<headius>
stuff like rake spec:ffi job ran in about 30s
<enebo[m]>
cool
<enebo[m]>
so wait for one job to do what all of them did and all others run quicker
<headius>
yeah
<headius>
looks like the bootstrap build takes 1:43 so the rest of the jobs are delayed starting by that much
<headius>
but once they go many of them run in seconds
<headius>
we should probably split the biggest suites too... test:mri and spec:ruby:core end up being the last jobs running and the former still takes over 30min
<headius>
ok got one failure that seems to be new so I'll investigate that, and I'll see if I can pull more slow tests out of test:mri:core
<headius>
test_any_hash_fixable in MRI test_hash.rb takes a minute and runs twice
<headius>
I'm not sure why it runs the same test subprocess 20 times
<headius>
probably going to just exclude this... I don't think it tests anything important for us
<headius>
some C issue with the keys or hashing or something