Another thing that seems likely to be useful, if somebody is up for it, is to grab the source from svn from a couple weeks ago, and run masscheck with that, and compare how long it takes to run to what's currently in svn. If it was a change in source, that should verify it. Then you could figure out which commit caused it to slow down. (And if we switched to git like everybody else, you could do that awesomely with a "git bisect run" script.)
Probably best to use a relatively small sample of emails for that testing. Are the "tagged_builds" we use for masscheck any different from svn snapshots? -- "Believe nothing, no matter where you read it or who has said it, even if I have said it, unless it agrees with your own reason and your own common sense." - Buddha, 563-483 B.C. http://www.ChaosReigns.com
