Cargo.toml

[workspace]
resolver = "2"
members = [
    "crates/plugmem-arena",
    "crates/plugmem-core",
    "crates/plugmem-host",
    "crates/plugmem-cli",
    "crates/plugmem-mcp",
    "crates/plugmem-napi",
    "crates/plugmem-py",
    "crates/plugmem-testgen",
    "tools/bench-matrix",
    "tools/bench-charts",
    "tools/wasm-probe",
]
# The fuzz targets are their own workspace: they build on nightly against a
# sanitizer runtime and must not enter this lockfile or lint set.
exclude = ["fuzz"]

[workspace.package]
version = "0.12.0"
edition = "2024"
authors = ["Mansur Azatbek <mansur62624@gmail.com>"]
license = "MIT"
repository = "https://github.com/m62624/plugmem"
homepage = "https://github.com/m62624/plugmem"

[workspace.dependencies]
# Local crates (version pinned so the crates can be published to crates.io)
plugmem-arena = { path = "crates/plugmem-arena", version = "0.12.0", default-features = false }
plugmem-core = { path = "crates/plugmem-core", version = "0.12.0", default-features = false }
plugmem-host = { path = "crates/plugmem-host", version = "0.12.0" }
# Internal, never published: path-only, so cargo strips it from published
# manifests when it appears as a dev-dependency.
plugmem-testgen = { path = "crates/plugmem-testgen" }

# Core dependencies (no_std; each entry build-verified against wasm32v1-none — specs/08)
thiserror = { version = "2.0.18", default-features = false }
xxhash-rust = { version = "0.8.17", default-features = false, features = ["xxh3"] }
hashbrown = { version = "0.17.1", default-features = false }
# Tokenizer decomposition helper retained for the search policy's
# diacritic/combining-mark rules.
unicode-normalization = { version = "0.1.25", default-features = false }
# ICU4X is the production Unicode backend: NFKC, locale-neutral lowercase,
# Unicode properties, and UAX #29 word segmentation with complex-script data.
icu-casemap = { version = "2.2.0", package = "icu_casemap", default-features = false, features = ["compiled_data"] }
icu-locale-core = { version = "2.2.0", package = "icu_locale_core", default-features = false }
icu-normalizer = { version = "2.2.0", package = "icu_normalizer", default-features = false, features = ["compiled_data", "utf8_iter"] }
icu-properties = { version = "2.2.0", package = "icu_properties", default-features = false, features = ["compiled_data"] }
icu-segmenter = { version = "2.2.0", package = "icu_segmenter", default-features = false, features = ["compiled_data", "auto"] }
writeable = { version = "0.6.3", default-features = false }
# Float math for no_std (BM25 idf/tf norms, recency decay).
libm = { version = "0.2.16", default-features = false }
# Optional serde support, gated behind each crate's `serde` feature (off by
# default). no_std-compatible: `derive` + `alloc`, no std. Used only to
# Serialize/Deserialize the public data types for JSON-facing consumers.
serde = { version = "1.0.229", default-features = false, features = ["derive", "alloc"] }

# Wrapper dependencies (std)
clap = { version = "4.6.2", features = ["derive"] }
# Canonical per-user config/data paths for Linux, macOS and Windows.
directories = "6.0.0"
# Cross-platform mmap for the host's zero-copy read-only open (specs/16).
# 0.9.11: MSRV 1.65, MIT/Apache-2.0, covers linux/macos/windows ×
# x86_64/aarch64. (The original `memmap` is unmaintained — RUSTSEC-2020-0077.)
memmap2 = "0.9.11"
# Safe self-referential struct: ReadOnlyDatabase owns the `Mmap` and a
# `Memory<'a>` borrowing it. Keeps the self-reference safe so the only
# `unsafe` on the read-only path is the (inherent) mmap call itself.
self_cell = "1.3.0"
ureq = "3.3.0"
serde_json = "1.0.150"
# config.toml parsing, shared by the host `config` loader and the CLI. Parse
# only — `toml::Table` via FromStr; no serializer needed.
toml = { version = "0.8", default-features = false, features = ["parse"] }

# Native Node.js addon (plugmem-napi): napi-rs bindings over plugmem-host.
# Node-API version pinned to 8 (Node 12.22+/14.17+/16+; we build for 26). No
# `async`/`tokio_rt` features — concurrency rides the libuv threadpool, not tokio.
napi = { version = "2.16", default-features = false, features = ["napi8"] }
napi-derive = "2.16"
napi-build = "2.1"

# Python extension module (plugmem-py): PyO3 over plugmem-host.
#
# `abi3-py310` is what makes one wheel per platform cover every GIL build of
# Python from 3.10 on: it restricts the binding to CPython's stable ABI, which
# does not change between versions. It costs nothing here because this crate
# touches no CPython API directly — everything goes through PyO3.
#
# `experimental-inspect` is what `maturin generate-stubs` reads to emit the
# `.pyi`, so the stubs come from the same macros as the binding instead of
# being written by hand and drifting.
#
# No `pyo3-async-runtimes`: the engine is CPU-bound and has no runtime to
# bridge. Concurrency is `Python::detach` plus the caller's own threads.
pyo3 = { version = "0.29.1", features = ["abi3-py310"] }
# Type stubs (`.pyi`) for the Python binding, which is what gives an IDE
# completion and a type checker something to check against a compiled module.
#
# Not PyO3's own `experimental-inspect` + `maturin generate-stubs`: measured on
# this crate at PyO3 0.29.1, that pair emits a stub containing nothing but
# `def __getattr__(name: str) -> Incomplete`, with the feature verifiably
# enabled. This crate reads the same macros the binding is written with and
# emits real signatures.
pyo3-stub-gen = { version = "0.23.0", default-features = false }

# Dev dependencies
proptest = "1.11.0"
criterion = "0.8.2"
insta = "1.48.0"
# Cross-platform resident-set-size probe (linux/macos/windows) for the host
# overlay RAM-win test (specs/16 §9): proves an mmap overlay open residents
# far less physical memory than an owned open of the same file. Dev-only.
memory-stats = "1.2.0"
# rand_xoshiro was reserved for testgen (specs/08) and dropped: the
# generator implements its own PRNG in place (specs/11 §3) so corpus
# reproducibility never depends on an external crate's stream stability.

# Every crate opts in via `[lints] workspace = true`. Documentation is not
# optional in this project: any public item without docs fails the build.
# All code and docs are English-only.
[workspace.lints.rust]
missing_docs = "deny"
unsafe_op_in_unsafe_fn = "deny"
# `cargo-tarpaulin` sets `--cfg=tarpaulin`; the hardening jobs pass
# `--cfg=valgrind` (valgrind sets no cfg of its own). Declare both so
# `#[cfg_attr(tarpaulin, ignore)]` / `#[cfg_attr(valgrind, ignore)]` on the
# timing-sensitive concurrency stress tests do not trip the unexpected-cfg lint.
# Those tests are covered by normal `cargo test`; under tarpaulin's ptrace or
# valgrind's ~100x slowdown their wall-clock watchdogs only time out (a false
# deadlock) without adding coverage. (`miri` is a built-in cfg — no declaration
# needed.)
unexpected_cfgs = { level = "warn", check-cfg = ['cfg(tarpaulin)', 'cfg(valgrind)'] }

[workspace.lints.rustdoc]
broken_intra_doc_links = "deny"

# Config for 'dist'
[workspace.metadata.dist]
# The preferred dist version to use in CI (Cargo.toml SemVer syntax)
cargo-dist-version = "0.31.0"
# CI backends to support
ci = "github"
# The installers to generate for each app
installers = ["shell", "powershell", "homebrew", "msi"]
# A GitHub repo to push Homebrew formulas to
tap = "m62624/homebrew-plugmem"
# Publish jobs to run in CI
publish-jobs = ["homebrew"]
# Skip checking whether the specified configuration files are up to date
allow-dirty = ["ci"]
# Target platforms to build apps for (Rust target-triple syntax)
targets = ["aarch64-apple-darwin", "aarch64-unknown-linux-gnu", "aarch64-pc-windows-msvc", "x86_64-apple-darwin", "x86_64-unknown-linux-gnu", "x86_64-pc-windows-msvc"]
# Which actions to run on pull requests
pr-run-mode = "plan"
# Path that installers should place binaries in
install-path = "CARGO_HOME"
# Whether to install an updater program
install-updater = false

# Build aarch64-windows on a NATIVE Windows runner. By default dist
# cross-builds it on Linux via cargo-xwin, but the `.msi` needs WiX v3's
# `candle.exe` (a Windows binary that can't run under xwin) — so it must
# build on Windows. This is the Windows-specific fix carried over from the
# elenchus release pipeline.
[workspace.metadata.dist.github-custom-runners]
aarch64-pc-windows-msvc = "windows-2022"

# The hot loops are split across crates by design — the arena holds the page
# walk and the byte moves, the core holds the scoring and the distance kernels
# — so without cross-crate inlining every posting decode and every slot read
# crosses a call boundary the optimizer cannot see through. `lto = "fat"` plus a
# single codegen unit is what lets it. Benches inherit this: a benchmark built
# with different inlining measures a different program.
[profile.release]
lto = "fat"
codegen-units = 1

[profile.bench]
inherits = "release"

# `dist` keeps `thin` deliberately: the release pipeline builds six targets and
# pays fat LTO six times for a binary whose cost is dominated by I/O, not by the
# scoring loops a benchmark exercises. `codegen-units` is restated for the same
# reason and not left to inheritance — the single unit above buys inlining a
# benchmark can measure, at the price of serializing codegen, and `dist` pays
# that price six times over for a binary that cannot show the gain.
[profile.dist]
inherits = "release"
lto = "thin"
codegen-units = 16