Replace repo-level CODE_OF_CONDUCT file with the inherited org-wide
[Code of
Conduct](https://github.com/apple/.github/blob/main/CODE_OF_CONDUCT.md)
in .github repo.
Co-authored-by: apple-ospo[bot] <326325757+apple-ospo[bot]@users.noreply.github.com>
## Motivation
`testBasicUnixSockets` binds a listener to a UNIX domain socket file
under `/tmp` (e.g. `/tmp/<uuid>_testBasicUnixSockets.sock`) but never
removes that file once the test finishes. Each run therefore leaves a
stale socket file behind, slowly littering `/tmp` on every test
invocation.
Fixes#53.
## Modifications
Added a `defer` block immediately after the socket path is computed that
removes the file with `FileManager.default.removeItem(atPath:)`. Because
`defer` blocks execute in LIFO order, this cleanup runs after the
existing `listener.close()`/`connection.close()` deferred closures, so
the socket is closed before the backing file is deleted. The removal is
best-effort (`try?`) so it never masks a genuine test failure.
## Result
`testBasicUnixSockets` no longer leaves a socket file in `/tmp` after
running. No behavioral change to the test's assertions.
Verified locally:
- `swift build` — Build complete.
- `swift test --filter NIOTSEndToEndTests/testBasicUnixSockets` — 1
test, 0 failures.
- Confirmed no `*_testBasicUnixSockets.sock` files remain in `/tmp`
after the run.
Motivation
* The shared `macos_tests.yml` workflow no longer accepts
`xcode_16_2_build_arguments_override` and the `xcode_16_3` input is a
no-op. Passing the removed input causes CI to fail.
* Older Xcodes are being removed from the runners; Swift version inputs
are the replacement.
Modifications
* Replace `xcode_16_2/16_3_build_arguments_override` with
`swift_6_1/6_2/6_3_build_arguments_override` carrying the same
`-Xswiftc -Xfrontend -Xswiftc -require-explicit-sendable` value.
Result
* macOS CI passes again using the Swift version inputs.
Only depend on NIOFoundationCompat on Apple platforms
### Motivation:
https://github.com/apple/swift-nio-transport-services/pull/211#issuecomment-4170778092
### Modifications:
Only depend on NIOFoundationCompat on Apple platforms because it's only
imported when Network can also be imported, so we don't unnecessarily
depend transitively on Foundation.
(We should use anyAppleOS as the condition if/when it becomes available
in SPM too: https://forums.swift.org/t/introducing-anyappleos/85728 )
### Result:
Stripped release binary is 20MB instead of 65MB in a project using
mqtt-nio
---------
Co-authored-by: Cory Benfield <lukasa@apple.com>
Motivation:
Swift 6.0 is no longer supported, we should bump the tools version and
remove it from our CI.
Modifications:
* Bump the Swift tools version to Swift 6.1
* Remove Swift 6.0 jobs where appropriate in main.yml, pull_request.yml
Result:
Code reflects our support window.
Motivation:
Swift 6.3 has been released, we should add it to our CI coverage.
Modifications:
Add additional Swift 6.3 jobs where appropriate in main.yml,
pull_request.yml
Result:
Improved test coverage.
Motivation
* Following [GitHub's
announcement](https://github.blog/changelog/2025-09-19-deprecation-of-node-20-on-github-actions-runners/),
Node 20 is being deprecated and runners will begin using Node 24 by
default starting March 4th, 2026. Node 20 will reach end-of-life in
April 2026.
Modifications
* Upgrade GitHub Actions to their latest versions to ensure
compatibility with Node 24.
* Replace `actions/checkout@v4` with `actions/checkout@v6` in all
workflow files.
Result
* Workflow files reference `actions/checkout@v6`.
Motivation
* `@main` references to `swiftlang/github-workflows` are deprecated.
Modifications
* Replace `@main` with `@0.0.7` in all workflow files.
Result
* Workflow files reference a pinned version of
`swiftlang/github-workflows`.
Motivation:
* More secure GitHub Actions workflows
Modifications:
Add explicit 'contents: read' permissions to workflows that did not have
explicit permissions defined. This follows GitHub Actions security best
practices by limiting the default GITHUB_TOKEN permissions.
Result:
An extra layer of security.
Fix NIOTS UDP premature close by not treating per‑datagram isComplete as
EOF; add multi‑datagram tests and NIOTSEchoClient example.
Motivation:
- While using [DNSClient
#40](https://github.com/orlandos-nl/DNSClient/issues/40), we observed
“one‑shot” behavior: a NIOTS connected‑UDP channel closed after the
first datagram, causing subsequent writes to fail or be dropped.
- Network.framework sets isComplete == true for each UDP datagram. This
indicates message completeness, not stream EOF. Treating that signal as
EOF caused the early close.
-
[PR#243](https://github.com/apple/swift-nio-transport-services/pull/243)
explored gating behavior differently (e.g., via
“completionClosesChannel”). This PR aligns EOF handling explicitly to
channel semantics (stream vs. datagram), which makes the reasoning
around transport semantics clear.
- Alternatives considered (brief):
- ContentContext.isFinal only: common stream half‑closes (FIN) arrive
with isComplete == true and isFinal == false, so relying solely on
isFinal would miss EOF on streams and risk hangs.
- protocolMetadata (e.g., NWProtocolUDP.Metadata): metadata
presence/shape varies and is often empty on final callbacks; channel
semantics (isDatagramChannel) are explicit and robust.
- Reproducer (pre‑fix): set a symbolic breakpoint `br s -n
nw_connection_cancel`; run `testBasicChannelCommunication()` test;
typical backtrace after the echoed datagram:
- StateManagedNWConnectionChannel.dataReceivedHandler(content=<N bytes>,
isComplete=true, error=nil) → didReadEOF() →
StateManagedChannel.close0(error=eof) → nw_connection_cancel.
- Note error=nil: close was triggered solely because isComplete was
true.
Modifications:
- Transport fix:
- StateManagedNWConnectionChannel.dataReceivedHandler now calls
didReadEOF() only when isComplete && !isDatagramChannel (streams
unchanged; datagrams no longer misinterpret isComplete as EOF).
- Tests:
- NIOTSDatagramBootstrapTests.testConnectedUDPEchoesTwoDatagrams
verifies two echoes on NIOTS connected UDP; both server and client
observe datagrams; bounded timeouts; proper cleanup.
-
NIOTSUDPMultiDatagramsPOSIXServerTests.testNIOTSClient_POSIXServer_MultipleDatagrams
verifies NIOTS client ↔ POSIX UDP echo; asserts two echoes to isolate
cross‑stack behavior.
- Example:
- NIOTSEchoClient executable example (Apple platforms): simple
connected‑UDP echo client using NIOTSDatagramConnectionBootstrap;
guarded with `canImport(Network)` and `@available`; minimal usage notes.
Result:
- Datagram channels (UDP) no longer misinterpret per‑packet completeness
as EOF; multi‑datagram send/receive works as expected.
- Stream channels (TCP/TLS) are unchanged; EOF continues to be detected
per existing behavior.
- Restores expected behavior so typical NIO UDP apps can be ported to
NIOTS by swapping event loops and datagram bootstraps (on Apple
platforms using Network.framework).
- Manual validation available via NIOTSEchoClient (pair with SwiftNIO’s
NIOUDPEchoServer).
Motivation:
Swift 6.2 has been released, we should add it to our CI coverage.
Modifications:
Add additional Swift 6.2 jobs where appropriate in main.yml,
pull_request.yml
Result:
Improved test coverage.
Since iOS 18, we're occasionally transitioned to the .setup state
### Motivation:
Closes#237.
We don't know how to reproduce this problem though, only that it occurs
extremely rarely on iOS 18 & later (including 26).
### Modifications:
Simply removed a precondition failure if we're informed of a transition
to the `.setup` state. I see no harm in this — it wasn't previously
happening, and if it happens now then we'll do nothing.
The only concern might be if this represents some kind of "reset", after
which we'll then pass back through states we've already visited — that
could conceivably break us or downstream consumers. I don't know how to
know that, test for it, or otherwise deal with it, though…
### Result:
Hopefully, we just won't crash any more.
### Motivation:
Some errors do not show up in debug builds. Enabling release mode builds
improves the CI coverage.
### Modifications:
Enable release mode builds for pull requests and scheduled builds on
main.
### Result:
Improved CI coverage.
Every EventLoop can also be a SerialExecutor. Let's opt into the default
NIO fast-path implementation.
SemVer patch as NIOTSEventLoop isn't public API.
Added buffer pool to NIOTSConnectionChannel for reading messages.
### Motivation:
In order to avoid creating a new `ByteBuffer` for every single message
read from the channel, we decided to reuse the
`NIOPooledRecvBufferAllocator` type from NIOCore to leverage a pool of
buffers.
### Modifications:
Used the buffer allocation mechanism provided by the
`NIOPooledRecvBufferAllocator` to reuse and adapt previously created
receive buffers.
### Result:
Channel reads can now reuse previously created buffers, avoiding
unnecessary overhead by creating new buffers every single time.
Motivation:
Swift 5.9 is no longer supported, we should bump the tools version and
remove it from our CI.
Modifications:
* Bump the Swift tools version to Swift 5.10
* Remove Swift 5.9 jobs where appropriate in main.yml, pull_request.yml
Result:
Code reflects our support window.
### Motivation:
The datagram-related types logically mirror the TCP-based ones
(`NIOTSDatagramListenerBootstrap` vs `NIOTSListenerBootstrap`;
`NIOTSListenerChannel` vs `NIOTSDatagramListenerChannel`;
`NIOTSDatagramChannel` vs `NIOTSDatagramConnectionChannel`;
`NIOTSDatagramBootstrap` vs `NIOTSConnectionBootstrap`).
However, some of the type names could more closely resemble their TCP
counterparts to make it easier to navigate the code/understand what
their purpose is.
Additionally, the docs for some of these types are wrong, as they've
been copy-pasted into the datagram versions without changes.
### Modifications:
There are separate commits for each of the following changes:
- Fix docs for `NIOTSDatagramListenerBootstrap` and rename the file to
match the type name.
- Rename `NIOTSDatagramConnectionChannelTests` to
`NIOTSDatagramBootstrapTests` (to match `NIOTSBootstrapTests`).
- Rename `NIOTSDatagramChannel` to `NIOTSDatagramConnectionChannel` (to
match `NIOTSConnectionChannel`)
- Fix docs for `NIOTSConnectionBootstrap`.
- Rename `NIOTSDatagramBootstrap` to `NIOTSDatagramConnectionBootstrap`
(to match `NIOTSConnectionBootstrap`). This one required a deprecate and
replace since it's a public type.
### Result:
Better docs and more consistency in our type names.
Allow users to provide a closure taking an `NWParameters` to customise
them before they're used to create `NWConnection`s.
### Motivation:
`NWParameters` are currently created using the provided TLS and UDP
options, and then passed over to new `NWConnection`s. However, there are
more ways in which `NWParameters` can be customised, so this new API
provides a way for users to do this.
### Modifications:
Introduce new `configureNWParameters` methods to the existing bootstraps
to allow configuring a closure for customising `NWParameters`.
### Result:
Users can now customise the `NWParameters` used to create new
`NWConnection`s.
Motivation:
Swift 6.1 has been released, we should add it to our CI coverage.
Modifications:
Add additional Swift 6.1 jobs where appropriate in main.yml,
pull_request.yml
Result:
Improved test coverage.
Current strict concurrency checks don't work consistently in Xcode. This
PR adds an additional flag (`-Xfrontend`) to our set of strict
concurrency flags to make sure nothing has been missed when building. It
also fixes an additional error uncovered after adding it.
Motivation:
* Improve test coverage
Modifications:
Enable macOS CI to be run on pull request commits and make the use of
the nightly runner pool for main.yml jobs explicit.
Result:
Improved test coverage.
Enable macOS CI on merge to main and daily timer
### Motivation:
* Improve test coverage
* Check test pass/fail status
* Monitor CI throughput
### Modifications:
Enable macOS CI to be run on all merges to main and on a daily timer.
### Result:
Improved test coverage run out-of-band at the moment so we can get a
feeling for if any changes need to be made in the repo or in the CI
pipelines to ensure timely and stable checks.
Motivation:
Dispatch wasn't imported in `NIOTSSingletons.swift` where `.default` was
used for the default QoS. This causes some build issues.
Modifications:
- Add missing Dispatch import
Result:
Fewer build issues
---------
Co-authored-by: Cory Benfield <lukasa@apple.com>
Enable MemberImportVisibility check on all targets. Use a standard
string header and footer to bracket the new block for ease of updating
in the future with scripts.
Remove the integration tests workflow from the main and scheduled jobs.
### Motivation:
There are no integration tests in this repository.
(see
https://github.com/apple/swift-nio-transport-services/actions/runs/11833082880)
### Modifications:
Remove the integration tests workflow from the main and scheduled jobs.
### Result:
No more scheduled run failures.
### Motivation:
* `main.yml` and `scheduled.yml` are mostly duplicative.
* Scheduled runs failed because of a deprecated reference to a Swift 5.8
pipeline
### Modifications:
* Unify `main.yml` and `scheduled.yml`
* Remove the reference to the 5.8 pipeline
### Result:
Working scheduled runs.
### Motivation:
To migrate to GitHub actions and centralised infrastructure.
### Modifications:
Changes of note:
* Adopt swift-format using rules from SwiftNIO. Modified so that it
ignores `NoAssignmentInExpressions` in some `XCTAssert` functions.
* Remove scripts and docker files which are no longer needed
### Result:
Feature parity with old CI.
Motivation:
Users are stuck with the hardcoded parameters for receiving data from a
connection.
Modifications:
- Add new options to `NIOTSChannelOptions` for configuring how to
receive data from a connection.
Result:
Users can configure how they receive data from a connection.
This reverts commit 40ffcdef46 in PR #209.
Unfortunately, Swift's behaviour around imports tends to be somewhat
"leaky". In this case, the leak is that the dependency on the
Package.swift allows downstream projects to import `NIOFoundationCompat`
without needing to actually specify the package dependency. This has
affected Hummingbird, which we noticed on our internal integration
testing functionality:
```
hummingbird/Sources/Hummingbird/Codable/JSON/JSONCoding.swift:18:8: error: no such module 'NIOFoundationCompat'
import NIOFoundationCompat
^
```
cc @Joannis @adam-fowler for the Hummingbird report.
Unfortunately, we can't make this change until we're willing to use a
semver major to achieve it. There may be some argument for doing that
now, as the semver major will be very cheap to adopt across the
ecosystem. But we'll need to do this in a considered way.
@Cyberbeni please feel free to reopen your PR targetting main, where we
can discuss whether this is worth issuing a semver major for.
Don't depend on NIOFoundationCompat in NIOTransportServices on Linux
### Motivation:
I'm trying to build a small utility tool based on mqtt-nio which depends
on NIOTransportServices. I noticed that NIOTransportServices depends on
NIOFoundationCompat on all platforms but only imports it where the
Network framework is available. Removing this dependency on non-Apple
platforms allows us to not import Foundations, significantly reducing
the size of the build product when using the Swift Static Linux SDK.
### Modifications:
This PR and a similar one to mqtt-nio.
### Result:
Hello world docker image depending on mqtt-nio went from 169MB to 85MB.
Motivation:
Swift Package Index builds our docs for us and will automatically insert
the dependency.
Modifications:
- Remove the docc-plugin dependency
Result:
Fewer dependencies
### Motivation:
Allow different devices to leverage the capabilities of Mutipath TCP
(MPTCP) to enhance the network reliability. Thanks to MPTCP, a
connection can for example automatically migrate to another interface if
it deteriorates on the first one. This can be especially interesting on
MacOS X and iOS, where devices may frequently benefit from multiple
interfaces (ethernet + Wi-Fi for Macs and Wi-fi + cellular for iOS).
Allowing developers to enable MPTCP on their connections seems thus like
a fine addition to this library.
### Modifications:
Added a function "withMultipath" on NIOTSConnectionBootstrap, that allow
to configure the type of service used for MPTCP (defaults to disabled).
This value will be stored in a field, and then propagated to the
underlying channel when the connect method will be called. Also updated
the parameters field of NIOTSConnectionChannel to set the
multipathServiceType accordingly.
### Result:
Users will now be able to easily enable Multipath TCP, allowing them to
benefit from seamless handover, interactive mode to automatically use
the lowest delay interface or aggregate mode to send data in parallel on
both interfaces.
Example:
```swift
let group = NIOTSEventLoopGroup()
defer {
try! group.syncShutdownGracefully()
}
let bootstrap = NIOTSConnectionBootstrap(group: group)
.withMultipath(.handover) // set handover mode
.channelInitializer { channel in
channel.pipeline.addHandler(MyChannelHandler())
}
try! bootstrap.connect(host: "example.org", port: 12345).wait()
/* the Channel is now connected */
```
Co-authored-by: Cory Benfield <lukasa@apple.com>
# Enhancement: Integration of Viability Handler in `Network.framework`
We have successfully integrated the `viabilityUpdateHandler` from
`Network.framework` into `NIOTransportServices`, enabling developers to
leverage the full potential of the `Network.framework` API.
This enhancement provides adopters with critical insights into the
viability of connections to remote endpoints. By being informed about
the current network status, developers can implement a more responsive
and adaptive approach to network behavior and conditions.
## Modifications
- Introduced two new methods along with corresponding calls to these
methods within the `Network.framework` handler.
- These handlers trigger the `NIO` *InboundEventsTriggered* method,
allowing consumers to respond to network changes effectively.
- Developed two structs within *NIOTSNetworkEvents* to accompany the
*InboundEventsTriggered* methods, ensuring seamless data transfer.
- Conformed these structs to Sendability for compatibility with the
required operating systems.
## Result
Clients can now receive viability events from `Network.framework`,
enhancing their ability to manage network connections dynamically
---------
Co-authored-by: George Barnett <gbarnett@apple.com>
Co-authored-by: Cory Benfield <lukasa@apple.com>
Motivation:
In some cases users may want to access APIs we haven't exposed
in NIOTransportServices. We should have a fallback that allows users
to do this.
Modifications:
- Add ChannelOptions for getting NWConnection and NWListener.
Result:
Users have an escape hatch
Motivation:
Looks like when we previously added syncOptions support to our
channels, we had a few issues. The listeners had code added, but
the code never worked. This is because the code was defined in
subclasses, but the protocol conformance comes from the parent class,
and that parent class did not have a customized protocol witness.
The datagram channel was also entirely missing its support.
Modifications:
- Added the missing unit tests for sync options.
- Added syncOptions to StateManagedListenerChannel base class.
- Added overrides to the listener subclasses.
- Added syncOptions to datagram channel
Result:
Sync options actually work across the board.