Re: [Haskell-cafe] Informal modelling and simulation of protocols

Hi,

On Tue, 5 Mar 2019 at 23:11, Will Yager wrote:

...

http://hackage.haskell.org/package/dejafu might do what you want

Thanks for the suggestion, that does look useful. From a quick look at the documentation... am I right in thinking that this test framework only covers concurrency when you're directly using IOVars and STM? Any concurrency that might involve other primitives, or IOVar/STM uses in libraries, wouldn't be testable? Unless the primitives/library had also been written to use the dejafu io classes and monads of course... which might be a good idea in and of itself.

Yes, that's the case. It's a bit of an unfortunate limitation but I don't currently have a way around it. You may find it interesting to look at adjoint.io's libraft, which uses dejafu for some tests: https://github.com/adjoint-io/raft/blob/master/test/TestDejaFu.hs The way they do it is by writing their library in terms of some "MonadRaft" classes, and give a concrete implementation of those classes in terms of dejafu's ConcIO monad for testing. -- Michael Walker (http://www.barrucadu.co.uk)