Ellie Hermaszewska

Just to recap on my train of thought, the below seems like a reasonable implementation: ```haskell -- | An instance of MonadNestedResource must satisfy the following laws -- -- -...

@snoyberg The definition in the preceding comment came out quite neat and tidy I think. How possible to you think it would be to get something like this included in...

One can observe similar behavior with the `test` command to the program described by `Setup.hs`

FWIW `kind-generics` supports deriving a `Generic`-like interface for GADT's

Possibly behind https://github.com/junegunn/fzf.vim/issues/1295

Ah, if I switch `command` here (https://github.com/junegunn/fzf/blob/94999101e358385f3ca67a6ec9512f549196b802/plugin/fzf.vim#L717) to `jobstart` then I get the expected contents in the fzf popup. (although obviously this doesn't allow capturing the results). `jobwait([jobstart(command)])` sadly doesn't...

Ah, this seems to work, great! ``` let fzf = {} let fzf.dict = a:dict let fzf.temps = a:temps function! fzf.on_exit(job_id, exit_status, event) dict call s:pushd(self.dict) let lines = s:collect(self.temps)...

Rebased onto 5.15: https://gist.github.com/expipiplus1/bdc5750447d979583702ce5d44592572

Thank you for the patient review, @RyanGlScott! Still to fix is the instance context generation, as it still wants to generate `Bifunctor` instances for any type variable `f : *...

Another alternative would be to allow configuring this behavior via `Options`, something like `preferFmap`, that way people who can tolerate writing their own instance head can use that and allow...