Let's talk the next release (0.5.0)

[travisbrown]

Previously: #486, #594, #834. I think we were planning to get the cats-kernel rearrangement into 0.5.0, but it's been two months since 0.4.0, a lot has happened since then, we have at least one person with a good reason to want a new stable version, and I'm planning to release circe 0.4.0 in the next day or two, so I think it's worth considering cutting a 0.5.0 release now.

(For what it's worth, none of my projects with Cats dependencies require any changes to update from 0.4.1 to the current 0.5.0-SNAPSHOT.)

I'd be happy just to set the cutoff as whatever makes it to master before I've got the release notes written, but we could also wait if there's anything anyone particularly wants to have included.

---

Okay, checklist time:

• Review MonadWriter: Monad writer, fixes #889 #911
• Provide a default implementation for ap (or product): Applicative doesn't have default implementation for map and product #904
• Consider adding Compose and Bicompose: Add Compose and Bicompose? #900
• Refactor MonadFilter and MonadCombine so that instances can be provided for e.g. XorT: Separate Strong/Weak Laws for MonadFilter/MonadCombine #694, Reinstate EitherT, OptionT MonadCombine/MonadFilter after #694 completes #695
• Publish a new Algebra release and depend on it here: updated to latest algebra #994

[adelbertc]

Things of interest to me:

• Define map and product in terms of ap (or vice versa) for Applicative until we have a solid proposal for breaking out default implementations (Applicative doesn't have default implementation for map and product #904 PR: product and map can be implemented in terms of ap, fixes #904 #967)
• MonadWriter Monad writer, fixes #889 #911
• Compose data type Add Compose and Bicompose? #900

[adelbertc]

Oh one thing that I think a lot of folks may want is a new Algebra release. We've been asked about the tuple stuff for Monoid and co. a couple times now which is in master but hasn't been released yet.

Though maybe we just want to do the whole cats-kernel thing..

[ShaneDelmore]

I would love to see #695 in 0.5.0. My team has been working around it just fine but it is causing new contributors to our code some confusion as they don't understand why "if" doesn't work in for comprehensions as they would expect.

[travisbrown]

I agree that all these things are valuable but I'm quickly becoming less convinced that I want the circe 0.4.0 release to wait for Cats 0.5.0.

[kailuowang]

I agree with these valuable things, but to make all the other already implemented features during the last 2 mo, shall we consider a 0.5.0 release now and a 0.6.0 later? Or just a 0.5.0-M1 release would help.

[non]

Should we just release what we have as 0.5.0?

EDIT: It seems like we have release notes now, so if there is consensus I can go ahead and cut the release.

[travisbrown]

@non That sounds reasonable to me.

[adelbertc]

I'd at least like:

Review MonadWriter: #911 (just needs one more +1)

and

Publish a new Algebra release and depend on it here. (people have been asking for tuple instances for a while)

If not the first one, then at least the second one.

[ceedubs]

I just updated the checkboxes. We opted to not wait for a couple of the items, but this release has happened, so I'm going to close out this issue.