Add IntoResponse impl for BytesMut and Chain #767
Conversation
This comment was marked as resolved.
This comment was marked as resolved.
|
You can just ignore the cargo deny CI failure. I'll get that fixed at some point 😊 |
Co-authored-by: David Pedersen <david.pdrsn@gmail.com>
AlisCode
force-pushed
the
more-into-response
branch
from
420e19e
to
907c8d6
Compare
axum/CHANGELOG.md
Outdated
| @@ -9,6 +9,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0 | |||
|
|
|||
| - **added:** `middleware::from_fn` for creating middleware from async functions. | |||
| This previously lived in axum-extra but has been moved to axum ([#719]) | |||
| - **added:** Implement `IntoResponse` for `bytes::BytesMut` and `bytes::Chain<T, U>` ([#767]) | |||
There was a problem hiding this comment.
Hm I knew this could come up at some point. This change was actually made in axum-core so it should at least be mentioned in its changelog as well. The question is whether it should be mentioned in axum's as well. The heuristic I've used so far is that only significant changes are mentioned in both. I kinda feel this isn't very significant so should only be in axum-core's changelog.
@jplatte wdyt?
There was a problem hiding this comment.
Didn't notice that there was a changelog specific to subcrates, I believe you're correct, this isn't worth a line in the "main" changelog. I'll move that there :)
There was a problem hiding this comment.
I know this won't be very helpful, but I don't care much either way 😅
|
Re-implemented Also squashed commits to keep a clean history :) |
AlisCode
force-pushed
the
more-into-response
branch
from
907c8d6
to
ed8433f
Compare
We squash merge all PRs so you actually don't have to. |
* Add IntoResponse impl for BytesMut and Chain Co-authored-by: David Pedersen <david.pdrsn@gmail.com> * Add CHANGELOG entry Co-authored-by: David Pedersen <david.pdrsn@gmail.com>
* Add IntoResponse impl for BytesMut and Chain Co-authored-by: David Pedersen <david.pdrsn@gmail.com> * Add CHANGELOG entry Co-authored-by: David Pedersen <david.pdrsn@gmail.com>
Adds two
IntoResponseimplementations for convenience to work onbytestypesMotivation
bytescratebytes::Chainseemed like a good opportunity to do so.Solution
Two new implementations of IntoResponse :
BytesMutwhich freezes (zero-cost operation) to get aByteshandle and use its IntoResponse's implementation.Chain<T,U>whereT: BufandU: Buf, which iterates over every collects everything into aBytes. Sadly the implementation ofFromIteratorseems to show that it's not exactly zero-cost.