struct DrainState<'cfg> {Show 20 fields
total_units: usize,
queue: DependencyQueue<Unit, Artifact, Job>,
messages: Arc<Queue<Message>>,
diag_dedupe: DiagDedupe<'cfg>,
warning_count: HashMap<JobId, WarningCount>,
active: HashMap<JobId, Unit>,
compiled: HashSet<PackageId>,
documented: HashSet<PackageId>,
scraped: HashSet<PackageId>,
counts: HashMap<PackageId, usize>,
progress: Progress<'cfg>,
next_id: u32,
timings: Timings<'cfg>,
tokens: Vec<Acquired>,
rustc_tokens: HashMap<JobId, Vec<Acquired>>,
to_send_clients: BTreeMap<JobId, Vec<Client>>,
pending_queue: Vec<(Unit, Job, usize)>,
print: DiagnosticPrinter<'cfg>,
finished: usize,
per_package_future_incompat_reports: Vec<FutureIncompatReportPackage>,
}
Expand description
This structure is backed by the DependencyQueue
type and manages the
actual compilation step of each package. Packages enqueue units of work and
then later on the entire graph is processed and compiled.
It is created from JobQueue when we have fully assembled the crate graph (i.e., all package dependencies are known).
Message queue
Each thread running a process uses the message queue to send messages back to the main thread. The main thread coordinates everything, and handles printing output.
It is important to be careful which messages use push
vs push_bounded
.
push
is for priority messages (like tokens, or “finished”) where the
sender shouldn’t block. We want to handle those so real work can proceed
ASAP.
push_bounded
is only for messages being printed to stdout/stderr. Being
bounded prevents a flood of messages causing a large amount of memory
being used.
push
also avoids blocking which helps avoid deadlocks. For example, when
the diagnostic server thread is dropped, it waits for the thread to exit.
But if the thread is blocked on a full queue, and there is a critical
error, the drop will deadlock. This should be fixed at some point in the
future. The jobserver thread has a similar problem, though it will time
out after 1 second.
Fields§
§total_units: usize
§queue: DependencyQueue<Unit, Artifact, Job>
§messages: Arc<Queue<Message>>
§diag_dedupe: DiagDedupe<'cfg>
Diagnostic deduplication support.
warning_count: HashMap<JobId, WarningCount>
Count of warnings, used to print a summary after the job succeeds
active: HashMap<JobId, Unit>
§compiled: HashSet<PackageId>
§documented: HashSet<PackageId>
§scraped: HashSet<PackageId>
§counts: HashMap<PackageId, usize>
§progress: Progress<'cfg>
§next_id: u32
§timings: Timings<'cfg>
§tokens: Vec<Acquired>
Tokens that are currently owned by this Cargo, and may be “associated” with a rustc process. They may also be unused, though if so will be dropped on the next loop iteration.
Note that the length of this may be zero, but we will still spawn work, as we share the implicit token given to this Cargo process with a single rustc process.
rustc_tokens: HashMap<JobId, Vec<Acquired>>
rustc per-thread tokens, when in jobserver-per-rustc mode.
to_send_clients: BTreeMap<JobId, Vec<Client>>
This represents the list of rustc jobs (processes) and associated clients that are interested in receiving a token.
pending_queue: Vec<(Unit, Job, usize)>
The list of jobs that we have not yet started executing, but have
retrieved from the queue
. We eagerly pull jobs off the main queue to
allow us to request jobserver tokens pretty early.
print: DiagnosticPrinter<'cfg>
§finished: usize
How many jobs we’ve finished
per_package_future_incompat_reports: Vec<FutureIncompatReportPackage>
Implementations§
source§impl<'cfg> DrainState<'cfg>
impl<'cfg> DrainState<'cfg>
fn spawn_work_if_possible<'s>(
&mut self,
cx: &mut Context<'_, '_>,
jobserver_helper: &HelperThread,
scope: &'s Scope<'s, '_>
) -> CargoResult<()>
fn has_extra_tokens(&self) -> bool
fn pop_waiting_client(&mut self) -> (JobId, Client)
fn grant_rustc_token_requests(&mut self) -> CargoResult<()>
fn handle_event(
&mut self,
cx: &mut Context<'_, '_>,
jobserver_helper: &HelperThread,
plan: &mut BuildPlan,
event: Message
) -> Result<(), ErrorToHandle>
fn wait_for_events(&mut self) -> Vec<Message> ⓘ
sourcefn drain_the_queue<'s>(
self,
cx: &mut Context<'_, '_>,
plan: &mut BuildPlan,
scope: &'s Scope<'s, '_>,
jobserver_helper: &HelperThread
) -> Option<Error>
fn drain_the_queue<'s>(
self,
cx: &mut Context<'_, '_>,
plan: &mut BuildPlan,
scope: &'s Scope<'s, '_>,
jobserver_helper: &HelperThread
) -> Option<Error>
This is the “main” loop, where Cargo does all work to run the compiler.
This returns an Option to prevent the use of ?
on Result
types
because it is important for the loop to carefully handle errors.
fn handle_error(
&self,
shell: &mut Shell,
err_state: &mut ErrorsDuringDrain,
new_err: impl Into<ErrorToHandle>
)
fn tick_progress(&mut self)
fn name_for_progress(&self, unit: &Unit) -> String
sourcefn run<'s>(
&mut self,
unit: &Unit,
job: Job,
cx: &Context<'_, '_>,
scope: &'s Scope<'s, '_>
)
fn run<'s>(
&mut self,
unit: &Unit,
job: Job,
cx: &Context<'_, '_>,
scope: &'s Scope<'s, '_>
)
Executes a job.
Fresh jobs block until finished (which should be very fast!), Dirty jobs will spawn a thread in the background and return immediately.
fn emit_warnings(
&mut self,
msg: Option<&str>,
unit: &Unit,
cx: &mut Context<'_, '_>
) -> CargoResult<()>
fn bump_warning_count(&mut self, id: JobId, emitted: bool, fixable: bool)
sourcefn report_warning_count(
&mut self,
config: &Config,
id: JobId,
rustc_workspace_wrapper: &Option<PathBuf>
)
fn report_warning_count(
&mut self,
config: &Config,
id: JobId,
rustc_workspace_wrapper: &Option<PathBuf>
)
Displays a final report of the warnings emitted by a particular job.
fn finish(
&mut self,
id: JobId,
unit: &Unit,
artifact: Artifact,
cx: &mut Context<'_, '_>
) -> CargoResult<()>
fn note_working_on(
&mut self,
config: &Config,
ws_root: &Path,
unit: &Unit,
fresh: &Freshness
) -> CargoResult<()>
fn back_compat_notice(&self, cx: &Context<'_, '_>, unit: &Unit) -> CargoResult<()>
Auto Trait Implementations§
impl<'cfg> !RefUnwindSafe for DrainState<'cfg>
impl<'cfg> !Send for DrainState<'cfg>
impl<'cfg> !Sync for DrainState<'cfg>
impl<'cfg> Unpin for DrainState<'cfg>
impl<'cfg> !UnwindSafe for DrainState<'cfg>
Blanket Implementations§
Layout§
Note: Most layout information is completely unstable and may even differ between compilations. The only exception is types with certain repr(...)
attributes. Please see the Rust Reference’s “Type Layout” chapter for details on type layout guarantees.
Size: 1240 bytes