Concurrent Task Extension
Clone or download
Type Name Latest commit message Commit time
Failed to load latest commit information.
examples Reworked XP sockets and TLS encryption. (#100) Jan 19, 2019
include Reworked XP sockets and TLS encryption. (#100) Jan 19, 2019
src Reworked XP sockets and TLS encryption. (#100) Jan 19, 2019
tests Reworked XP sockets and TLS encryption. (#100) Jan 19, 2019
thirdparty Full refactoring of all internals including task scheduling. Dec 10, 2018
.gitattributes Added composer.json file and license to the project. Jul 13, 2018
.gitignore Integration of libuv (#53) Jul 27, 2018
.travis.yml Reworked valgrind support. Switched to stable PHP 7.3 Dec 10, 2018
Dockerfile Reworked valgrind support. Switched to stable PHP 7.3 Dec 10, 2018
LICENSE Added composer.json file and license to the project. Jul 13, 2018
Makefile.frag Full refactoring of all internals including task scheduling. Dec 10, 2018 Reworked XP sockets and TLS encryption. (#100) Jan 19, 2019
Vagrantfile Support for async file IO (#82) Nov 8, 2018
config.m4 Implemented Channels. (#98) Jan 5, 2019
config.w32 Implemented Channels. (#98) Jan 5, 2019
package.php Full refactoring of all internals including task scheduling. Dec 10, 2018
package.xml Implemented Channels. (#98) Jan 5, 2019
php_async.c Reworked XP sockets and TLS encryption. (#100) Jan 19, 2019
php_async.h Optimized operation reuse, error handling and callback referencing. Jan 12, 2019 Reworked valgrind support. Switched to stable PHP 7.3 Dec 10, 2018

Async Extension for PHP

Build Status Coverage Status

Provides concurrent Zend VM executions using native C fibers in PHP.


At least PHP 7.3.0 RC5 is required to compile and use the extension with async file IO support!

The async extension is not published as pecl extension (yet).

Linux / MacOS

You have to install libuv using Homebrew in order to install the extension for Mac OS. It is advisable to also install openssl in order to establish encrypted TCP socket connections.

brew install libuv
brew install openssl

Installing ext-async from source works like this (assuming you have PHP 7.3 installed):

mkdir ext-async
curl -LSs | tar -xz -C "ext-async" --strip-components 1

pushd ext-async
make install

These are the options supported by configure:

Option Description
--with-openssl-dir=DIR Allows you to specify the directory where libssl-dev is installed.
--with-valgrind[=DIR] Can be used to enable Valgrind support and (optional) specify the valgrind directory.


You can download a pre-compiled DLL working with PHP 7.3+ (VC15 x64 Thread Safe only). Just drop the DLL file in your ext directory and add extension=php_async.dll in your php.ini file.

Building the extension on Windows requires the PHP-SDK for Windows. After the SDK is installed you have to use the starter batch file for you platform and Visual Studio version. In the command prompt navigate to the php-src directory and execute these commands and build PHP with the async extension as DLL file:

configure --disable-all --enable-cli --enable-async=shared

You can leave out =shared to compile the extension into PHP (faster compilation and no need to load it in php.ini) which is nice for testing but does not create a DLL file that can be distributed. You should add --enable-debug to allow for debugging when you are working on the source code of the extension.

INI Settings

Setting Description
async.dns Replaces some internal function (gethostbyname() and gethostbynamel()) with async implementations.
async.filesystem Replaces PHP's file stream wrapper with an async implementation.
async.tcp (experimental) Replaces PHP's tcp and tls stream wrappers with async implementations.
async.timer Replaces PHP's sleep() function with an async implementation.
async.udp (experimental) Replaces PHP's udp stream wrapper with an async implementation.

Async API

The async extension exposes a public API that can be used to create, run and interact with fiber-based async executions. You can obtain the API stub files for code completion in your IDE by installing concurrent-php/async-api via Composer.

Stream Wrappers

The async extension provides async-tcp, async-tls and async-udp stream wrappers that can be used to create async PHP stream resources. Use async-tcp://{server}:{port}/ with stream_socket_client() to establish a PHP stream that is backed by ext-async and does non-blocking IO (this is not related to stream_set_blocking()). You can also use INI settings async.tcp and async.udp to replace PHP's default stream implementations with their async counterpart which eliminates the need to prefix protocol names with async-.

Async stream wrappers have (limited) support for TLS encryption using stream context options:

Option Implementation Status
crypto_method Ignored, ext async will always use TLS 1.2 and automatically detect server / client mode.
peer_name Supported, fallback to host name being used to establish the connection.
verify_peer Always enabled!
verify_peer_name Always enabled!
allow_self_signed Supported, defaults to false.
cafile Not supported
capath Not supported
local_cert Supported but requires a the private key to be in a file specified by local_pk.
local_pk Supported
passphrase Supported
CN_match Ignored (use peer_name instead).
verify_depth Supported
ciphers Not supported
capture_peer_cert Not supported
capture_peer_cert_chain Not supported
SNI_enabled Supported, defaults to false.
SNI_server_name Ignored (use peer_name instead).
disable_compression Ignored, compression is always disabled to protect against CRIME attacks.
peer_fingerprint Not supported


This interface cannot be implemented directly by userland classes, implementations are provided by Deferred::awaitable() and Task. Awaitable is exposed as a union type to enable proper type hinting.

namespace Concurrent;

interface Awaitable { }


A deferred is a placeholder for an async operation that can be succeeded or failed from userland. It can be used to implement combinator function that operate on multiple Awaitable and expose a single Awaitable as result. The value returned from awaitable() is meant to be consumed by other tasks (or deferreds). The Deferred object itself must be kept private to the async operation because it can either succeed or fail the awaitable.

Each Deferred may specify a cancellation callback as constructor argument. The callback will be triggered when the current Context is cancelled. It receives the Deferred object as first argument and the cancellation error as second argument. You must not throw an error from the callback, doing so will trigger a fatal error and terminate the script.

namespace Concurrent;

final class Deferred
    public function __construct(callable $cancel = null) { }
    public function awaitable(): Awaitable { }
    public function resolve($val = null): void { }
    public function fail(\Throwable $e): void { }
    public static function value($val = null): Awaitable { }
    public static function error(\Throwable $e): Awaitable { }
    public static function combine(array $awaitables, callable $continuation): Awaitable { }
    public static function transform(Awaitable $awaitable, callable $transform): Awaitable { }


A task is a fiber-based object that executes a PHP function or method on a separate call stack. Tasks are created using Task::async() or TaskScheduler::run() (and their contextual counterparts). All tasks are associated with a task scheduler as they are created, there is no way to migrate tasks between different schedulers.

Calling Task::await() will suspend the current task and await resolution of the given Awaitable. If the awaited object is another Task it has to be run on the same scheduler, otherwise await() will throw an error.

namespace Concurrent;

final class Task implements Awaitable
    public static function isRunning(): bool { }
    /* Should be replaced with async keyword if merged into PHP core. */
    public static function async(callable $callback, ...$args): Task { }
    /* Should be replaced with extended async keyword expression if merged into PHP core. */
    public static function asyncWithContext(Context $context, callable $callback, ...$args): Task { }
    /* Should be replaced with await keyword if merged into PHP core. */
    public static function await(Awaitable $awaitable): mixed { }


The task scheduler manages a queue of ready-to-run tasks and a (shared) event loop that provides support for timers and async IO. It will also keep track of suspended tasks to allow for proper cleanup on shutdown. There is an implicit default scheduler that will be used when Task::async() or Task::asyncWithContext() is used in PHP code that is not run using one of the public scheduler methods. It is neighter necessary (nor advisable) to create a task scheduler instance yourself. The only exception to that rule are unit tests, each test should use a dedicated task scheduler to ensure proper test isolation.

You can use run() or runWithContext() to have the given callback be executed as root task within an isolated task scheduler. The run methods will return the value returned from your task callback or throw an error if your task callback throws. The scheduler will allways run all scheduled tasks to completion, even if the callback task you passed is completed before other tasks. The optional inspection callback will be called as soon as the root task (= the callback) is completed and receive an array containing information about all tasks that have not been completed yet.

namespace Concurrent;

final class TaskScheduler
    public static function run(callable $callback, ?callable $inspect = null): mixed { }
    public static function runWithContext(Context $context, callable $callback, ?callable $inspect = null): mixed { }


You can use Channel objects to exchange messages between different tasks. Every call to send() will pause the calling task until the message has been received by another task. A channel can buffer a number of messages specified by $capacity within the constructor. If capacity is bigger than 0 calls to send() will succedd immediately if there is some space left in the channel's buffer.

Reading from a channel is done using PHP's iterator API. You can use foreach to iterate over a channel object. Be sure to call getIterator() and return just the created ChannelIterator if you want to expose the contents of a channel. This way you prevent thirdparty code from calling close() or sending messages into your channel.

You can use isReadyForSend() to check if a value can be sent into the channel without blocking, this is possible when there is space left in the channel's buffer or a receive operation is pending. Likewise you can use isReadyForReceive() to check if a message can be received without blocking (this is true when a message has been buffered or a send operation is pending).

namespace Concurrent;

final class Channel implements \IteratorAggregate
    public function __construct(int $capacity = 0) { }
    public function close(?\Throwable $e = null): void { }
    public function isClosed(): bool { }
    public function isReadyForSend(): bool { }
    public function isReadyForReceive(): bool { }
    public function send($message): void { }


Working with multiple channels at once can be done by using a ChannelGroup. The group provides the very useful select() call that allows to read from multiple channels concurrently, it will return once a message has been received from any channel that is part of the group. You can use the $timeout parameter to specify a maximum wait time (in milliseconds). If no message has been received select() will return NULL. Whenever a message has been received select() will return the index of the channel in the wrapped $channel array. You have to pass a variable by reference to select() if you want to receive the message payload (this is not necessary if you just want to know the origin of the message but you do not care about the actual payload).

The constructor accepts a $channels array that must contain eighter Channel objects or objects that implement IteratorAggregate and return a ChannelIterator object when getIterator() is called. You can call count() to check how many of the wrapped channels are still open and therefore considered to be readable. Keep in mind that you need to call select() before checking the count because channels can only be checked for closed state during select() (consider using a do / while loop and check count() as the loop condition). Closed channels are silently removed from the group if they are not closed with an error. If any of the input channels is closed with an error it will be forwarded exactly once by select() after that the closed channel will be removed!

namespace Concurrent;

final class ChannelGroup implements \Countable
    public function __construct(array $channels, ?int $timeout = null, bool $shuffle = false) { }
    public function select(& $value = null) { }


Each async operation is associated with a Context object that provides a logical execution context. The context can be used to carry execution-specific variables and (in a later revision) cancellation signals across API boundaries. The context is immutable, a new context must be derived whenever anything needs to be changed for the current execution. You can pass a Context to Task::asyncWithContext() or TaskScheduler::runWithContext() that will become the current context for the new task. It is also possible to enable a context for the duration of a callback execution using the run() method. Every call to Task::await() will backup the current context and restore it when the task is resumed.

namespace Concurrent;

final class Context
    public readonly bool $cancelled;

    public function with(ContextVar $var, $value): Context { }
    public function withTimeout(int $milliseconds): Context { }
    public function withCancel(& $cancel): Context { }
    public function shield(): Context { }
    public function throwIfCancelled(): void { }
    public function run(callable $callback, ...$args): mixed { }
    public static function current(): Context { }
    public static function background(): Context { }


You can access contextual data using a ContextVar object. Calling get() will lookup the variable's value from the context (passed as argument, current context by default). You have to use Context::with() to derive a new Context that has a value bound to the variable.

namespace Concurrent;

final class ContextVar
    public function get(?Context $context = null) { }


You have to call Context::withCancel(& $cancel) to create a new Context (returned value) and a CancellationHandler placed in $cancel (which has to be passed by reference). Invoking the handler will cancel the associated context, the optional error argument will registered as previous error with the thrown CancellationException.

namespace Concurrent;

final class CancellationHandler
    public function __invoke(?\Throwable $e = null): void { }

Watcher API


The Timer class is used to schedule timers with the integrated event loop. Timers do not make use of callbacks, instead they will suspend the current task during awaitTimeout() and continue when the next timeout is exceeded. The first call to awaitTimeout() will start the timer. If additional tasks await an active the timer they will share the same timeout (which could be less than the value passed to the constructor). A Timer can be closed by calling close() which will fail all pending timeout subscriptions and prevent any further operations.

namespace Concurrent;

final class Timer
    public function __construct(int $milliseconds) { }
    public function close(?\Throwable $e = null): void { }
    public function awaitTimeout(): void { }


A StreamWatcher observes a PHP stream or socket for readability or writability. Only a single stream watcher is allowed for any PHP resource. The watcher should be closed when it is no longer needed to free internal resources. The StreamWatcher will suspend the current task during awaitReadable() and awaitWritable() and continue once the watched stream becomes readable or is closed by the remote peer. A StreamWatcher can be used simultaneously (by multiple tasks) to await both read and write events, all tasks will be continued (in the same order as they entered await) once the stream is readable / writable. A StreamWatcher can be closed by calling close() which will fail all pending read & write subscriptions and prevent any further operations.

namespace Concurrent;

final class Watcher
    public function __construct($resource) { }
    public function close(?\Throwable $e = null): void { }
    public function awaitReadable(): void { }
    public function awaitWritable(): void { }


A SignalWatcher observes UNIX signals (limited support on Windows). The watcher should be closed when it is no longer needed to free internal resources. The current task will be suspended during calls to awaitSignal() and continue once the signal has been received. Multiple tasks can await a signal at the same time, all of them will be continued when the signal has been received. You can use isSupported() to check if the passed signal can be observed. Windows systems only support SIGHUP (console window closed) and SIGINT (CTRL + C) handling.

namespace Concurrent;

final class SignalWatcher
    public const SIGHUP;
    public const SIGINT;
    public const SIGQUIT;
    public const SIGKILL;
    public const SIGTERM;
    public const SIGUSR1;
    public const SIGUSR2;

    public function __construct(int $signum) { }
    public function close(?\Throwable $e = null): void { }
    public function awaitSignal(): void { }
    public static function isSupported(int $signum): bool { }

Stream API

The stream API provides an object-oriented interface to arbitrary byte streams. The stream API is closely aligned with fclose(), fread() and fwrite() functions.


A readable stream provides access to chunks of incoming data. There is no method to check for EOF, a call to read() will return null when the stream is at EOF. The (optional) $length argument can be used to specify the maximum number of bytes to be returned, a stream might return fewer bytes depending on network IO or internal buffers. Every call to read() must return at least one bytes, or null if no more bytes can be read (EOF). The optional error argument of close() allows to pass in an error that will be set as previous error when failing a read operation. Calling close() will fail all pending read operations and prevent any further reads from the stream by throwing a StreamClosedException.

Only one pending read operation is allowed on a ReadableStream at any time. Calls to read() must throw a PendingReadException if an attempt is made to read from a stream before all previous reads have completed.

namespace Concurrent\Stream;

interface ReadableStream
    public function close(?\Throwable $e = null): void;
    public function read(?int $length = null): ?string;


A writable stream allows to write chunks of data in sequence. Multiple calls to write() from different tasks at the same time are allowed, stream implementations must preserve order of write operations. The optional error argument of close() allows to pass in an error that will be set as previous error when failing a write operation. Calling close() will fail all pending write operations and prevent any further writes from the stream.

namespace Concurrent\Stream;

interface WritableStream
    public function close(?\Throwable $e = null): void;
    public function write(string $data): void;


The duplex stream implements both ReadableStream and WritableStream. Streams backed by a socket will usually be compatible with (and implement) this interface. You can call getReadableStream() or getWritableStream() to aquire a stream that is restricted to one of the combined interfaces. This is especially useful if you want calls to close() to result in a half-closed stream.

namespace Concurrent\Stream;

interface DuplexStream extends ReadableStream, WritableStream
    public function getReadableStream(): ReadableStream;
    public function getWritableStream(): WritableStream;

Network API

The network API provides access to stream and datagram sockets.


Defines the basic API that every socket-based component exposes.

namespace Concurrent\Network;

interface Socket
    public function close(?\Throwable $e = null): void;

    public function getAddress(): string;

    public function getPort(): ?int;
    public function setOption(int $option, $value): bool;


Defines the API being provided by reliable connected socket streams.

namespace Concurrent\Network;

use Concurrent\Stream\DuplexStream;

interface SocketStream extends Socket, DuplexStream
    public function getRemoteAddress(): string;

    public function getRemotePort(): ?int;
    public function writeAsync(string $data): int;
    public function getWriteQueueSize(): int;


Provides the minimal API being exposed by reliable socket servers.

namespace Concurrent\Network;

interface Server extends Socket
    public function accept(): SocketStream;


A TcpSocket wraps a TCP network conneciton. It implements DuplexStream to provide access based on the stream API. Closing a TCP socket will close both read and write sides of the stream. You can use getWritableStream() to aquire the writer and call close() on it to signal the remote peer that the stream is half-closed, you can still read data from the remote peer until the stream is closed by the remote peer.

namespace Concurrent\Network;

final class TcpSocket implements SocketStream
    public const NODELAY;
    public const KEEPALIVE;

    public static function connect(string $host, int $port, ?TlsClientEncryption $tls = null): TcpSocket { }
    public static function pair(): array { }
    public function encrypt(): void { }


A TcpServer listens on a local port for incoming TCP connection attempts until close() is called to terminate the server socket. You have to call accept() to accept the next pending connection attempt. Each accepted connection is wrapped in a TcpSocket that can be used to communicate with the remote peer. Accepted socket connections are not closed when the server is closed, they have to be closed individually by calling close() on the TcpSocket object.

namespace Concurrent\Network;

final class TcpServer implements Server
    public const SIMULTANEOUS_ACCEPTS;
    public static function listen(string $host, int $port, ?TlsServerEncryption $tls = null): TcpServer { }


Configures an encrypted (TLS) socket client.

namespace Concurrent\Network;

final class TlsClientEncryption
    public function withAllowSelfSigned(bool $allow): TlsClientEncryption { }
    public function withVerifyDepth(int $depth): TlsClientEncryption { }
    public function withPeerName(string $name): TlsClientEncryption { }


Configures an encrypted (TLS) socket server.

namespace Concurrent\Network;

final class TlsServerEncryption
    public function withDefaultCertificate(string $cert, string $key, ?string $passphrase = null): TlsServerEncryption { }
    public function withCertificate(string $host, string $cert, string $key, ?string $passphrase = null): TlsServerEncryption { }


Provides UDP networking capabilities.

namespace Concurrent\Network;

final class UdpSocket implements Socket
    public const TTL;
    public const MULTICAST_LOOP;
    public const MULTICAST_TTL;
    public static function bind(string $address, int $port): UdpSocket { }
    public static function multicast(string $group, int $port): UdpSocket { }
    public function receive(): UdpDatagram { }
    public function send(UdpDatagram $datagram): void { }
    public function sendAsync(UdpDatagram $datagram): int { }


Wraps a UDP datagram into a single object to allow for better type-hinting and a single return value.

namespace Concurrent\Network;

final class UdpDatagram
    public readonly string $data;
    public readonly string $address;
    public readonly int $port;

    public function __construct(string $data, string $address, int $port) { }
    public function withData(string $data): UdpDatagram { }
    public function withPeer(string $address, int $port): UdpDatagram { }

Process API

The process API provides tools to spawn processes and communicate with them. This includes setting the work directory, setting environment variables, dealing with input / output, support for signals (limited support on Windows) and awaiting termination (including access to the exit code).


The ProcessBuilder is used to configure the execution environment of a spawned process. The constructor takes the command to be executed, additional arguments can be passed to execute() or start(). The current working directory of the spawned process can be changed using setDirectory(), the process will inherit the current working directory by default. You can specify env vars to be passed to the process using setEnv(), inheritance of all current env vars can be configured using inheritEnv() (inheritance is enabled by default, variables specified in setEnv() will be added to the inherited vars).

Each process is spawned with access to three anonymous pipes (STDIN, STDOUT and STDERR). The behavior of these pipes can be configured using the configure*($mode, $fd) methods of the ProcessBuilder. By default all pipes will be ignored (redirected to eighter /dev/null or NUL) providing no access to process IO. It is possible to have pipes use the same pipe as the process that is spawning a new process, use STDIO_INHERIT as $mode and one of the ProcessBuilder class constants (STDIN, STDOUT and STDERR) to configure inheritance. Both STDIO_IGNORE and STDIO_INHERIT handle IO automatically, there is no way to control or access the pipes. If you need to write or read data to / from a pipe you need to configure it using STDIO_PIPE as mode. Doing so will provide access to the various pipes as async streams that can be accessed using the Process object.

A call to execute() will run the spawned process to completion and return with the exit code of the process. This method can only be used if all pipes are configured with mode STDIO_IGNORE or STDIO_INHERIT. If you need to do pipe IO, access the process PID or signal the process you need to call start() instead which will return a Process object that provides access to a running process.

namespace Concurrent\Process;

final class ProcessBuilder
    public const STDIN;
    public const STDOUT;
    public const STDERR;
    public const STDIO_IGNORE;
    public const STDIO_INHERIT;
    public const STDIO_PIPE;
    public function __construct(string $command, string ...$args) { }
    public function setDirectory(string $dir): void { }
    public function setEnv(array $env): void { }
    public function inheritEnv(bool $inherit): void { }
    public function configureStdin(int $mode, ?int $fd = null): void { }
    public function configureStdout(int $mode, ?int $fd = null): void { }
    public function configureStderr(int $mode, ?int $fd = null): void { }
    public function execute(string ...$args): int {}
    public function start(string ...$args): Process {}


The Process class provides access to a started process. You can use isRunning() to check if the process has terminated yet. The process identifier can be accessed using getPid(). Iy any pipe was configured using STDIO_PIPE it will be accessible via the corresponding getter method. You can send a signal to the process using signal(), on Windows systems only SIGHUP and SIGINT are supported (you should use the class constants defined in SignalWatcher to avoid magic numbers). Calling join() will suspend the current task until the process has terminated and return the exit code of the process.

namespace Concurrent\Process;

use Concurrent\Stream\ReadableStream;
use Concurrent\Stream\WritableStream;

final class Process
    public function isRunning(): bool { }
    public function getPid(): int { }
    public function getStdin(): WritableStream { }
    public function getStdout(): ReadableStream { }
    public function getStderr(): ReadableStream { }
    public function signal(int $signum): void { }
    public function join(): int { }


A lower-level API for concurrent callback execution is available through the Fiber API. The underlying stack-switching is the same as in the Task implementation but fibers do not come with a scheduler or a higher level abstraction of continuations. A fiber must be started and resumed by the caller in PHP userland. Calling Fiber::yield() will suspend the fiber and return the yielded value to start(), resume() or throw(). The status() method is needed to check if the fiber has been run to completion yet.

namespace Concurrent;

final class Fiber
    public const STATUS_INIT;
    public const STATUS_SUSPENDED;
    public const STATUS_RUNNING;
    public const STATUS_FINISHED;
    public const STATUS_FAILED;
    public readonly int $status;
    public function __construct(callable $callback, ?int $stack_size = null) { }
    public function start(...$args): mixed { }
    public function resume($val = null): mixed { }
    public function throw(\Throwable $e): mixed { }
    public static function yield($val = null): mixed { }
    public static function isRunning(): bool { }
    public static function backend(): string { }

Async / Await Keyword Transformation

The extension provides Task::async() and Task::await() static methods that are implemented in a way that allows for a very simple transformation to the keywords async and await which could be introduced into PHP some time in the future.

$task = async $this->sendRequest($request, $timeout);
$response = await $task;

// The above code would be equivalent to the following:
$task = Task::async(Closure::fromCallable($this, 'sendRequest'), $request, $timeout);
$response = Task::await($task);

The example shows a possible syntax for a keyword-based async execution model. The async keyword can be prepended to any function or method call to create a Task object instead of executing the call directly. The calling scope should be preserved by this operation, hence being consistent with the way method calls work in PHP (no need to create a closure in userland code). The await keyword is equivalent to calling Task::await() but does not require a function call, it can be implemented as an opcode handler in the Zend VM.

$context = Context::inherit(['foo' => 'bar']);

$task = async $context => doSomething($a, $b);
$result = await $task;

// The above code would be equivalent to the following:
$task = Task::asyncWithContext($context, 'doSomething', [$a, $b]);
$result = Task::await($task);

The second example shows how passing a new context to a task would also be possible using the async keyword. This would allow for a very simple and readable way to setup tasks in a specific context using a keyword-based syntax.

PHP with support for async & await keywords

You can install a patched version of PHP that provides native support for async and await as described in the transformation section. To get up and running with it you can execute this in your shell:

mkdir php-src
curl -LSs | tar -xz -C "php-src" --strip-components 1

pushd php-src
./buildconf --force
./configure --prefix=/usr/local/php/cli --with-config-file-path=/usr/local/php/cli --without-pear
make -j4
make install

mkdir ext-async
curl -LSs | tar -xz -C "ext-async" --strip-components 1

pushd ext-async
make install

This will install a modified version of PHP's master branch that has full support for async and await. It will also install the async extension that is required for the actual async execution model.

Source Transformation Examples

The source transformation needs to consider namespaces and preserve scope. Dealing with namespaces is a problem when it comes to function calls because there is a fallback to the global namespace involved and there is no way to determine the called function in all cases during compilation. Here are some examples of code using async / await syntax and the transformed source code:

$task = async max(1, 2, 3);
$result = await $task;

$task = \Concurrent\Task::async('max', 1, 2, 3);
$result = \Concurrent\Task::await($task);

Function calls in global namespace only have to check imported functions, the correct function can be determined at compile time.

namespace Foo;

$task = async bar(1, 2);

$task = \Concurrent\Task::async(\function_exists('Foo\\bar') ? 'Foo\\bar' : 'bar', 1, 2);

Unqualified function calls in namespace require runtime evaluation of the function to be called (unless the function is imported via use statement).

namespace Foo;

$context = \Concurrent\Context::inherit(['num' => 321]);
$work = function (int $a): int { return $a + Context::var('num'); };

$result = await async $context => $work(42);

$result = \Concurrent\Task::await(\Concurrent\Task::asyncWithContext(\Closure::fromCallable($work), 42));

Calling functions stored in variables requires to keep track of the calling scope because $work might contain a method call (or an object with __invoke() method) with a visibility other than public.