See #346 to get the full picture. Local runs were failing while CI was not. The reasoning behind:

Take this run for example. The whole step takes 2m59s to execute. Out of these, 2m26s are spent in the main regression and 13s in unit tests. It seems unlikely that the stdlib is being codegen'd using the remaining 20s. Opening an issue to investigate.

We should investigate what is going on here. What is the usual runtime?