otplike.process
This namespace implements core process concepts such as spawning, linking, monitoring, message passing and exiting.
Process context
All the calls made from process function directly or indirectly after it has been spawned happen in the context of the process (i.e., are issued by the process).
Process exit
A process exits when:
- it receives exit signal with reason
:kill
, - it receives exit signal with the reason other than
:kill
, and it doesn’t trap exits, - its initial function ends (returning a value or with exception).
As there is no way to force process function to stop execution after its process has exited, a process can be alive or exiting:
- a process is alive until it exits for any reason,
- a process becomes exiting after it exited until it’s initial function returns.
There can be cases when exiting process tries to communicate with other processes. In such cases exception with the reason :noproc
is thrown.
The following happens when a process exits:
- its mailbox becomes closed so that no future messages can be received,
- all linked/monitoring processes receive exit/down signal,
- it can not be reached using its pid,
- it is no longer registered.
Signals (control messages)
Signals are used internally to manage processes. Exiting, monitoring, linking and some other operations require sending signals (not messages) to involved processes.
!
(! dest message)
Sends a message
to dest
. dest
can be a process identifier, or a registered name.
Returns true
if message
was sent (dest
process existed), false otherwise.
Throws if any of arguments is nil
.
alive?
(alive?)
(alive? pid)
Returns true
if the process exists and is alive, that is, is not exiting and has not exited. Otherwise returns false
.
When called without arguments, returns information about the calling process.
async
macro
(async & body)
Executes body asynchronously. Like go-block but propagates exceptions.
The returned value is to be passed to await!
.
async?
(async? x)
Returns true
if x
is async value (i.e. is returned by async
), otherwise returns false
.
await!
macro
(await! x)
Returns the value of the async operation represented by x
or exits with the same reason the operation exited. Parks until the operation is completed if required.
It is illegal to pass the same async value to await!
more than once.
Throws if x
is not async value (i.e. is not returned by async
).
await?!
macro
(await?! x)
If x
is returned by async
, returns the value of the corresponding async operation (parks if needed). If x
is a regular value, returns x
.
demonitor
(demonitor mref)
(demonitor mref {flush? :flush})
If mref
is a reference that the calling process obtained by calling monitor, this monitoring is turned off. If the monitoring is already turned off, nothing happens. If mref
is created by other process, nothing happens.
Once demonitor has returned, it is guaranteed that no [:DOWN monitor-ref _ _ _]
message, because of the monitor, will be placed in the caller message queue in the future. A [:DOWN monitor-ref _ _ _]
message can have been placed in the caller message queue before the call, though. It is therefore usually advisable to remove such a :DOWN
message from the message queue after monitoring has been stopped. (demonitor mref {:flush true})
can be used instead of (demonitor mref)
if this cleanup is wanted.
When :flush
option is true
, removes (one) :DOWN
message, if there is one, from the caller message queue after monitoring has been stopped. This is equivalent to the following:
(demonitor mref)
(selective-receive!
[_ mref _ _ _] true
(after 0
true))
Returns true
.
Throws when called not in process context, or calling process is not alive, or mref
is not a ref.
ex-catch
macro
(ex-catch expr)
Executes expr
. Returns either result of execution or [:EXIT reason]
.
exit
(exit reason)
(exit pid reason)
When called with one argument (reason)
Throws special exception (which can be caught). When the exception leaves process’ initial function, it causes the process to exit with the specified reason.
When called with two arguments (pid and reason)
Sends an exit signal with the reason reason
to the process identified by pid
.
If reason is any term, except :normal
or :kill
:
- if
pid
is not trapping exits,pid
itself exits with exit reason. - if
pid
is trapping exits, the exit signal is transformed into a message[:EXIT from reason]
and delivered to the message queue ofpid
.from
is the process identifier of the process that sent the exit signal.
If reason is :normal
, pid
does not exit. If pid
is trapping exits, the exit signal is transformed into a message [:EXIT from :normal]
and delivered to its message queue.
If reason is :kill
, an untrappable exit signal is sent to pid, which unconditionally exits with reason :killed
.
Notice that process can exit with other reason before exit signal is processed.
Returns true
if exit signal was sent (dest
process existed), false
otherwise.
Throws when called not in process context, if calling process is not alive, if pid
is not a pid, or reason is nil
.
flag
(flag flag value)
Sets the value of a process’ flag. See description of each flag below.
Flags:
:trap-exit
. When set totrue
, exit signals arriving to a process are converted to[:EXIT from reason]
messages, which can be received as ordinary messages. If is set tofalse
, the process exits if it receives an exit signal other than:normal
and the exit signal is propagated to its linked processes.
Returns the old value of a flag
.
Throws when called not in process context.
link
(link pid)
Creates a link between the calling process and another process identified by pid
, if there is not such a link already. If a process attempts to create a link to itself, nothing is done.
If pid does not exist and the calling process
- is trapping exits - the calling process receives message
[:EXIT pid :noproc]
. - is not trapping exits - process exits with reason
:noproc
.
Returns true
.
Throws when called not in process context, or calling process is not alive, or pid
is not a pid.
map-async
(map-async f async-val)
Creates a copy of async-val
adding f
to the list of its transformation functions.
monitor
(monitor pid-or-name)
Sends a monitor request to the entity identified by pid-or-name
. If the monitored entity does not exist or when it dies, the caller of monitor will be notified by a message of the following format:
[tag monitor-ref type object info]
type
can be one of the following keywords: :process
.
A monitor is triggered only once, after that it is removed from both monitoring process and the monitored entity. Monitors are fired when the monitored process terminates, or does not exist at the moment of creation. The monitoring is also turned off when demonitor
is called.
When monitoring by name please note, that the registered-name is resolved to pid only once at the moment of monitor instantiation, later changes to the name registration will not affect the existing monitor.
When a monitor is triggered, a :DOWN
message that has the following pattern
[:DOWN monitor-ref type object info]
is sent to the monitoring process.
In monitor message monitor-ref
and type
are the same as described earlier, and:
object
- the monitored entity, which triggered the event. That is the argument of monitor call.info
- either the exit reason of the process, or:noproc
(process did not exist at the time of monitor creation).
Making several calls to monitor
for the same pid-or-name
is not an error; it results in as many independent monitoring instances.
Monitoring self does nothing.
Returns monitor-ref
.
Throws when called not in process context, or if calling process is not alive.
pid->str
(pid->str pid)
Returns a string corresponding to the text representation of pid
.
Throws if pid
is not a process identifier.
Warning: this function is intended for debugging and is not to be used in application programs.
proc-defn
macro
(proc-defn fname doc-string? args & body)
The same as (def fname (proc-fn args body))
.
proc-defn-
macro
(proc-defn- fname args & body)
The same as proc-defn, but defines a private var.
proc-fn
macro
(proc-fn name-or-args & args-body)
Creates process function which can be passed to spawn
.
process-info
(process-info pid)
(process-info pid item-or-list)
(process-info pid)
Returns a map containing information about the process identified by pid
, or nil
if the process is not alive.
All items are not mandatory. The set of info-tuples being part of the result can be changed without prior notice.
The following info-tuples are part of the result: :initial-call
, :status
, :message-queue-len
, :links
, :flags
.
If the process identified by pid
has a registered name, also an info-tuple for :registered-name
is included.
Warning! This function is intended for debugging only. For all other purposes, use
(process-info pid item-or-list)
.
Throws if pid
is not a pid.
(process-info pid item-or-list)
Returns information about the process identified by pid
, as specified by into-key or info-key list. Returns nil
if the process is not alive.
If the process is alive and a single info-key is specified, the returned value is the corresponding info-tuple.
(process-info pid :messages)
=> [:messages [:msg1 [:msg2] {:msg3 3}]]
If a list of info-keys is specified, the result is a list of info-tuples. The info-tuples in the list are included in the same order as the keys were included in info-key list. Valid items can be included multiple times in item-key list.
Info-tuples:
[:initial-call [fn-symbol arity]]
fn-symbol
, arity
is the initial function call with which the process was spawned.
[:links pids]
pids
is a list of process identifiers, with processes to which the process has a link.
[:message-queue-len message-queue-len]
message-queue-len
is the number of messages currently in the message queue of the process. This is the length of the list message-queue
returned as the information item messages (see below).
[:messages message-queue]
message-queue
is a list of the messages to the process,which have not yet been processed.
[:monitored-by pids]
A list of process identifiers monitoring the process.
[:monitors monitors]
A list of monitors that are active for the process. The list consists of pids and registered names.
[:registered-name reg-name]
reg-name
is the registered process name or nil
if the process has no registered name.
[:status status]
status
is the status of the process and is one of the following:
:exiting
:waiting
(for a message):running
[:trace trace-flags]
A map of trace flags set for the process. This info-tuple is not available now but it is reserved for future.
[:flags flags]
A map of flags set for the process (e.g., {:trap-exit true}
).
Throws if pid
is not a pid, or specified info-key doesn’t exist.
processes
(processes)
Returns a sequence of process identifiers corresponding to all the processes currently existing.
Notice that an exiting process exists, but is not alive. That is, (alive? pid)
returns false
for an exiting process, but its process identifier is part of the result returned from (processes)
.
receive!
macro
(receive! & clauses)
Receives and removes from the inbox the first message sent to the process using the !
function:
(receive!
pattern1 pattern-expr1
pattern2 pattern-expr2
...)
The message is matched using clojure.core.match/match
against the patterns. If a match succeeds, the corresponding expression is evaluated, otherwise throws. It is illegal to use a receive!
with no patterns.
If there are no messages in the inbox, the execution is suspended, possibly indefinitely, until the first message arrives.
The receive expression can be augmented with a timeout:
(receive!
pattern pattern-expr
...
(after timeout
timeout-expr))
There are two special cases for the timeout
value: :infinity
- the process is to wait indefinitely for a matching message. This is the same as not using a timeout. This can be useful for timeout values that are calculated at runtime. 0
- if there is no messages in the mailbox, or the first message doesn’t match, the timeout occurs immediately.
Returns the value of the evaluated expression.
resolve-pid
(resolve-pid pid-or-name)
If pid-or-name
is a pid - returns pid. If a registered name - returns the pid of registered process. Else returns nil
.
selective-receive!
macro
(selective-receive! & clauses)
The same as receive!
but doesn’t crash if the first message doesn’t match. Instead waits for the matching message, removes it from the mailbox leaving all the rest messages in the original order.
When the timeout is 0
, checks all the messages in the mailbox and not the first one only.
self
(self)
Returns the process identifier of the calling process. Throws when called not in process context, or process is not alive.
spawn
(spawn proc-func)
(spawn proc-func args)
Returns the process identifier of a new process started by the application of proc-fun
to args
.
spawn-link
(spawn-link proc-func)
(spawn-link proc-func args)
Returns the process identifier of a new process started by the application of proc-fun
to args
. A link is created between the calling process and the new process, atomically. Otherwise works like spawn
.
Throws when called not in process context, or calling process is not alive.
spawn-opt
(spawn-opt proc-func opts)
(spawn-opt proc-func args opts)
Returns the process identifier of a new process started by the application of proc-fun
to args
.
options
argument is a map of option names (keywords) to their values.
The following options are allowed:
:flags
- a map of process’ flags (e.g.{:trap-exit true}
):link
- iftrue
, sets a link to the parent process:register
- name to register the process, can not be pid, if name isnil
process will not be registered
Throws
- when there is another process registered under the same name,
- on invalid arguments.
unlink
(unlink pid)
Removes the link, if there is one, between the calling process and the process referred to by pid
.
Returns true
.
Does not fail if there is no link to pid
, if pid
is self pid, or if pid
does not exist.
Once unlink
has returned, it is guaranteed that the link between the caller and the entity referred to by pid
has no effect on the caller in the future (unless the link is setup again).
If the caller is trapping exits, an [:EXIT pid _]
message from the link can have been placed in the caller’s message queue before the call.
Notice that the [:EXIT pid _]
message can be the result of the link, but can also be the result of pid calling exit. Therefore, it can be appropriate to clean up the message queue when trapping exits after the call to unlink.
Throws when called not in process context, or calling process is not alive, or pid
is not a pid.
whereis
(whereis reg-name)
Returns the process identifier with the registered name reg-name
, or nil
if the name is not registered.
with-async
macro
(with-async [binding-form async-expr :as bindings] & body)
Wraps body
into a function with binding-form as its single argument.
Returns (map-async body-fn async-expr-result)
.
Works as a recursion point for body
.