Draft of 0.7.0 release notes #1259
Conversation
Current coverage is 90.63% (diff: 100%)@@ master #1259 diff @@
==========================================
Files 237 237
Lines 3725 3725
Methods 3663 3663
Messages 0 0
Branches 58 58
==========================================
Hits 3376 3376
Misses 349 349
Partials 0 0
|
| And some name changes: | ||
|
|
||
| * [#1140](https://github.com/typelevel/cats/pull/1140): `cats.std` is now `cats.instances` | ||
| * [#1201](https://github.com/typelevel/cats/pull/1201): Many `*Unsafe` methods are now `unsafe*` |
There was a problem hiding this comment.
Ah, thanks @fthomas. There's a possibility there's also stuff still missing—I'll do a closer pass this afternoon.
travisbrown
force-pushed
the
topic/0.7.0-release-notes
branch
from
b35f2ba
to
10b2443
Compare
|
I went ahead and added commits adding new contributors and updating versions as described in the process document. |
|
@kailuowang Yeah, I wanted to wait for all three |
|
|
||
| Version 0.7.0 is the seventh Cats release, and includes several major rearrangements and changes to names. | ||
|
|
||
| ### Migration notes |
There was a problem hiding this comment.
Thank you for adding this section @travisbrown. I think this will be really helpful for people.
|
I'd love to finish #1280 before 0.7.0. It would be a shame to have 0.7.0 introduce FlatMapRec only to remove it in the next release. I can get 1280 passing in a day or two if we can wait. |
travisbrown
force-pushed
the
topic/0.7.0-release-notes
branch
from
80ad807
to
6037658
Compare
travisbrown
changed the title
|
|
||
| * All references to `cats.std` will need to be changed to `cats.instances` ([#1140](https://github.com/typelevel/cats/pull/1140)). If you're using `cats.std.all` or the other `cats.std` objects with wildcard imports, this is likely to be the only change you need to make. If you are importing or referring to instance definitions by name, you'll need to be aware that the naming convention has changed (see [#1066](https://github.com/typelevel/cats/pull/1066), [#1068](https://github.com/typelevel/cats/pull/1068), [#1110](https://github.com/typelevel/cats/pull/1110), and [#1122](https://github.com/typelevel/cats/pull/1122)). | ||
| * `NonEmptyList` and `NonEmptyVector` are no longer type aliases for `OneAnd`, so any code using `OneAnd` to construct or pattern match on these types will need to be changed to use `NonEmptyList` or `NonEmptyVector` directly. There are also some API changes; for example, `unwrap` calls will need to be replaced by `toList` or `toVector`, and `NonEmptyList(1, 2, 3)` is now `NonEmptyList.of(1, 2, 3)`. | ||
| * `pureEval` has been removed from `Applicative` ([#1234](https://github.com/typelevel/cats/pull/1234)), and has not been replaced, so if you are relying on it for laziness or effect capturing (which wasn't enforced or guaranteed), you'll need to find another approach. |
There was a problem hiding this comment.
What I suggested was to require a NaturalTransformation[Eval, M] if you need to create an M[T] from an Eval[T]. For M[T] with laziness, you can do something non-trivial.
travisbrown
force-pushed
the
topic/0.7.0-release-notes
branch
from
6037658
to
edbcb43
Compare
|
LGTM 👍 |
|
Since this is documentation I'm going to merge this after one +1. |
I still think we should include an update guide, but this is a first stab at #1253.This is ready to be reviewed, apart from the fact that
#1264, #1265,#1266(saving for next release),and #1267aren't yet included.