with > 'harvested' data and add the Big List to the SVN so other can join in > the fun and we can start defining 'serious' test-sets for benchmark > and other tests....
That would be great! As we're seeing with DMAP-29 its unlikely that all implementations will result the exact same results, so we might want to provide for exception definitions in a master test file. Something like "ideally this should be 'Samsung GT54240 but we know the java classifier outputs 'Samsung Android'", so that we can cope with slight differences while having an ongoing check of regressions and improvements. -Bertrand
