On Thu, Mar 09, 2023 at 04:21:20PM -0800, Felix Lechner wrote: > I did mention the Cuirass question on IRC as you had asked, but no one > seemed to remember. Then I addressed the Haskell failure by responding > to a recent thread about that upload. [2] By your reply, I know that > you saw my message. Thanks for working on those leads! In my reply to your message regarding Haskell, I CC-ed Mathieu Othacehe, who is relatively active developing Cuirass and assists with operations on ci.guix.gnu.org. Let's see if he replies. I do see that on ci.guix.gnu.org, we are running Cuirass 1.1.0-13.1341725, which should respect the max-silent-time property. According to NEWS, this was added in Cuirass 1.0.0. One might consider trying to debug a local installation of Cuirass, to see if it actually respects this property. Imagine a test package that sleeps for 10 seconds in a build phase but has a max-silent-time of 5. In the meantime, I would check that gocryptfs is working satisfactorily based on the branch. I was surprised that Cuirass tried to build GHC and all these R packages as a result of these patches. Do you know if these packages (e.g. GHC) depend on Go somehow?