pub struct Cx<'a, A: 'static> { /* private fields */ }
Expand description
Implementations§
source§impl<'a, A> Cx<'a, A>
impl<'a, A> Cx<'a, A>
sourcepub fn this(&self) -> &Actor<A>
pub fn this(&self) -> &Actor<A>
Borrow the current actor reference temporarily. If you need a
longer-lived reference to the actor, then use
cx.this().clone()
.
sourcepub fn id(&self) -> LogID
pub fn id(&self) -> LogID
Get the logging-ID of the current actor. If the logger feature isn’t enabled, returns 0.
sourcepub fn stop(&mut self)
pub fn stop(&mut self)
Indicate successful termination of the actor. As soon as the
currently-running actor call finishes, the actor will be
terminated. Actor state will be dropped, and any further
calls to this actor will be discarded. The termination status
is passed back to the StopCause
handler provided when the
actor was created. See also the stop!
macro.
Examples found in repository?
69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86
pub fn switch(&mut self, cx: CX![], on: bool) {
// Change the light state
call!([self.light], set(on));
self.count -= 1;
if self.count != 0 {
// Call switch again after a delay
after!(self.interval, [cx], switch(!on));
} else {
// Terminate the actor successfully, causing StopCause handler to run
cx.stop();
}
// Query the light state, receiving the response in the method
// `recv_state`, which has both fixed and forwarded arguments.
let ret = ret_some_to!([cx], recv_state(self.count) as (bool));
call!([self.light], query(ret));
}
sourcepub fn fail(&mut self, e: impl Error + 'static)
pub fn fail(&mut self, e: impl Error + 'static)
Indicate failure of the actor. As soon as the
currently-running actor call finishes, the actor will be
terminated. Actor state will be dropped, and any further
calls to this actor will be discarded. The termination status
is passed back to the StopCause
handler provided when the
actor was created. See also the fail!
macro.
sourcepub fn fail_str(&mut self, e: &'static str)
pub fn fail_str(&mut self, e: &'static str)
Indicate failure of the actor. As soon as the
currently-running actor call finishes, the actor will be
terminated. Actor state will be dropped, and any further
calls to this actor will be discarded. The termination status
is passed back to the StopCause
handler provided when the
actor was created. See also the fail!
macro.
sourcepub fn fail_string(&mut self, e: impl Into<String>)
pub fn fail_string(&mut self, e: impl Into<String>)
Indicate failure of the actor. As soon as the
currently-running actor call finishes, the actor will be
terminated. Actor state will be dropped, and any further
calls to this actor will be discarded. The termination status
is passed back to the StopCause
handler provided when the
actor was created. See also the fail!
macro.
sourcepub fn access_actor(&self) -> &Actor<A>
pub fn access_actor(&self) -> &Actor<A>
Used in macros to get an Actor
reference
sourcepub fn access_log_id(&self) -> LogID
pub fn access_log_id(&self) -> LogID
Used in macros to get the actor’s logging-ID. If the logger feature isn’t enabled, returns 0.
Methods from Deref<Target = Core>§
sourcepub fn now(&self) -> Instant
pub fn now(&self) -> Instant
Our view of the current time. Actors should use this in
preference to Instant::now()
for speed and in order to work
in virtual time.
Examples found in repository?
24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42
pub fn init(cx: CX![]) -> Option<Self> {
// Use cx.now() instead of Instant::now() to allow execution
// in virtual time if supported by the environment.
let start = cx.now();
Some(Self { start, on: false })
}
// Methods that may be called once the actor is "Ready" have a
// `&mut self` or `&self` first argument.
pub fn set(&mut self, cx: CX![], on: bool) {
self.on = on;
let time = cx.now() - self.start;
println!(
"{:04}.{:03} Light on: {}",
time.as_secs(),
time.subsec_millis(),
on
);
}
sourcepub fn systime(&self) -> SystemTime
pub fn systime(&self) -> SystemTime
Get the current SystemTime
. Normally this returns the same
as SystemTime::now()
, but if running in virtual time, it
would return the virtual SystemTime
instead (as provided to
Stakker::set_systime
by the virtual time main loop). Note
that this time is not suitable for timing things, as it may go
backwards if the user or a system process adjusts the clock.
It is just useful for showing or recording “human time” for
the user, and for recording times that are meaningful on a
longer scale, e.g. from one run of a process to the next.
sourcepub fn start_instant(&self) -> Instant
pub fn start_instant(&self) -> Instant
Get the Instant
which was passed to Stakker::new
when this
runtime was started.
sourcepub fn defer(&mut self, f: impl FnOnce(&mut Stakker) + 'static)
pub fn defer(&mut self, f: impl FnOnce(&mut Stakker) + 'static)
Defer an operation to be executed later. It is put on the
main queue, and run as soon all operations preceding it have
been executed. See also the call!
macro.
sourcepub fn lazy(&mut self, f: impl FnOnce(&mut Stakker) + 'static)
pub fn lazy(&mut self, f: impl FnOnce(&mut Stakker) + 'static)
Defer an operation to executed soon, but lazily. It goes onto
a lower priority queue executed once the normal defer queue
has been completely cleared (including any further deferred
items added whilst clearing that queue). This can be used for
flushing data generated in this batch of processing, for
example. See also the lazy!
macro.
sourcepub fn idle(&mut self, f: impl FnOnce(&mut Stakker) + 'static)
pub fn idle(&mut self, f: impl FnOnce(&mut Stakker) + 'static)
Defer an operation to be executed when this process next
becomes idle, i.e. when all other queues are empty and there
is no I/O to process. This can be used to implement
backpressure on incoming streams, i.e. only fetch more data
once there is nothing else left to do. See also the idle!
macro.
sourcepub fn after(
&mut self,
dur: Duration,
f: impl FnOnce(&mut Stakker) + 'static
) -> FixedTimerKey
pub fn after( &mut self, dur: Duration, f: impl FnOnce(&mut Stakker) + 'static ) -> FixedTimerKey
Delay an operation to be executed after a duration has passed.
This is the same as adding it as a fixed timer. Returns a key
that can be used to delete the timer. See also the after!
macro.
sourcepub fn timer_add(
&mut self,
expiry: Instant,
f: impl FnOnce(&mut Stakker) + 'static
) -> FixedTimerKey
pub fn timer_add( &mut self, expiry: Instant, f: impl FnOnce(&mut Stakker) + 'static ) -> FixedTimerKey
Add a fixed timer that expires at the given time. Returns a
key that can be used to delete the timer. See also the
at!
macro.
sourcepub fn timer_del(&mut self, key: FixedTimerKey) -> bool
pub fn timer_del(&mut self, key: FixedTimerKey) -> bool
Delete a fixed timer. Returns true
on success, false
if
timer no longer exists (i.e. it expired or was deleted)
sourcepub fn timer_max_add(
&mut self,
expiry: Instant,
f: impl FnOnce(&mut Stakker) + 'static
) -> MaxTimerKey
pub fn timer_max_add( &mut self, expiry: Instant, f: impl FnOnce(&mut Stakker) + 'static ) -> MaxTimerKey
Add a “Max” timer, which expires at the greatest (latest)
expiry time provided. See MaxTimerKey
for the
characteristics of this timer. Returns a key that can be used
to delete or modify the timer.
See also the timer_max!
macro, which may be more
convenient as it combines Core::timer_max_add
and
Core::timer_max_upd
.
sourcepub fn timer_max_upd(&mut self, key: MaxTimerKey, expiry: Instant) -> bool
pub fn timer_max_upd(&mut self, key: MaxTimerKey, expiry: Instant) -> bool
Update a “Max” timer with a new expiry time. It will be used as the new expiry time only if it is greater than the current expiry time. This call is designed to be very cheap to call frequently.
Returns true
on success, false
if timer no longer exists
(i.e. it expired or was deleted)
See also the timer_max!
macro, which may be more
convenient as it combines Core::timer_max_add
and
Core::timer_max_upd
.
sourcepub fn timer_max_del(&mut self, key: MaxTimerKey) -> bool
pub fn timer_max_del(&mut self, key: MaxTimerKey) -> bool
Delete a “Max” timer. Returns true
on success, false
if
timer no longer exists (i.e. it expired or was deleted)
sourcepub fn timer_max_active(&mut self, key: MaxTimerKey) -> bool
pub fn timer_max_active(&mut self, key: MaxTimerKey) -> bool
Check whether a “Max” timer is active. Returns true
if it
exists and is active, false
if it expired or was deleted or
never existed
sourcepub fn timer_min_add(
&mut self,
expiry: Instant,
f: impl FnOnce(&mut Stakker) + 'static
) -> MinTimerKey
pub fn timer_min_add( &mut self, expiry: Instant, f: impl FnOnce(&mut Stakker) + 'static ) -> MinTimerKey
Add a “Min” timer, which expires at the smallest (earliest)
expiry time provided. See MinTimerKey
for the
characteristics of this timer. Returns a key that can be used
to delete or modify the timer.
See also the timer_min!
macro, which may be more
convenient as it combines Core::timer_min_add
and
Core::timer_min_upd
.
sourcepub fn timer_min_upd(&mut self, key: MinTimerKey, expiry: Instant) -> bool
pub fn timer_min_upd(&mut self, key: MinTimerKey, expiry: Instant) -> bool
Update a “Min” timer with a new expiry time. It will be used as the new expiry time only if it is earlier than the current expiry time. This call is designed to be very cheap to call frequently, so long as the change is within the wiggle-room allowed. Otherwise it causes the working timer to be deleted and added again, readjusting the wiggle-room accordingly.
Returns true
on success, false
if timer no longer exists
(i.e. it expired or was deleted)
See also the timer_min!
macro, which may be more
convenient as it combines Core::timer_min_add
and
Core::timer_min_upd
.
sourcepub fn timer_min_del(&mut self, key: MinTimerKey) -> bool
pub fn timer_min_del(&mut self, key: MinTimerKey) -> bool
Delete a “Min” timer. Returns true
on success, false
if
timer no longer exists (i.e. it expired or was deleted)
sourcepub fn timer_min_active(&mut self, key: MinTimerKey) -> bool
pub fn timer_min_active(&mut self, key: MinTimerKey) -> bool
Check whether a “Min” timer is active. Returns true
if it
exists and is active, false
if it expired or was deleted or
never existed
sourcepub fn anymap_set<T: Clone + 'static>(&mut self, val: T)
pub fn anymap_set<T: Clone + 'static>(&mut self, val: T)
Put a value into the anymap
. This can be accessed using the
Core::anymap_get
or Core::anymap_try_get
call. An
anymap can store one value for each Rust type. The value must
implement Clone
, i.e. it must act something like an Rc
or
else be copyable data.
This is intended to be used for storing certain global
instances which actors may need to get hold of, for example an
access-point for the I/O poll implementation that Stakker is
running under. In other words the anymap
is intended to
represent the environment.
There’s nothing I can do to stop you using this like an
inefficient global variable store, but doing that would be a
terrible idea. Using the anymap
that way breaks the actor
model and makes your code harder to reason about. Really it
would be cleaner to use a Share
if you need to break the
actor model and share data, because at least then the
interconnection between actors would be explicit, and trying
to move an interconnected actor to a remote machine would fail
immediately.
sourcepub fn anymap_get<T: Clone + 'static>(&mut self) -> T
pub fn anymap_get<T: Clone + 'static>(&mut self) -> T
Gets a clone of a value from the Stakker anymap
. This is
intended to be used to access certain global instances, for
example the I/O poll implementation that this Stakker is
running inside. Panics if the value is not found.
sourcepub fn anymap_try_get<T: Clone + 'static>(&mut self) -> Option<T>
pub fn anymap_try_get<T: Clone + 'static>(&mut self) -> Option<T>
Tries to get a clone of a value from the Stakker anymap
.
This is intended to be used to access certain global
instances, for example the I/O poll implementation that this
Stakker is running inside. Returns None
if the value is
missing.
sourcepub fn shutdown(&mut self, cause: StopCause)
pub fn shutdown(&mut self, cause: StopCause)
Request that the event loop terminate. For this to work, the
event loop must check Core::not_shutdown
each time through
the loop. See also the ret_shutdown!
macro which can be
used as the StopCause
handler for an actor, to shut down
the event loop when that actor terminates. The event loop
code can obtain the StopCause
using
Core::shutdown_reason
.
sourcepub fn not_shutdown(&self) -> bool
pub fn not_shutdown(&self) -> bool
Should the event loop continue running? Returns true
if
there is no active shutdown in progress.
Examples found in repository?
93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127
fn main() {
// Contains all the queues and timers, and controls access to the
// state of all the actors.
let mut stakker0 = Stakker::new(Instant::now());
let stakker = &mut stakker0;
// Create and initialise the Light and Flasher actors. The
// Flasher actor is given a reference to the Light. Use a
// StopCause handler to shutdown when the Flasher terminates.
let light = actor!(stakker, Light::init(), ret_nop!());
let _flasher = actor!(
stakker,
Flasher::init(light.clone(), Duration::from_secs(1), 6),
ret_shutdown!(stakker)
);
// Since we're not in virtual time, we use `Instant::now()` in
// this loop, which is then passed on to all the actors as
// `cx.now()`. (If you want to run time faster or slower you
// could use another source of time.) So all calls in a batch of
// processing get the same `cx.now()` value. Also note that
// `Instant::now()` uses a Mutex on some platforms so it saves
// cycles to call it less often.
stakker.run(Instant::now(), false);
while stakker.not_shutdown() {
// Wait for next timer to expire. Here there's no I/O polling
// required to wait for external events, so just `sleep`
let maxdur = stakker.next_wait_max(Instant::now(), Duration::from_secs(60), false);
std::thread::sleep(maxdur);
// Run queue and timers
stakker.run(Instant::now(), false);
}
}
sourcepub fn shutdown_reason(&mut self) -> Option<StopCause>
pub fn shutdown_reason(&mut self) -> Option<StopCause>
Get the reason for shutdown, if shutdown was requested. After
calling this, the shutdown flag is cleared,
i.e. Core::not_shutdown
will return true
and the event
loop could continue to run.
sourcepub fn waker(&mut self, cb: impl FnMut(&mut Stakker, bool) + 'static) -> Waker
pub fn waker(&mut self, cb: impl FnMut(&mut Stakker, bool) + 'static) -> Waker
Register a wake handler callback, and obtain a Waker
instance which can be passed to another thread. The wake
handler will always be executed in the main thread. When
Waker::wake
is called in another thread, a wake-up
is scheduled to occur in the main thread, using the wake-up
mechanism provided by the I/O poller. Then when that wake-up
is received, the corresponding wake handler is executed. Note
that this is efficient – if many wake handlers are scheduled
around the same time, they share the same main thread wake-up.
The wake handler is called in the main thread with arguments
of (stakker, deleted)
. Note that there is a small chance of
a spurious wake call happening occasionally, so the wake
handler code must be ready for that. If deleted
is true
then the Waker
was dropped, and this wake handler is
also just about to be dropped.
This call panics if no I/O poller has yet set up a waker using
Stakker::set_poll_waker
.
Borrow two Share
instances mutably at the same time. This
will panic if they are the same instance.
Borrow three Share
instances mutably at the same time.
This will panic if any two are the same instance.
sourcepub fn log(
&mut self,
id: LogID,
level: LogLevel,
target: &str,
fmt: Arguments<'_>,
kvscan: impl Fn(&mut dyn LogVisitor)
)
pub fn log( &mut self, id: LogID, level: LogLevel, target: &str, fmt: Arguments<'_>, kvscan: impl Fn(&mut dyn LogVisitor) )
Pass a log-record to the current logger, if one is active and
if the log-level is enabled. Otherwise the call is ignored.
id
should be the logging-ID (obtained from actor.id()
or
cx.id()
, or core.log_span_open()
for non-actor spans) or 0
if the log-record doesn’t belong to any span. The arguments
are used to form the LogRecord
that is passed to the
logger.
Normally you would use a logging macro which wraps this call,
which would be provided by an external crate such as
stakker_log
.
This call does nothing unless the logger feature is enabled.
sourcepub fn log_check(&self, level: LogLevel) -> bool
pub fn log_check(&self, level: LogLevel) -> bool
Check whether a log-record with the given LogLevel
should
be logged
sourcepub fn log_span_open(
&mut self,
tag: &str,
parent_id: LogID,
kvscan: impl Fn(&mut dyn LogVisitor)
) -> LogID
pub fn log_span_open( &mut self, tag: &str, parent_id: LogID, kvscan: impl Fn(&mut dyn LogVisitor) ) -> LogID
Allocate a new logging-ID and write a LogLevel::Open
record to the logger. tag
will be included as the record’s
text, and should indicate what kind of span it is, e.g. the
type name for an actor. The tag would not normally contain
any dynamic information. If parent_id
is non-zero, then a
parent
key will be added with that value. kvscan
will be
called to add any other key-value pairs as required, which is
where the dynamic information should go.
This is used by actors on startup to allocate a logging-ID for
the span of the actor’s lifetime. However other code that
needs to do logging within a certain identifiable span can
also make use of this call. The new span should be related to
another span using parent_id
(if possible), and
Core::log_span_close
should be called when the span is
complete.
In the unlikely event that a program allocates 2^64 logging IDs, the IDs will wrap around to 1 again. If this is likely to cause a problem downstream, the logger implementation should detect this and warn or terminate as appropriate.
This call does nothing unless the logger feature is enabled.
sourcepub fn log_span_close(
&mut self,
id: LogID,
fmt: Arguments<'_>,
kvscan: impl Fn(&mut dyn LogVisitor)
)
pub fn log_span_close( &mut self, id: LogID, fmt: Arguments<'_>, kvscan: impl Fn(&mut dyn LogVisitor) )
Write a LogLevel::Close
record to the logger. fmt
is a
message which may give more information, e.g. the error
message in the case of a failure. kvscan
will be called to
add key-value pairs to the record.
This call does nothing unless the logger feature is enabled.
sourcepub fn access_core(&mut self) -> &mut Core
pub fn access_core(&mut self) -> &mut Core
Used in macros to get a Core
reference
sourcepub fn access_deferrer(&self) -> &Deferrer
pub fn access_deferrer(&self) -> &Deferrer
Used in macros to get a Deferrer
reference
sourcepub fn access_log_id(&self) -> LogID
pub fn access_log_id(&self) -> LogID
Used in macros to get the LogID
in case this is an actor
or context. Since it isn’t, this call returns 0.