Nothing random there, just the ability to continue from the point of failure. Your build log tells you what commands cabal used. One of them failed. You edit the command line to see whether you can fix the issue. But to run and test the modified commandline, you need the temporary build
#427: how to access temporary build-files/logs? ----------------------------+----------------------------------------------- Reporter: claus | Owner: Type: enhancement | Status: new Priority: normal | Milestone: Component: Cabal library | Version: 1.6.0.1 Severity: normal | Resolution: Keywords: | Difficulty: normal Ghcversion: 6.8.3 | Platform: ----------------------------+----------------------------------------------- Comment (by duncan): Replying to [comment:2 claus]: directory in the state in which that previous command failed. Once your modified command succeeds, you know whether you need to fix your package or cabal or ghc or .. Normally we should use a new randomly generated temp directory, since we're creating tmp dirs inside of a globally writable dir. So the first issue for restarting a build like that is knowing what build dir to use. The next issue is that we would be re-running commands so that's not continuing from the point of failure. -- Ticket URL: http://hackage.haskell.org/trac/hackage/ticket/427#comment:3 Hackage http://haskell.org/cabal/ Hackage: Cabal and related projects