2026-02-07 17:39:02 +00:00
|
|
|
[workspace]
|
feat: command interface — frontend-agnostic dispatch crate
Add swactor-command crate that decouples command dispatch from the
dashboard, enabling the same commands from REPL, REST, TUI, or any
future frontend.
- CommandRouter with dispatch(), CommandHandler trait, CommandMeta
- CommandRequest/CommandResponse (serde-serializable JSON protocol)
- CommandContext with Arc<Runtime> + optional StatsEnricher trait
- 9 built-in commands extracted from investigate.rs:
help, overview, workers, worker, actors, actor, hot, phases, diff,
shutdown
- REPL line parser (parse_line) with positional arg mapping
- REST adapter (from_query_params) for HTTP query parameters
- Dashboard integration: StatsEnricher impl for StatsCollector,
investigate.rs refactored to thin CommandRouter wrapper,
server.rs updated to share CommandRouter across handler threads
- 18 behavioral tests in crates/command/tests/command_api.rs
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-12 18:03:30 +00:00
|
|
|
members = [".", "crates/python", "crates/wasm", "crates/wasm-actor", "crates/simulation", "crates/runtime-dashboard", "crates/distribution", "crates/simulation-dashboard", "crates/std", "crates/command"]
|
2026-02-07 17:39:02 +00:00
|
|
|
exclude = ["tools/depgraph"]
|
|
|
|
|
|
2025-11-25 00:39:17 +00:00
|
|
|
[package]
|
|
|
|
|
name = "swactor"
|
|
|
|
|
version = "0.1.0"
|
|
|
|
|
edition = "2024"
|
feat: Stress tests, benchmarking, and non-failing queues and inboxes #3
Adds some basic benchmarking, stress tests. They still need to be properly examined to ensure they are testing the correct properties, but fit for "good enough". Implements the HybridChannel type, which features a channel buffer that can withstand overflows. It does so by providing a dequeue behind a mutex. Without overflow, will push messages into the lock free ArrayQueue implemented by crossbeam_queue; when that buffer fills, will use the locking portion provided by the Mutex<VecDequeue>.
In the future we can even further optimize this, perhaps with some linked list implementations of lock-free channels, but, like the benchmarks, this fits the "good enough" bar for now.
2026-01-26 07:14:00 +00:00
|
|
|
autobenches = false
|
2025-11-25 00:39:17 +00:00
|
|
|
|
2026-02-09 14:42:25 +00:00
|
|
|
[profile.bench]
|
|
|
|
|
debug = true
|
|
|
|
|
strip = false
|
|
|
|
|
|
2025-11-25 00:39:17 +00:00
|
|
|
[lib]
|
2026-02-07 17:39:02 +00:00
|
|
|
crate-type = ["rlib"]
|
2025-11-25 00:39:17 +00:00
|
|
|
|
2026-01-23 05:00:57 +00:00
|
|
|
[features]
|
|
|
|
|
default = ["getrandom"]
|
|
|
|
|
getrandom = ["dep:getrandom"]
|
2026-02-08 16:18:39 +00:00
|
|
|
serde = ["dep:serde"]
|
2026-02-09 09:04:57 +00:00
|
|
|
tracing = ["dep:tracing"]
|
feat: Stress tests, benchmarking, and non-failing queues and inboxes #3
Adds some basic benchmarking, stress tests. They still need to be properly examined to ensure they are testing the correct properties, but fit for "good enough". Implements the HybridChannel type, which features a channel buffer that can withstand overflows. It does so by providing a dequeue behind a mutex. Without overflow, will push messages into the lock free ArrayQueue implemented by crossbeam_queue; when that buffer fills, will use the locking portion provided by the Mutex<VecDequeue>.
In the future we can even further optimize this, perhaps with some linked list implementations of lock-free channels, but, like the benchmarks, this fits the "good enough" bar for now.
2026-01-26 07:14:00 +00:00
|
|
|
no_random = [] # compile without access to a source of randomness
|
2026-02-09 19:05:37 +00:00
|
|
|
transport = [] # transport-agnostic messaging (no mandatory deps; codec is user-provided)
|
2026-01-23 05:00:57 +00:00
|
|
|
|
2025-11-25 00:39:17 +00:00
|
|
|
[dependencies]
|
2026-01-23 05:00:57 +00:00
|
|
|
getrandom = { version = "0.2", optional = true }
|
2026-02-08 16:18:39 +00:00
|
|
|
serde = { version = "1", features = ["derive"], optional = true }
|
2026-02-09 09:04:57 +00:00
|
|
|
tracing = { version = "0.1", optional = true }
|
2026-01-23 05:00:57 +00:00
|
|
|
crossbeam-queue = "0.3.12"
|
2026-02-06 11:25:37 +00:00
|
|
|
crossbeam-utils = "0.8.21"
|
feat: Stress tests, benchmarking, and non-failing queues and inboxes #3
Adds some basic benchmarking, stress tests. They still need to be properly examined to ensure they are testing the correct properties, but fit for "good enough". Implements the HybridChannel type, which features a channel buffer that can withstand overflows. It does so by providing a dequeue behind a mutex. Without overflow, will push messages into the lock free ArrayQueue implemented by crossbeam_queue; when that buffer fills, will use the locking portion provided by the Mutex<VecDequeue>.
In the future we can even further optimize this, perhaps with some linked list implementations of lock-free channels, but, like the benchmarks, this fits the "good enough" bar for now.
2026-01-26 07:14:00 +00:00
|
|
|
|
2026-02-06 11:25:37 +00:00
|
|
|
[dev-dependencies]
|
|
|
|
|
criterion = { version = "0.5", features = ["html_reports"] }
|
2026-02-13 07:11:24 +00:00
|
|
|
proptest = "1"
|
|
|
|
|
proptest-state-machine = "0.3"
|
|
|
|
|
swactor-std = { path = "crates/std" }
|
2026-02-06 11:25:37 +00:00
|
|
|
|
|
|
|
|
[[bench]]
|
|
|
|
|
name = "runtime_benchmarks"
|
|
|
|
|
harness = false
|
2026-02-06 14:45:19 +00:00
|
|
|
|
2026-02-09 14:02:40 +00:00
|
|
|
[[bench]]
|
|
|
|
|
name = "mt_benchmarks"
|
|
|
|
|
harness = false
|
2026-02-11 15:23:26 +00:00
|
|
|
|
2026-02-13 07:11:24 +00:00
|
|
|
[[bench]]
|
|
|
|
|
name = "hasher_benchmarks"
|
|
|
|
|
harness = false
|
|
|
|
|
|
2026-02-11 15:23:26 +00:00
|
|
|
[[example]]
|
|
|
|
|
name = "tcp_ping_pong"
|
|
|
|
|
required-features = ["transport"]
|