Module job 

Re-exports

use std::io::Write; 🔒

use std::ops::ControlFlow; 🔒

use std::sync::Arc; 🔒

use std::iter; 🔒

use std::mem; 🔒

use rustc_data_structures::fx::FxHashMap; 🔒

use rustc_data_structures::fx::FxHashSet; 🔒

use rustc_data_structures::hash_table::HashTable; 🔒

use rustc_data_structures::sync::DynSend; 🔒

use rustc_data_structures::sync::DynSync; 🔒

use rustc_middle::queries::TaggedQueryKey; 🔒

use rustc_middle::query::ActiveKeyStatus; 🔒

use rustc_middle::query::QueryCache; 🔒

use rustc_middle::query::QueryCycle; 🔒

use rustc_middle::query::QueryJob; 🔒

use rustc_middle::query::QueryJobId; 🔒

use rustc_middle::query::QueryKey; 🔒

use rustc_middle::query::QueryLatch; 🔒

use rustc_middle::query::QueryStackFrame; 🔒

use rustc_middle::query::QueryVTable; 🔒

use rustc_middle::query::QueryWaiter; 🔒

use rustc_middle::ty::TyCtxt; 🔒

use rustc_span::DUMMY_SP; 🔒

use rustc_span::Span; 🔒

use tracing::debug; 🔒

use crate::query_vtables::for_each_query_vtable; 🔒

Structs

AbstractedWaiter 🔒

This abstracts over non-resumable waiters which are found in QueryJob’s parent field and resumable waiters are in latch field.

QueryJobInfo 🔒

QueryJobMap

Map from query job IDs to job information collected by collect_active_query_jobs.

Enums

CollectActiveJobsKind

Functions

abstracted_waiters_of 🔒

Returns all the non-resumable and resumable waiters of a query. This is used so we can uniformly loop over both non-resumable and resumable waiters.

break_query_cycle

Detects query cycles by using depth first search over all active query jobs. If a query cycle is found it will break the cycle by finding an edge which uses a query latch and then resuming that waiter.

collect_active_query_jobs

Returns a map of currently active query jobs, collected from all queries.

collect_active_query_jobs_inner 🔒

Internal plumbing for collecting the set of active jobs for this query.

connected_to_root 🔒

Finds out if there’s a path to the compiler root (aka. code which isn’t in a query) from query without going through any of the queries in visited. This is achieved with a depth first search.

find_and_process_cycle 🔒

Looks for a query cycle starting at query. Returns a waiter to resume if a cycle is found.

find_cycle 🔒

Looks for a query cycle by doing a depth first search starting at query. span is the reason for the query to execute. This is initially DUMMY_SP. If a cycle is detected, this initial value is replaced with the span causing the cycle. stack will contain just the cycle on return if detected.

find_cycle_in_stack 🔒

find_dep_kind_root 🔒

Finds the query job closest to the root that is for the same query method as id (but not necessarily the same query key), and returns information about it.

print_query_stack

process_cycle 🔒

Processes a found query cycle into a Cycle

Type Aliases

ResumableWaiterLocation 🔒

The locaton of a resumable waiter. The usize is the index into waiters in the query’s latch. We’ll use this to remove the waiter if we’re waking it up in find_and_process_cycle.