BlockedIndefinitelyOnMVar on GHC 7.8 about acid-state

[meteficha]

I forgot to mention that the failed test was built with GHC 7.8.4. The Travis build from the same commit using GHC 7.10.1 passed.

[meteficha]

There definitely is something fishy going on with GHC 7.8. While GHC 7.10 is able to finish my test suite using 600 MiB, I can't make GHC 7.8 work because not even 2 GiB is enough for it and Travis "only" allows 3 GiB (note that the process was killed before GHC could print statistics).

Is it known that GHC 7.8 requires a lot more memory than GHC 7.10? The other related packages on the same build work fine on both 7.8 and 7.10, leading me to assume it isn't a mistake on my code.

[sdx23]

Just a guess, as I'm not aware of how Travis does this: If they handle disk-storage as a tmpfs the overall memory-usage may be too high with GHC 7.8 and as a result the thing run out of "disk space".

If you have the time, please check whether pull-request #58 fixes this for you. (should still fail, but not hanging in an MVar operation)

[adamgundry]

This issue is now rather old and it sounds like it was only observed on GHC 7.8 in particular circumstances, so I'm closing. Feel free to reopen if you have further information or think this is still a problem.