PackageEvaluator is back online

PackageEvaluator, sometimes referred to as “PkgEval,” is a core piece of the Julia infrastructure. It runs the tests for every registered package, updates our packages webpage (, and creates status badges. As you may or may not have noticed, PackageEvaluator has been offline since last October due to a variety of issues.

Well friends, I’m pleased to announce that PackageEvaluator is now back online. This means that you’ll finally get your shiny new 0.7 status badges for your packages’ READMEs. In case you aren’t sure what I’m talking about, it’s these things:

Note that in the past, PackageEvaluator was run daily, which means that there was a quick turnaround for package tags having the status reflected in badges and on the packages website. Due to the number of registered packages and the amount of time it takes to run all of the tests, we’ve reduced the frequency to every other day to ensure things don’t get backed up.

Further, since Julia v0.5 has been declared unmaintained, we’ve removed testing on 0.5 from PackageEvaluator to increase capacity for supported versions. So tests are only run on 0.6 and 0.7, even if your package still supports earlier versions.

Now that you can show the world your testing status on 0.7, it might be a good time to get your packages updated for 0.7. :wink:


Thanks Alex, also for all the work you have been putting into this and other infrastructure projects! If anyone in the community wants to lend a hand and help out, we would be more than happy about any help!


Thank you @ararslan!!

I would love to. How can I know the list of projects under the infrastructure side for Julia?

Most of the infrastructure projects are at Julia CI (Continuous Integration) · GitHub a few noteworthy are:

  • Nanosoldier.jl (continuous performance testing based on BenchmarkTools and BaseBenchmarks)
  • PackageEvaluator (runs all package test against nightly and release branches)
  • julia-buildbot (binary building and testing infrastructure)

there are a couple of other projects like AttoBot and generally we would love to have more automation. PackageEvaluator as an example is at the limit of what it can do (I think one run now takes more than 24h) and making it scale up and useable on demand (like Nanosoldier) would be great. There is also a slack channel #infrastructure where people hang out.