Callan McGill

It was kindly pointed out to me by Li-yao that I get the correct behaviour if I pass `--sertop-arg=--no_prelude`

Yes, that is fair enough! I'd be happy to try to improve the situation when I get the chance if you are interested?

I think that'd be a good idea and it would allow us to add functions like the one suggested in this issue too: https://github.com/haskell/containers/issues/798

It would be better performance as we don't need to do any of the checks necessary to see if the key is already present in the map (e.g. if we...

To be more specific the code is: ```haskell insert :: Key -> a -> IntMap a -> IntMap a insert !k x t@(Bin p m l r) | nomatch k...

One branch less for a heavy workload of inserts when the writer knows the check is unnecessary seems like a free win to me. With regards to the specifics, I...

Shall I try to take this up and finish it - anyone know what there is to do left?

Here is how it looks on the original benchmark, the same benchmark with 50% hits, with 90% hits and with all failures: ```terminal original fast-fail version: benchmarked lookup time 183.5...

I can't see any good reason that `find` should not be this variant since it otherwise throws an error.

@nomeata I did not do anything sensible. The current benchmark has keys `12..2^12` so I just shifted them all by `2^12`. In general the `IntMap` benchmarks could do with some...