Closed subttle closed 3 years ago
I changed this to remove the unncessary -XRankNTypes
requirement in the source, and must have forgotten about this little rearrangement. That said, I am reading the PVP spec and realize these changes probably warranted a major version bump, which I can do ASAP this week, and I agree with your assessment that the Alternative
probably requires an application more often than otherwise.
If you want to get in your Equivalence
work, i'm happy to release a patch Monday or Tuesday.
Closed via #26
thanks @subttle!
Hello again!
Thank you again for all your work on this great package! I've noticed in the latest version (
smash-0.1.2
) there's a slight change in the type signature from the version prior (smash-0.1.1.0
) for thepartitionWedges
,partitionCans
, andpartitionSmashes
functions.In the earlier version:
Because the
f
was specified first, it was arguably easier to useTypeApplications
to specifyf
and let thet
be inferred. For example, in GHCi, one could:vs. the newer type signatures:
For which one now has to specify an extra type application:
So as it stands, the current code is:
but I think the type signature would be more preferable with
Foldable t
andAlternative f
flipped:And similarly for
partitionCans
andpartitionSmashes
. It is perhaps a minor inconvenience, so perhaps you prefer to leave it (or perhaps I'm missing a reason for the changed ordering), but I thought I'd bring it to your attention in case you'd like to restore the previously available convenience. I apologize for such a long post for such a little issue. Thank you for your time and consideration!