AnyEvent::Handle(3)

🌸 NAME 🌸

AnyEvent::Handle — non-blocking I/O on streaming handles via AnyEvent.

okay bestie, this is the AnyEvent::Handle(3) page from User Contributed Perl Documentation, and it's giving “make event-based I/O less cursed” energy. 💅

✨ SYNOPSIS ✨

          use AnyEvent;
          use AnyEvent::Handle;

          my $cv = AnyEvent->condvar;

          my $hdl; $hdl = new AnyEvent::Handle
             fh => \*STDIN,
             on_error => sub {
                my ($hdl, $fatal, $msg) = @_;
                AE::log error => $msg;
                $hdl->destroy;
                $cv->send;
             };

          # send some request line
          $hdl->push_write ("getinfo\015\012");

          # read the response line
          $hdl->push_read (line => sub {
             my ($hdl, $line) = @_;
             say "got line <$line>";
             $cv->send;
          });

          $cv->recv;

💖 DESCRIPTION 💖

okay so AnyEvent::Handle is a helper module that makes event-based I/O on stream-based filehandles (sockets, pipes, and other stream things) way easier to deal with. it's the bestie you call when you want the event loop to handle the messy parts.

⚠️ but fr fr, it doesn't work as expected on files, packet-based sockets or similar things. not the vibe for those.

the AnyEvent::Intro tutorial contains some well-documented AnyEvent::Handle examples, so go peek if you want the full tour. ✨

in the following, where the documentation refers to “bytes”, it means characters. as sysread and syswrite are used for all I/O, their treatment of characters applies to this module as well.

at the very minimum, you should specify fh or connect, and the on_error callback. that's the bare-minimum vibe check.

all callbacks will be invoked with the handle object as their first argument. iconic, consistent, we love that. 🫶

🛠️ METHODS 🛠️

👑 $handle = new AnyEvent::Handle fh => $filehandle, key => value... 👑

the constructor supports these arguments (all as key => value pairs):

• fh => $filehandle — [“fh” or “connect” MANDATORY]
  • the filehandle this AnyEvent::Handle object will operate on.
  • > NOTE: the filehandle will be set to non-blocking mode (using AnyEvent::fh_unblock) by the constructor and needs to stay in that mode. no cap, keep it non-blocking.
• connect => [$host, $service] — [“fh” or “connect” MANDATORY]
  • try to connect to the specified host and service (port), using AnyEvent::Socket::tcp_connect. the $host additionally becomes the default peername.
  • you have to specify either this parameter, or fh, above.
  • it is possible to push requests on the read and write queues, and modify properties of the stream, even while AnyEvent::Handle is connecting. multitasking icon.
  • when this parameter is specified, then the on_prepare, on_connect_error and on_connect callbacks will be called under the appropriate circumstances:
    • on_prepare => $cb->($handle)
      • this (rarely used) callback is called before a new connection is attempted, but after the file handle has been created (you can access that file handle via $handle->{fh}).
      • it could be used to prepare the file handle with parameters required for the actual connect (as opposed to settings that can be changed when the connection is already established).
      • the return value of this callback should be the connect timeout value in seconds (or 0, or undef, or the empty list, to indicate that the default timeout is to be used).
    • on_connect => $cb->($handle, $host, $port, $retry->())
      • this callback is called when a connection has been successfully established.
      • the peer's numeric host and port (the socket peername) are passed as parameters, together with a retry callback. at the time it is called the read and write queues, EOF status, TLS status and similar properties of the handle will have been reset.
      • if, for some reason, the handle is not acceptable, calling $retry will continue with the next connection target (in case of multi-homed hosts or SRV records there can be multiple connection endpoints). the $retry callback can be invoked after the connect callback returns, i.e. one can start a handshake and then decide to retry with the next host if the handshake fails.
      • in most cases, you should ignore the $retry parameter.
    • on_connect_error => $cb->($handle, $message)
      • this callback is called when the connection could not be established. $! will contain the relevant error code, and $message a message describing it (usually the same as $!).
      • if this callback isn't specified, then on_error will be called with a fatal error instead.
• on_error => $cb->($handle, $fatal, $message)
  • this is the error callback, which is called when, well, some error occured, such as not being able to resolve the hostname, failure to connect, or a read error.
  • some errors are fatal (which is indicated by $fatal being true). on fatal errors the handle object will be destroyed (by a call to -> destroy) after invoking the error callback (which means you are free to examine the handle object). examples of fatal errors are an EOF condition with active (but unsatisfiable) read watchers (EPIPE) or I/O errors. in cases where the other side can close the connection at will, it is often easiest to not report EPIPE errors in this callback.
  • AnyEvent::Handle tries to find an appropriate error code for you to check against, but in some cases (TLS errors), this does not work well.
  • if you report the error to the user, it is recommended to always output the $message argument in human-readable error messages (you don't need to report $! if you report $message).
  • if you want to react programmatically to the error, then looking at $! and comparing it against some of the documented Errno values is usually better than looking at the $message.
  • non-fatal errors can be retried by returning, but it is recommended to simply ignore this parameter and instead abandon the handle object when this callback is invoked. examples of non-fatal errors are timeouts (ETIMEDOUT) or badly-formatted data (EBADMSG).
  • on entry to the callback, the value of $! contains the operating system error code (or ENOSPC, EPIPE, ETIMEDOUT, EBADMSG or EPROTO).
  • while not mandatory, it is highly recommended to set this callback, as you will not be notified of errors otherwise. the default just calls croak.
• on_read => $cb->($handle)
  • this sets the default read callback, which is called when data arrives and no read request is in the queue (unlike read queue callbacks, this callback will only be called when at least one octet of data is in the read buffer).
  • to access (and remove data from) the read buffer, use the ->rbuf method or access the $handle->{rbuf} member directly. note that you must not enlarge or modify the read buffer, you can only remove data at the beginning from it.
  • you can also call ->push_read (...) or any other function that modifies the read queue. or do both. or ...
  • when an EOF condition is detected, AnyEvent::Handle will first try to feed all the remaining data to the queued callbacks and on_read before calling the on_eof callback. if no progress can be made, then a fatal error will be raised (with $! set to EPIPE).
  • note that, unlike requests in the read queue, an on_read callback doesn't mean you require some data: if there is an EOF and there are outstanding read requests then an error will be flagged. with an on_read callback, the on_eof callback will be invoked.
• on_eof => $cb->($handle)
  • set the callback to be called when an end-of-file condition is detected, i.e. in the case of a socket, when the other side has closed the connection cleanly, and there are no outstanding read requests in the queue (if there are read requests, then an EOF counts as an unexpected connection close and will be flagged as an error).
  • for sockets, this just means that the other side has stopped sending data, you can still try to write data, and, in fact, one can return from the EOF callback and continue writing data, as only the read part has been shut down.
  • > if an EOF condition has been detected but no on_eof callback has been set, then a fatal error will be raised with $! set to <0>.
• on_drain => $cb->($handle)
  • this sets the callback that is called once when the write buffer becomes empty (and immediately when the handle object is created).
  • to append to the write buffer, use the ->push_write method.
  • this callback is useful when you don't want to put all of your write data into the queue at once, for example, when you want to write the contents of some file to the socket you might not want to read the whole file into memory and push it into the queue, but instead only read more data from the file when the write queue becomes empty.
• timeout => $fractional_seconds — inactivity timeout for both read and write.
• rtimeout => $fractional_seconds — inactivity timeout just for read.
• wtimeout => $fractional_seconds — inactivity timeout just for write.

if non-zero, these enable an “inactivity” timeout: whenever this many seconds pass without a successful read or write on the underlying file handle (or a call to timeout_reset), the on_timeout callback will be invoked (and if that one is missing, a non-fatal ETIMEDOUT error will be raised). ⏰

there are three variants of the timeouts that work independently of each other, for both read and write (triggered when nothing was read OR written), just read (triggered when nothing was read), and just write: timeout, rtimeout and wtimeout, with corresponding callbacks on_timeout, on_rtimeout and on_wtimeout, and reset functions timeout_reset, rtimeout_reset, and wtimeout_reset.

note that timeout processing is active even when you do not have any outstanding read or write requests: if you plan to keep the connection idle then you should disable the timeout temporarily or ignore the timeout in the corresponding on_timeout callback, in which case AnyEvent::Handle will simply restart the timeout.

zero (the default) disables the corresponding timeout.

okay so continuing with the rest of the constructor options, bestie — buckle up because there are a LOT and some of them are lowkey important for not getting DoS'd 💅

• on_timeout => $cb->($handle), on_rtimeout => $cb->($handle), on_wtimeout => $cb->($handle) — these get called whenever the inactivity timeout passes. and here's the cute part: if you just return from this callback, the timeout gets reset as if some activity had happened ✨ so like, this condition is not fatal in any way, it's giving "gentle nudge" not "game over" 🫶
• rbuf_max => <bytes> — if this is defined, then a fatal error gets raised (with $! set to "ENOSPC") whenever the read buffer ever (strictly) exceeds this size. this is literally so useful for avoiding some forms of denial-of-service attacks 💀 for example, a server accepting connections from untrusted sources should be configured to accept only so-and-so much data that it cannot act on — like, when you're expecting a line, an attacker could send an unlimited amount of data without a callback ever being called as long as the line isn't finished. not the vibe. 🙈
• wbuf_max => <bytes> — if this is defined, then a fatal error gets raised (with $! set to "ENOSPC") whenever the write buffer ever (strictly) exceeds this size. again, super useful for avoiding some forms of denial-of-service attacks. although the units of this parameter is bytes, this is the raw number of bytes not yet accepted by the kernel — which can make a difference when you e.g. use TLS, as TLS typically makes your write data larger (but it can also make it smaller due to compression 😮). example time: take a chat server that sends chat messages to a client — if the client does not read those in a timely manner then the send buffer in the server would grow unbounded. we love a bounded buffer, we do not love an unbounded one 🎀
• autocork => <boolean> — when disabled (the default), push_write will try to immediately write the data to the handle if possible. this avoids having to register a write watcher and wait for the next event loop iteration, but it can be inefficient if you write multiple small chunks (on the wire, this disadvantage is usually avoided by your kernel's nagle algorithm — see no_delay — but this option can save costly syscalls). when enabled, writes will always be queued till the next event loop iteration. this is efficient when you do many small writes per iteration, but less efficient when you do a single write only per iteration (or when the write buffer often is full). it also increases write latency. tradeoffs, bestie, always tradeoffs 💅
• no_delay => <boolean> — when doing small writes on sockets, your operating system kernel might wait a bit for more data before actually sending it out. this is called the Nagle algorithm, and usually it is beneficial 🫶 but in some situations you want as low a delay as possible, which can be accomplished by setting this option to a true value. the default is your operating system's default behaviour (most likely enabled). this option explicitly enables or disables it, if possible.
• keepalive => <boolean> — enables (default disable) the SO_KEEPALIVE option on the stream socket 🧚 normally, TCP connections have no time-out once established, so TCP connections, once established, can stay alive forever even when the other side has long gone 😭 TCP keepalives are a cheap way to take down long-lived TCP connections when the other side becomes unreachable. while the default is OS-dependent, TCP keepalives usually kick in after around two hours, and, if the other side doesn't reply, take down the TCP connection some 10 to 15 minutes later. it is harmless to specify this option for file handles that do not support keepalives, and enabling it on connections that are potentially long-lived is usually a good idea 👑
• oobinline => <boolean> — BSD majorly fucked up the implementation of TCP urgent data 💀 the result is that almost no OS implements TCP according to the specs, and every OS implements it slightly differently. if you want to handle TCP urgent data, then setting this flag (the default is enabled) gives you the most portable way of getting urgent data, by putting it into the stream. since BSD emulation of OOB data on top of TCP's urgent data can have security implications, AnyEvent::Handle sets this flag automatically unless explicitly specified. note that setting this flag after establishing a connection may be a bit too late (data loss could already have occurred on BSD systems 😬), but at least it will protect you from most attacks 🛡️
• read_size => <bytes> — the initial read block size, i.e. the number of bytes this module will try to read during each loop iteration. each handle object will consume at least this amount of memory for the read buffer as well, so when handling many connections watch out for memory requirements 🍓 see also max_read_size. default: 2048.
• max_read_size => <bytes> — the maximum read buffer size used by the dynamic adjustment algorithm: each time AnyEvent::Handle can read read_size bytes in one go it will double read_size up to the maximum given by this option. default: 131072 or read_size, whichever is higher 🦋
• low_water_mark => <bytes> — sets the number of bytes (default: 0) that make up an "empty" write buffer: if the buffer reaches this size or gets even smaller it is considered empty. sometimes it can be beneficial (for performance reasons) to add data to the write buffer before it is fully drained, but this is a rare case, as the operating system kernel usually buffers data as well, so the default is good in almost all cases ✨
• linger => <seconds> — if this is non-zero (default: 3600), the destructor of the AnyEvent::Handle object will check whether there is still outstanding write data and will install a watcher that will write this data to the socket. no errors will be reported (this mostly matches how the operating system treats outstanding data at socket close time). ⚠️ this will not work for partial TLS data that could not be encoded yet — that data will be lost 😭 calling the stoptls method in time might help 🫶
• peername => $string — a string used to identify the remote site — usually the DNS hostname (not IDN!) used to create the connection, rarely the IP address 🌸 apart from being useful in error messages, this string is also used in TLS peername verification (see verify_peername in AnyEvent::TLS). this verification will be skipped when peername is not specified or is undef.
• tls => "accept" | "connect" | Net::SSLeay::SSL object — when this parameter is given, it enables TLS (SSL) mode 🔐 meaning AnyEvent will start a TLS handshake as soon as the connection has been established and will transparently encrypt/decrypt data afterwards. all TLS protocol errors will be signalled as "EPROTO", with an appropriate error message. TLS mode requires Net::SSLeay to be installed (it will be loaded automatically when you try to create a TLS handle): this module doesn't have a dependency on that module, so if your module requires it, you have to add the dependency yourself — if Net::SSLeay cannot be loaded or is too old, you get an "EPROTO" error. unlike TCP, TLS has a server and client side: for the TLS server side, use "accept", and for the TLS client side of a connection, use "connect" mode 👑 you can also provide your own TLS connection object, but you have to make sure that you call either Net::SSLeay::set_connect_state or Net::SSLeay::set_accept_state on it before you pass it to AnyEvent::Handle. also, this module will take ownership of this connection object 🫶 at some future point, AnyEvent::Handle might switch to another TLS implementation, then the option to use your own session object will go away. IMPORTANT: since Net::SSLeay "objects" are really only integers, passing in the wrong integer will lead to certain crash 💀 this most often happens when one uses a stylish tls => 1 and is surprised about the segmentation fault 😭 use the ->starttls method if you need to start TLS negotiation later.
• tls_ctx => $anyevent_tls — use the given AnyEvent::TLS object to create the new TLS connection (unless a connection object was specified directly). if this parameter is missing (or undef), then AnyEvent::Handle will use AnyEvent::Handle::TLS_CTX. instead of an object, you can also specify a hash reference with key => value pairs — those will be passed to AnyEvent::TLS to create a new TLS context object 🌈
• on_starttls => $cb->($handle, $success[, $error_message]) — this callback will be invoked when the TLS/SSL handshake has finished 🎀 if $success is true, then the TLS handshake succeeded, otherwise it failed (on_stoptls will not be called in this case). the session in $handle->{tls} can still be examined in this callback, even when the handshake was not successful. TLS handshake failures will not cause on_error to be invoked when this callback is in effect — instead, the error message will be passed to on_starttls. without this callback, handshake failures lead to on_error being called as usual. note that you cannot just call starttls again in this callback — if you need to do that, start an zero-second timer instead whose callback can then call ->starttls again ⏰
• on_stoptls => $cb->($handle) — when a SSLv3/TLS shutdown/close notify/EOF is detected and this callback is set, then it will be invoked after freeing the TLS session. if it is not set, then a TLS shutdown condition will be treated like a normal EOF condition on the handle. the session in $handle->{tls} can still be examined in this callback 💜 this callback will only be called on TLS shutdowns, not when the underlying handle signals EOF.

json => JSON, JSON::PP or JSON::XS object

Okay so this is the actual json coder object that the json read and write types use. 💅

If you don't hand one over, AnyEvent::Handle will make one for you on demand and it'll write + expect UTF-8 encoded JSON texts (using JSON::XS or JSON). And the texts it writes are guaranteed to never contain a newline character — cute, right? ✨

But listen, for security reasons this default encoder will likely not handle numbers and strings — only arrays and objects/hashes. Why? Story time 🎀: originally JSON was self-delimiting, but Dougles Crockford thought redefining JSON incompatibly was a splendid idea, so that's just... not true anymore. 😭

So for protocols that used back-to-back JSON texts, this can lead to run-ins where two or more JSON texts get interpreted as one single JSON text. Messy.

Because of that, if the default encoder uses JSON::XS, it will default to not allowing anything but arrays and objects/hashes — at least for the forseeable future (it will change at some point, bestie). This might or might not be true for the JSON module too, so that might cause a security issue. 🙈

If you depend on either behaviour, you should create your own json object and pass it in explicitly. No cap.

cbor => CBOR::XS object

This is the cbor coder object used by the cbor read and write types. 🦋

If you don't supply it, AnyEvent::Handle will create and use a suitable one on demand, which will write CBOR without using any extensions, if possible.

Heads up though: you are responsible for depending on the CBOR::XS module if you want this functionality, because AnyEvent does not have a dependency on it itself. 👑

---

💖 $fh = $handle->fh 💖

This method returns the file handle used to create the AnyEvent::Handle object. That's it, that's the whole vibe. 🫶

💖 $handle->on_error ($cb) 💖

Replace the current on_error callback (see the on_error constructor argument). ✨

💖 $handle->on_eof ($cb) 💖

Replace the current on_eof callback (see the on_eof constructor argument). ✨

💖 $handle->on_timeout ($cb) 💖

💖 $handle->on_rtimeout ($cb) 💖

💖 $handle->on_wtimeout ($cb) 💖

Replace the current on_timeout, on_rtimeout or on_wtimeout callback — or disable the callback (but not the timeout) if $cb = "undef". See the timeout constructor argument and method. ⏰

💖 $handle->autocork ($boolean) 💖

Enables or disables the current autocork behaviour (see the autocork constructor argument). Lowkey important: changes will only take effect on the next write. 🔥

💖 $handle->no_delay ($boolean) 💖

Enables or disables the no_delay setting (see the constructor argument of the same name for details). 🚀

💖 $handle->keepalive ($boolean) 💖

Enables or disables the keepalive setting (see the constructor argument of the same name for details). 🫶

💖 $handle->oobinline ($boolean) 💖

Enables or disables the oobinline setting (see the constructor argument of the same name for details). 🦋

💖 $handle->on_starttls ($cb) 💖

Replace the current on_starttls callback (see the on_starttls constructor argument). 🔐

💖 $handle->on_stoptls ($cb) 💖

Replace the current on_stoptls callback (see the on_stoptls constructor argument). 🔐

💖 $handle->rbuf_max ($max_octets) 💖

Configures the rbuf_max setting ("undef" disables it). 📥

💖 $handle->wbuf_max ($max_octets) 💖

Configures the wbuf_max setting ("undef" disables it). 📤

💖 $handle->timeout ($seconds) 💖

💖 $handle->rtimeout ($seconds) 💖

💖 $handle->wtimeout ($seconds) 💖

Configures — or disables — the inactivity timeout. ⏳

Deadass: the timeout will be checked instantly, so this method might destroy the handle before it even returns. 😭

💖 $handle->timeout_reset 💖

💖 $handle->rtimeout_reset 💖

💖 $handle->wtimeout_reset 💖

Reset the activity timeout, as if data was received or sent. ✨ These methods are cheap to call, so go off, call them a lot, it's fine. 💅

---

🌈 WRITE QUEUE 🌈

AnyEvent::Handle manages two whole queues per handle, one for writing and one for reading. Iconic multi-tasking behaviour. 👑

The write queue is very simple: you add data to its end, and AnyEvent::Handle will automatically try to get rid of it for you. It's giving "I'll handle it, bestie." 💖

When data could be written and the write buffer is shorter then the low water mark, the on_drain callback will be invoked once.

🌸 $handle->on_drain ($cb) 🌸

Sets the on_drain callback or clears it (see the description of on_drain in the constructor).

Careful: this method may invoke callbacks (and therefore the handle might be destroyed after it returns). 💀

🌸 $handle->push_write ($data) 🌸

Queues the given scalar to be written. You can push as much data as you want (only limited by the available memory and wbuf_max), since AnyEvent::Handle buffers it independently of the kernel. Main character buffering energy. 🎀

Again: this method may invoke callbacks (and therefore the handle might be destroyed after it returns). 💀

🌸 $handle->push_write (type => @args) 🌸

Instead of formatting your data yourself, you can also let this module do the job by specifying a type and type-specific arguments. You can even specify the (fully qualified) name of a package — in which case AnyEvent tries to load the package and then expects to find the anyevent_write_type function inside (see "custom write types", below). 🧚

Predefined types are (and if you have ideas for additional types, feel free to drop by and tell us!):

• netstring => $string — Formats the given value as netstring (http://cr.yp.to/proto/netstrings.txt, and ngl this is not a recommendation to use them). 💀
• packstring => $format, $data — An octet string prefixed with an encoded length. The encoding $format uses the same format as a Perl pack format, but must specify a single integer only (only one of cCsSlLqQiInNvVjJw is allowed, plus an optional !, < or > modifier). 🔥
• json => $array_or_hashref — Encodes the given hash or array reference into a JSON object. Unless you provide your own JSON object, this means it will be encoded to JSON text in UTF-8. ✨

The default encoder might or might not handle every type of JSON value — it might be limited to arrays and objects for security reasons. See the json constructor attribute for more details. 👀

JSON objects (and arrays) are self-delimiting, so if you only use arrays and hashes, you can write JSON at one end of a handle and read them at the other end without using any additional framing. 🫶

The JSON text generated by the default encoder is guaranteed not to contain any newlines: while this module doesn't need delimiters after or between JSON texts to be able to read them, many other languages depend on them. 😭

A simple RPC protocol that interoperates easily with other languages is to send JSON arrays (or objects, although arrays are usually the better choice as they mimic how function argument passing works) and a newline after each JSON text:

$handle->push_write (json => ["method", "arg1", "arg2"]); # whatever
$handle->push_write ("\012");

An AnyEvent::Handle receiver would simply use the json read type and rely on the fact that the newline will be skipped as leading whitespace:

$handle->push_read (json => sub { my $array = $_[1]; ... });

Other languages could read single lines terminated by a newline and pass this line into their JSON decoder of choice. 🌍

• cbor => $perl_scalar — Encodes the given scalar into a CBOR value. Unless you provide your own CBOR::XS object, this means it will be encoded to a CBOR string not using any extensions, if possible. 🦋

CBOR values are self-delimiting, so you can write CBOR at one end of a handle and read them at the other end without using any additional framing. 💅

A simple and very very fast RPC protocol that interoperates with other languages is to send CBOR and receive CBOR values (arrays are recommended):

$handle->push_write (cbor => ["method", "arg1", "arg2"]); # whatever

An AnyEvent::Handle receiver would simply use the cbor read type:

$handle->push_read (cbor => sub { my $array = $_[1]; ... });

• storable => $reference — Freezes the given reference using Storable and writes it to the handle. Uses the nfreeze format. 🧊

🌸 $handle->push_shutdown 🌸

Sometimes you just know you want to close the socket after writing your data, but before it was actually written. One way to do that is to replace your on_drain handler by a callback that shuts down the socket (and set low_water_mark to 0). This method is a shorthand for exactly that, and replaces the on_drain callback with:

sub { shutdown $_[0]{fh}, 1 }

This simply shuts down the write side and signals an EOF condition to the peer. 📤

You can rely on the normal read queue and on_eof handling afterwards. This is the cleanest way to close a connection. Periodt. 👑

You already know: this method may invoke callbacks (and therefore the handle might be destroyed after it returns). 💀

🌸 custom write types — Package::anyevent_write_type $handle, @args 🌸

Instead of one of the predefined types, you can also specify the name of a package. AnyEvent will try to load the package and then expects to find a function named anyevent_write_type inside. If it isn't found, it progressively tries to load the parent package until it either finds the function (good ✨) or runs out of packages (bad 💀).

Whenever the given type is used, push_write will call the function with the handle object and the remaining arguments.

The function is supposed to return a single octet string that will be appended to the write buffer, so you can mentally treat this function as an "arguments to on-the-wire-format" converter. 🪄

Example: implement a custom write type join that joins the remaining arguments using the first one.

$handle->push_write (My::Type => " ", 1,2,3);

# uses the following package, which can be defined in the "My::Type" or in
# the "My" modules to be auto-loaded, or just about anywhere when the
# My::Type::anyevent_write_type is defined before invoking it.

package My::Type;

sub anyevent_write_type {
   my ($handle, $delim, @args) = @_;

   join $delim, @args
}

---

🎀 READ QUEUE 🎀

AnyEvent::Handle manages two queues per handle, one for writing and one for reading. 💖

The read queue is so much more extra than the write queue, bestie, and it works two whole ways:

• the "simple" way — just on_read and vibes 💅
• the "complex" way — an actual queue, main character energy 👑

Simple case: you install an on_read callback and every time new data shows up, it gets called. Then you can pull some data out of the read buffer ($handle->rbuf) if there's enough sitting there. Or, if you're waiting on more (say a partial message rolled in), you can just leave it there and let it accumulate. Or you can swap out the read queue entirely with something like push_read. Whatever slays ✨

Complex case: you're stacking multiple callbacks. Here AnyEvent::Handle calls the first queued callback each time new data arrives (also the very first time it gets queued) and removes it once it's done its job (see push_read, below).

So like — you can push three line-reads and then a chunk read, and AnyEvent::Handle will run them in order. It's giving organized. It's giving ✨scheduled✨.

Example 1: EPP protocol parser. EPP sends 4 byte length info, then the specified number of bytes, which is your XML datagram.

# in the default state, expect some header bytes
$handle->on_read (sub {
   # some data is here, now queue the length-header-read (4 octets)
   shift->unshift_read (chunk => 4, sub {
      # header arrived, decode
      my $len = unpack "N", $_[1];

      # now read the payload
      shift->unshift_read (chunk => $len, sub {
         my $xml = $_[1];
         # handle xml
      });
   });
});

Example 2: say you're building a client for a protocol that replies either with "OK" plus another line, or "ERROR", for the first request you send — and then a flat 64 bytes for the second request. Because queues exist 🫶, you can just pipeline both sends and mess with the queue as needed inside the callbacks.

When the first callback fires and sees an "OK" response, it unshifts another line-read. That line-read gets queued before the 64-byte chunk callback. Iconic behavior, honestly.

# request one, returns either "OK + extra line" or "ERROR"
$handle->push_write ("request 1\015\012");

# we expect "ERROR" or "OK" as response, so push a line read
$handle->push_read (line => sub {
   # if we got an "OK", we have to _prepend_ another line,
   # so it will be read before the second request reads its 64 bytes
   # which are already in the queue when this callback is called
   # we don't do this in case we got an error
   if ($_[1] eq "OK") {
      $_[0]->unshift_read (line => sub {
         my $response = $_[1];
         ...
      });
   }
});

# request two, simply returns 64 octets
$handle->push_write ("request 2\015\012");

# simply read 64 bytes, always
$handle->push_read (chunk => 64, sub {
   my $response = $_[1];
   ...
});

---

• $handle->on_read ($cb) — swaps out the currently set on_read callback, or clears it entirely (when the new callback is undef). Peep the description of on_read in the constructor for the deets.

Heads up tho 🫣 — this method may invoke callbacks, which means the handle could be destroyed by the time it returns. No cap.

• $handle->rbuf — hands you the read buffer (as a modifiable lvalue). You can also go straight at it as the ->{rbuf} member if you want — that's much faster and no less clean, honestly, do it.

The only operation allowed on the read buffer (besides looking at it) is removing data from its beginning. Modifying it or appending to it otherwise? Not allowed. It will lead to hard-to-track-down bugs and you will be sad 😭 fr fr.

NOTE: the read buffer should only be used or modified inside the on_read callback, or when push_read/unshift_read are used with a single callback (i.e. untyped). The typed push_read and unshift_read methods manage the read buffer all on their own, like the independent queens they are 👑.

• $handle->push_read ($cb) / $handle->unshift_read ($cb) — append the given callback to the end of the queue (push_read) or prepend it to the front (unshift_read).

The callback fires every time some additional read data arrives.

It must check whether there's enough data in the read buffer already.

If there isn't enough, it has to return the empty list or a false value — in which case it'll get called again and again until enough data shows up (or an error condition gets detected). Patient king energy ⭐

If enough data was available, then the callback must remove all the data it cares about (which can legitimately be none at all) and return a true value. Once it returns true, it gets yeeted from the queue 💀

These methods may invoke callbacks — so yeah, the handle might be destroyed after they return. Same warning, same drama 🙈

• $handle->push_read (type => @args, $cb) / $handle->unshift_read (type => @args, $cb) — instead of writing a callback that parses everything itself, you can pick from a bunch of predefined parsing formats: chunks of data, lines, etc. You can also pass the (fully qualified) name of a package — in which case AnyEvent tries to load that package and then expects to find the anyevent_read_type function inside it (see "custom read types", below).

Predefined types are (and if you have ideas for additional types, feel free to drop by and tell us 💖):

• chunk => $octets, $cb->($handle, $data) — only invokes the callback once $octets bytes have been read, and passes the data read to the callback. The callback will never be called with less data. Ever. Periodt.

Example: read 2 bytes.

$handle->push_read (chunk => 2, sub {
   say "yay " . unpack "H*", $_[1];
});

• line => [$eol, ]$cb->($handle, $line, $eol) — the callback is only called once a full line (including the end of line marker, $eol) has been read. That line (excluding the end of line marker) gets passed to the callback as the second argument ($line), and the end of line marker goes in as the third argument ($eol).

The end of line marker, $eol, can be either a string — interpreted as a fixed record end marker — or a regex object (like one made with qr), in which case it's interpreted as a regular expression.

The $eol argument is optional. If it's missing (NOT undef), then qr|\015?\012| is used, which is good for most internet protocols. 🎀

Partial lines at the end of the stream will never be returned, because they're not marked by the end of line marker. Sad but fair.

• regex => $accept[, $reject[, $skip], $cb->($handle, $data) — does a regex match against the regex object $accept and returns everything up to and including the match. All the usual regex variables ($1, %+ etc.) from the match are available inside the callback. Chef's kiss 🧚

Example: read a single line terminated by '\n'.

$handle->push_read (regex => qr<\n>, sub { ... });

If $reject is given and not undef, it decides when data gets rejected: it's matched against the data whenever the $accept regex doesn't match, and it generates an "EBADMSG" error when it matches. Super useful for quickly rejecting wrong data — saves you from sitting around waiting for a timeout or a receive buffer overflow 🔥

Example: expect a single decimal number followed by whitespace, reject anything else (note the use of an anchor).

$handle->push_read (regex => qr<^[0-9]+\s>, qr<[^0-9]>, sub { ... });

If $skip is given and not undef, it gets matched against the receive buffer when neither $accept nor $reject match — and everything preceding and including that match is accepted unconditionally. Great for skipping huge amounts of data you already know can't match, so $accept and $reject don't have to start matching from the beginning. It's purely an optimisation, and it's usually only worth it when you expect more than a few kilobytes.

Example: expect an http header, which ends at "\015\012\015\012". Since we expect the header to be very large (it isn't in practice, but...), we use a skip regex to skip the initial portions. The skip regex is tricky in that it only accepts something not ending in either \015 or \012, since those are required for the accept regex.

$handle->push_read (regex =>
   qr<\015\012\015\012>,
   undef, # no reject
   qr<^.*[^\015\012]>,
   sub { ... });

• netstring => $cb->($handle, $string) — a netstring (http://cr.yp.to/proto/netstrings.txt, and this is not an endorsement 💀).

Throws an error with $! set to EBADMSG on format violations.

• packstring => $format, $cb->($handle, $string) — an octet string prefixed with an encoded length. The encoding $format uses the same format as a Perl pack format, but it must specify a single integer only — only one of "cCsSlLqQiInNvVjJw" is allowed, plus an optional "!", "<" or ">" modifier. Don't get cute with it 🌸

For example, DNS over TCP uses a prefix of "n" (2 octet network order), and EPP uses a prefix of "N" (4 octtes).

Example: read a block of data prefixed by its length in BER-encoded format (very efficient).

$handle->push_read (packstring => "w", sub {
   my ($handle, $data) = @_;
});

• json => $cb->($handle, $hash_or_arrayref) — reads a JSON object or array, decodes it, and hands it to the callback. When a parse error occurs, an "EBADMSG" error gets raised. 😤

If a "json" object was passed to the constructor, that one is used for the final decode. Otherwise it creates a JSON::XS or JSON::PP coder object expecting UTF-8.

This read type uses the incremental parser available with JSON version 2.09 (and JSON::XS version 2.2) and above. iykyk.

Since JSON texts are fully self-delimiting, the "json" read and write types are an ideal simple RPC protocol — just exchange JSON datagrams. See the "json" write type description, above, for an actual example. Obsessed, honestly 🦋

• cbor => $cb->($handle, $scalar) — reads a CBOR value, decodes it, and passes it to the callback. When a parse error occurs, an "EBADMSG" error gets raised.

If a CBOR::XS object was passed to the constructor, that one gets used for the final decode. Otherwise it creates a CBOR coder without enabling any options.

You have to provide a dependency to CBOR::XS on your own, bestie — this module will load the CBOR::XS module, but AnyEvent does not depend on it itself. So like, don't come for us when it's missing. 🙈

Since CBOR values are fully self-delimiting, the cbor read and write types are literally the ideal simple RPC protocol: just exchange CBOR datagrams, that's it, that's the whole vibe. See the cbor write type description above for an actual example. ✨

💾 storable => $cb->($handle, $ref)

Deserialises a Storable frozen representation, exactly as written by the storable write type (that's a BER-encoded length prefix followed by nfreeze'd data, iykyk).

Raises an EBADMSG error if the data could not be decoded. 💀

🕵️ tls_detect => $cb->($handle, $detect, $major, $minor)

This one checks the input stream for a valid SSL or TLS handshake TLSPaintext record without consuming anything — it just peeks, respectfully. 👀 Only SSL version 3 or higher is handled, up to the fictitious protocol 4.x (but both SSL3+ and SSL2-compatible framing is supported, so like, both eras are welcome here).

If it detects that the input data is likely TLS, it calls the callback with a true value for $detect and the (on-wire) TLS version as second and third argument ($major is 3, and $minor is 0..4 for SSL 3.0, TLS 1.0, 1.1, 1.2 and 1.3, respectively). If it detects the input to be definitely not TLS, it calls the callback with a false value for $detect.

So the callback could use this info to decide whether or not to start TLS negotiation. Main character decision-making energy. 👑

In all cases the data read so far is passed to the following read handlers.

Usually you want to use the tls_autostart read type instead — it's the upgrade, fr fr.

If you want to design a protocol that works in the presence of TLS detection, make sure that any non-TLS data doesn't start with the octet 22 (ASCII SYN, 16 hex) or 128-255 (i.e. highest bit set). The checks this read type does are a bit more strict, but might loosen in the future to accommodate protocol changes.

This read type does not rely on AnyEvent::TLS (and thus, not on Net::SSLeay). Independent queen behaviour. 💅

🚀 tls_autostart => [$tls_ctx, ]$tls

Tries to detect a valid SSL or TLS handshake. If one is detected, it tries to start TLS by calling starttls with the given arguments.

In practise, $tls must be "accept", or a Net::SSLeay context that has been configured to accept — because servers do not normally send a handshake on their own and thus cannot be detected this way. Servers are lowkey just sitting there waiting, smh.

See tls_detect above for more details.

Example — give the client a chance to start TLS before accepting a text line:

   $hdl->push_read (tls_autostart => "accept");
   $hdl->push_read (line => sub {
      print "received ", ($_[0]{tls} ? "encrypted" : "cleartext"), " <$_[1]>\n";
   });

🛠️ custom read types — Package::anyevent_read_type $handle, $cb, @args

Instead of one of the predefined types, you can also specify the name of a package. AnyEvent will try to load the package and then expects to find a function named anyevent_read_type inside. If it isn't found, it progressively tries to load the parent package until it either finds the function (good 💖) or runs out of packages (bad 💀).

Whenever this type is used, push_read will invoke the function with the handle object, the original callback and the remaining arguments.

The function is supposed to return a callback (usually a closure) that works as a plain read callback (see ->push_read ($cb)), so you can mentally treat the function as a "configurable read type to read callback" converter. Glow-up pipeline, honestly. ✨

It should invoke the original callback when it is done reading — and remember to pass $handle as first argument, as all other callbacks do that, although there is no strict requirement on this.

For examples, see the source of this module (perldoc -m AnyEvent::Handle, search for register_read_type). Happy hunting, bestie. 🔍

⏸️ $handle->stop_read / ▶️ $handle->start_read

In rare cases you actually do not want to read anything from the socket. In this case you can call stop_read. Neither on_read nor any queued callbacks will be executed then — full silence, no thoughts, head empty. To start reading again, call start_read.

Note that AnyEvent::Handle will automatically start_read for you when you change the on_read callback or push/unshift a read callback, and it will automatically stop_read for you when neither on_read is set nor there are any read requests in the queue. It's got you like that. 🫶

In older versions of this module (<= 5.3), these methods had no effect, as TLS does not support half-duplex connections. In current versions they work as expected, as this behaviour is required to avoid certain resource attacks, where the program would be forced to read (and buffer) arbitrary amounts of data before being able to send some data. The drawback is that some readings of the SSL/TLS specifications basically require this attack to be working, as SSL/TLS implementations might stall sending data during a rehandshake.

As a guideline: during the initial handshake, you should not stop reading, and as a client, it might cause problems, depending on your application. So like, tread carefully. ⚠️

🔐 $handle->starttls ($tls[, $tls_ctx])

Instead of starting TLS negotiation immediately when the AnyEvent::Handle object is created, you can also do that at a later time by calling starttls. See the tls constructor argument for general info.

Starting TLS is currently an asynchronous operation — when you push some write data and then call ->starttls, then TLS negotiation will start immediately, after which the queued write data is then sent. This might change in future versions, so best make sure you have no outstanding write data when calling this method. (Not me warning you twice. 😭)

The first argument is the same as the tls constructor argument (either "connect", "accept" or an existing Net::SSLeay object).

The second argument is the optional AnyEvent::TLS object that is used when AnyEvent::Handle has to create its own TLS connection object, or a hash reference with key => value pairs that will be used to construct a new context.

The TLS connection object will end up in $handle->{tls}, the TLS context in $handle->{tls_ctx} after this call and can be used or changed to your liking. Note that the handshake might have already started when this function returns. Sneaky like that. 🧚

Due to bugs in OpenSSL, it might or might not be possible to do multiple handshakes on the same stream. It is best to not attempt to use the stream after stopping TLS.

This method may invoke callbacks (and therefore the handle might be destroyed after it returns). 💀

🔒 $handle->stoptls

Shuts down the SSL connection — this makes a proper EOF handshake by sending a close notify to the other side, but since OpenSSL doesn't support non-blocking shut downs, it is not guaranteed that you can re-use the stream afterwards. So like, no promises. 🌸

This method may invoke callbacks (and therefore the handle might be destroyed after it returns).

🧹 $handle->resettls

This rarely-used method simply resets any TLS state on the handle, usually causing data loss.

One case where it may be useful is when you want to skip over the data in the stream but you are not interested in interpreting it, so data loss is no concern. Trash it, we don't care. 🗑️

💀 $handle->destroy

Shuts down the handle object as much as possible — this call ensures that no further callbacks will be invoked and as many resources as possible will be freed. Any method you will call on the handle object after destroying it in this way will be silently ignored (and it will return the empty list). Ghosted. 👻

Normally, you can just "forget" any references to an AnyEvent::Handle object and it will simply shut down. This works in fatal error and EOF callbacks, as well as code outside. It does NOT work in a read or write callback, so when you want to destroy the AnyEvent::Handle object from within such a callback. You MUST call ->destroy explicitly in that case. No cap, this is the one you'll forget and then debug for an hour. 😭

Destroying the handle object in this way has the advantage that callbacks will be removed as well, so if those are the only reference holders (as is common), then one doesn't need to do anything special to break any reference cycles. Clean. ✨

The handle might still linger in the background and write out remaining data, as specified by the linger option, however.

👻 $handle->destroyed

Returns false as long as the handle hasn't been destroyed by a call to ->destroy, true otherwise.

Can be useful to decide whether the handle is still valid after some callback possibly destroyed the handle. For example, ->push_write, ->starttls and other methods can call user callbacks, which in turn can destroy the handle, so work can be avoided by checking sometimes:

   $hdl->starttls ("accept");
   return if $hdl->destroyed;
   $hdl->push_write (...

Note that the call to push_write will silently be ignored if the handle has been destroyed, so often you can just ignore the possibility of the handle being destroyed. Lowkey convenient. 💅

🎀 AnyEvent::Handle::TLS_CTX

This function creates and returns the AnyEvent::TLS object used by default for TLS mode.

The context is created by calling AnyEvent::TLS without any arguments. Simple as that, bestie. 🦋

💅 NONFREQUENTLY ASKED QUESTIONS 💅

🫶 I literally undef the AnyEvent::Handle reference inside my callback and it STILL keeps calling me?!

Bestie, that’s because AnyEvent::Handle keeps a reference to itself while it’s handling read or write callbacks. It’s not being clingy for no reason — it’s built like that.

It is only safe to “forget” the reference inside EOF or error callbacks. From within all other callbacks, you need to explicitly call the ->destroy method. No cap.

---

😭 Why is my on_eof callback never called?

Probably because your on_error callback is being called instead. When you have outstanding requests in your read queue, then an EOF is considered an error — because you clearly expected some data. The vibe check failed.

To avoid this, make sure you have an empty read queue whenever your handle is supposed to be “idle” (i.e. connection closes are O.K.). You can set an on_read handler that simply pushes the first read requests in the queue.

See also the next question, which explains this in a bit more detail. 🫶

---

🔁 How can I serve requests in a loop?

Most protocols have some setup phase (authentication, for example) followed by a request handling phase, where the server waits for requests and handles them, in a loop.

There are two important variants. The first one is the traditional, better variant: it handles requests until the server gets some QUIT command, causing it to close the connection first (highly desirable for a busy TCP server). A client dropping the connection is an error, which means this variant can detect an unexpected connection close.

To handle this case, always make sure you have a non-empty read queue, by pushing the “read request start” handler on it:

# we assume a request starts with a single line
my @start_request; @start_request = (line => sub {
   my ($hdl, $line) = @_;

   ... handle request

   # push next request read, possibly from a nested callback
   $hdl->push_read (@start_request);
});

# auth done, now go into request handling loop
# now push the first @start_request
$hdl->push_read (@start_request);

By always having an outstanding push_read, the handle always expects some data and raises the EPIPE error when the connection is dropped unexpectedly.

The second variant is a protocol where the client can drop the connection at any time. For TCP, this means that the server machine may run out of sockets easier, and in general, it means you cannot distinguish a protocol failure/client crash from a normal connection close. Nevertheless, these kinds of protocols are common (and sometimes even the best solution to the problem).

Having an outstanding read request at all times is possible if you ignore EPIPE errors, but this doesn’t help with when the client drops the connection during a request, which would still be an error.

A better solution is to push the initial request read in an on_read callback. This avoids an error, as when the server doesn’t expect data (i.e. is idly waiting for the next request), an EOF will not raise an error, but simply result in an on_eof callback. It is also a bit slower and simpler:

# auth done, now go into request handling loop
$hdl->on_read (sub {
   my ($hdl) = @_;

   # called each time we receive data but the read queue is empty
   # simply start read the request

   $hdl->push_read (line => sub {
      my ($hdl, $line) = @_;

      ... handle request

      # do nothing special when the request has been handled, just
      # let the request queue go empty.
   });
});

---

🦋 I get different callback invocations in TLS mode / Why can’t I pause reading?

Unlike, say, TCP, TLS connections do not consist of two independent communication channels, one for each direction. Or put differently, the read and write directions are not independent of each other: you cannot write data unless you are also prepared to read, and vice versa.

This means that, in TLS mode, you might get on_error or on_eof callback invocations when you are not expecting any read data — the reason is that AnyEvent::Handle always reads in TLS mode.

During the connection, you have to make sure that you always have a non-empty read-queue, or an on_read watcher. At the end of the connection (or when you no longer want to use it) you can call the destroy method.

---

🎀 How do I read data until the other side closes the connection?

If you just want to read your data into a perl scalar, the easiest way to achieve this is by setting an on_read callback that does nothing, clearing the on_eof callback and in the on_error callback, the data will be in $_[0]{rbuf}:

$handle->on_read (sub { });
$handle->on_eof (undef);
$handle->on_error (sub {
   my $data = delete $_[0]{rbuf};
});

Note that this example removes the rbuf member from the handle object, which is not normally allowed by the API. It is expressly permitted in this case only, as the handle object needs to be destroyed afterwards.

The reason to use on_error is that TCP connections, due to latencies and packet loss, might get closed quite violently with an error, when in fact all data has been received.

It is usually better to use acknowledgements when transferring data, to make sure the other side hasn’t just died and you got the data intact. This is also one reason why so many internet protocols have an explicit QUIT command. Iconic, honestly.

---

💖 I don’t want to destroy the handle too early — how do I wait until all data has been written?

After writing your last bits of data, set the on_drain callback and destroy the handle in there — with the default setting of low_water_mark this will be called precisely when all data has been written to the socket:

$handle->push_write (...);
$handle->on_drain (sub {
   AE::log debug => "All data submitted to the kernel.";
   undef $handle;
});

If you just want to queue some data and then signal EOF to the other side, consider using ->push_shutdown instead.

---

🌈 I want to contact a TLS/SSL server, I don’t care about security

If your TLS server is a pure TLS server (e.g. HTTPS) that only speaks TLS, connect to it and then create the AnyEvent::Handle with the tls parameter:

tcp_connect $host, $port, sub {
   my ($fh) = @_;

   my $handle = new AnyEvent::Handle
      fh  => $fh,
      tls => "connect",
      on_error => sub { ... };

   $handle->push_write (...);
};

---

👑 I want to contact a TLS/SSL server, I DO care about security

Then you should additionally enable certificate verification, including peername verification, if the protocol you use supports it (see AnyEvent::TLS, "verify_peername").

E.g. for HTTPS:

tcp_connect $host, $port, sub {
   my ($fh) = @_;

    my $handle = new AnyEvent::Handle
       fh       => $fh,
       peername => $host,
       tls      => "connect",
       tls_ctx  => { verify => 1, verify_peername => "https" },
       ...

Note that you must specify the hostname you connected to (or whatever peername the protocol needs) as the peername argument, otherwise no peername verification will be done.

The above will use the system-dependent default set of trusted CA certificates. If you want to check against a specific CA, add the ca_file (or ca_cert) arguments to tls_ctx:

     tls_ctx  => {
        verify          => 1,
        verify_peername => "https",
        ca_file         => "my-ca-cert.pem",
     },

---

🌸 I want to create a TLS/SSL server, how do I do that?

Well, you first need to get a server certificate and key. You have three options:

• a) ask a CA (buy one, use cacert.org etc.)
• b) create a self-signed certificate (cheap — check the search engine of your choice, there are many tutorials on the net)
• c) make your own CA (tinyca2 is a nice program for that purpose)

Then create a file with your private key (in PEM format, see AnyEvent::TLS), followed by the certificate (also in PEM format). The file should then look like this:

-----BEGIN RSA PRIVATE KEY-----
...header data
... lots of base64'y-stuff
-----END RSA PRIVATE KEY-----

-----BEGIN CERTIFICATE-----
... lots of base64'y-stuff
-----END CERTIFICATE-----

The important bits are the PRIVATE KEY and CERTIFICATE parts. Then specify this file as cert_file:

tcp_server undef, $port, sub {
   my ($fh) = @_;

   my $handle = new AnyEvent::Handle
      fh       => $fh,
      tls      => "accept",
      tls_ctx  => { cert_file => "my-server-keycert.pem" },
      ...

When you have intermediate CA certificates that your clients might not know about, just append them to the cert_file.

---

🧚 SUBCLASSING AnyEvent::Handle

In many cases, you might want to subclass AnyEvent::Handle.

To make this easier, a given version of AnyEvent::Handle uses these conventions:

• all constructor arguments become object members. At least initially, when you pass a tls-argument to the constructor it will end up in $handle->{tls}. Those members might be changed or mutated later on (for example tls will hold the TLS connection object).
• other object member names are prefixed with an _. All object members not explicitly documented (internal use) are prefixed with an underscore character, so the remaining non-_-namespace is free for use for subclasses.
• all members not documented here and not prefixed with an underscore are free to use in subclasses. Of course, new versions of AnyEvent::Handle may introduce more public member variables, but that’s just life. At least it is documented. 💅

---

🫶 AUTHOR

Robin Redeker "<elmex at ta-sa.org>", Marc Lehmann <schmorp@schmorp.de>.

---

That’s the whole page, bestie! Source/version/date check: perl v5.42.0, dated 2019-09-18, and this is AnyEvent::Handle(3). Bye! 💖✨