RegexSolver v1

.github/workflows/deploy.yml

@@ -3,17 +3,65 @@ name: Publish to crates.io
 on:
   push:
     tags:
-      - 'v*'
+      - "v*"
+
+permissions:
+  contents: read
+
+env:
+  CARGO_TERM_COLOR: always
+  RUSTDOCFLAGS: -D warnings
 
 jobs:
-  build:
+  publish:
     runs-on: ubuntu-latest
+    # Gate publishing behind a protected GitHub environment so the crates.io
+    # token is only exposed to reviewed, tag-triggered runs. Configure required
+    # reviewers on this environment in the repository settings.
+    environment: crates-io
     steps:
-      - uses: actions/checkout@v4
-      - uses: actions-rs/toolchain@v1
+      - uses: actions/checkout@34e114876b0b11c390a56381ad16ebd13914f8d5 # v4
+      - uses: dtolnay/rust-toolchain@fa04a1451ff1842e2626ccb99004d0195b455a88 # master
         with:
-            toolchain: stable
-            override: true
-      - uses: katyo/publish-crates@v2
+          toolchain: stable
+          components: rustfmt, clippy
+
+      # Never publish a crate whose version disagrees with the pushed tag.
+      - name: Verify tag matches Cargo.toml version
+        run: |
+          crate_version="$(cargo metadata --no-deps --format-version 1 | jq -r '.packages[0].version')"
+          tag_version="${GITHUB_REF_NAME#v}"
+          if [ "$crate_version" != "$tag_version" ]; then
+            echo "::error::Tag ${GITHUB_REF_NAME} (${tag_version}) does not match Cargo.toml version ${crate_version}"
+            exit 1
+          fi
+
+      - name: Format
+        run: cargo fmt --all -- --check
+      - name: Lint
+        run: |
+          cargo clippy --locked --all-targets -- -D warnings
+          cargo clippy --locked --no-default-features --all-targets -- -D warnings
+      - name: Test
+        run: |
+          cargo test --locked
+          cargo test --locked --no-default-features
+      - name: Docs
+        run: cargo doc --locked --no-deps
+
+      # Also verify the declared MSRV before publishing, like CI does.
+      - name: Read MSRV from Cargo.toml
+        id: msrv
+        run: echo "msrv=$(grep -m1 '^rust-version' Cargo.toml | cut -d'"' -f2)" >> "$GITHUB_OUTPUT"
+      - uses: dtolnay/rust-toolchain@fa04a1451ff1842e2626ccb99004d0195b455a88 # master
         with:
-            registry-token: ${{ secrets.CRATES_API_TOKEN }}
+          toolchain: ${{ steps.msrv.outputs.msrv }}
+      - name: Check on MSRV
+        run: |
+          cargo +${{ steps.msrv.outputs.msrv }} check --locked --all-features
+          cargo +${{ steps.msrv.outputs.msrv }} check --locked --no-default-features
+
+      - name: Publish
+        env:
+          CARGO_REGISTRY_TOKEN: ${{ secrets.CRATES_API_TOKEN }}
+        run: cargo +stable publish --locked

.github/workflows/rust.yml

@@ -2,23 +2,68 @@ name: Rust
 
 on:
   push:
-    branches: [ "main" ]
+    branches: ["main", "v1"]
   pull_request:
-    branches: [ "main" ]
+  workflow_dispatch:
+
+permissions:
+  contents: read
 
 env:
   CARGO_TERM_COLOR: always
+  RUSTDOCFLAGS: -D warnings
 
 jobs:
   build:
+    runs-on: ubuntu-latest
+
+    steps:
+      - uses: actions/checkout@34e114876b0b11c390a56381ad16ebd13914f8d5 # v4
+      - uses: dtolnay/rust-toolchain@fa04a1451ff1842e2626ccb99004d0195b455a88 # master
+        with:
+          toolchain: stable
+          components: rustfmt, clippy
+      - name: Format
+        run: cargo fmt --all -- --check
+      - name: Build
+        run: cargo build --locked --verbose
+      - name: Test & Lint
+        run: |
+          cargo test --locked
+          cargo clippy --locked --all-targets -- -D warnings
+      - name: Test & Lint (no default features)
+        run: |
+          cargo test --locked --no-default-features
+          cargo clippy --locked --no-default-features --all-targets -- -D warnings
+      - name: Docs
+        run: cargo doc --locked --no-deps
+
+  # Keep the declared `rust-version` (MSRV) honest, for both feature configs.
+  msrv:
+    runs-on: ubuntu-latest
+
+    steps:
+      - uses: actions/checkout@34e114876b0b11c390a56381ad16ebd13914f8d5 # v4
+      # Single source of truth: the MSRV comes from Cargo.toml.
+      - name: Read MSRV from Cargo.toml
+        id: msrv
+        run: echo "msrv=$(grep -m1 '^rust-version' Cargo.toml | cut -d'"' -f2)" >> "$GITHUB_OUTPUT"
+      - uses: dtolnay/rust-toolchain@fa04a1451ff1842e2626ccb99004d0195b455a88 # master
+        with:
+          toolchain: ${{ steps.msrv.outputs.msrv }}
+      - name: Check on MSRV
+        run: |
+          cargo check --locked --all-features
+          cargo check --locked --no-default-features
 
+  audit:
     runs-on: ubuntu-latest
+    permissions:
+      contents: read
+      issues: write # audit-check files an issue when a vulnerability is found
 
     steps:
-    - uses: actions/checkout@v4
-    - name: Build
-      run: cargo build --all-features --verbose
-    - name: Test & Lint
-      run: |
-        cargo test --all-features
-        cargo clippy --all-features
+      - uses: actions/checkout@34e114876b0b11c390a56381ad16ebd13914f8d5 # v4
+      - uses: rustsec/audit-check@69366f33c96575abad1ee0dba8212993eecbe998 # v2.0.0
+        with:
+          token: ${{ secrets.GITHUB_TOKEN }}

.gitignore

@@ -3,10 +3,6 @@
 debug/
 target/
 
-# Remove Cargo.lock from gitignore if creating an executable, leave it for libraries
-# More information here https://doc.rust-lang.org/cargo/guide/cargo-toml-vs-cargo-lock.html
-Cargo.lock
-
 # These are backup files generated by rustfmt
 **/*.rs.bk
 
@@ -18,4 +14,7 @@ Cargo.lock
 #  be found at https://github.com/github/gitignore/blob/main/Global/JetBrains.gitignore
 #  and can be added to the global gitignore or merged into this file.  For a more nuclear
 #  option (not recommended) you can uncomment the following to ignore the entire idea folder.
-#.idea/
+#.idea/
+
+# cargo mutants output
+mutants.out*/

CHANGELOG.md

@@ -0,0 +1,128 @@
+# Changelog
+
+All notable changes to this crate are documented in this file.
+
+The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
+and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
+
+## [Unreleased] - upcoming 1.0.0
+
+This is a major redesign of the public API around the `Term` enum (wrapping
+either a `RegularExpression` or a `FastAutomaton`), which dispatches each
+operation to the cheaper representation when possible. See the crate-level
+docs for the architecture. Almost the entire public surface changed; highlights
+below.
+
+### Added
+- New `Term` constructors and conversions: `new_empty`, `new_total`,
+  `new_empty_string`, `from_pattern`, `from_regex(RegularExpression)`,
+  `from_automaton(FastAutomaton)`, plus `From<RegularExpression>`,
+  `From<FastAutomaton>`, `FromStr`, `Display`, and `Default` (= `new_empty`).
+- New `Term` operations: `concat`, `complement`, `determinize`, `minimize`,
+  `matches`, `is_deterministic`, `is_minimal`, `is_finite`, `to_pattern`, and
+  `iter_strings` (lazy `StringGenerator` iterator).
+- `RegularExpression::MAX_NESTING_DEPTH` and a depth check in `to_automaton`
+  that returns the new `EngineError::RegexTooDeeplyNested` for hand-built trees
+  nested past the limit, instead of overflowing the stack.
+- `to_automaton` returns `Cow` to avoid unnecessary cloning; `to_regex`
+  returns `Result<Cow<…>, EngineError>` (see Changed — it is now fallible).
+- `FastAutomaton` gained corresponding low-level constructors/operations
+  (`new_empty`, `new_total`, `new_empty_string`, `determinize`, `minimize`
+  using Hopcroft's algorithm, `is_minimal`, `unaccept`, `print_dot`,
+  `try_add_transition`) and inspection helpers (`states`, `direct_states`,
+  `transitions_from`, `transitions_to_vec`, `has_transition`, ...).
+- New `EngineError` variants: `InvalidRepetitionBounds`,
+  `IncompatibleSpanningSet`, `DeterministicAutomatonRequired`,
+  `UnsupportedRegexFeature`, `RegexTooDeeplyNested`; the enum is now
+  `#[non_exhaustive]`.
+- `tracing` instrumentation on the core `Term`, `FastAutomaton`, and
+  `RegularExpression` operations (concat, union, intersection, difference,
+  complement, repeat, determinize, minimize, equivalence/subset checks,
+  string generation, conversions). No-op unless a `tracing` subscriber is
+  installed.
+- Parallel (Rayon-backed) variants of union/intersection for >3 operands,
+  gated behind the default-on `parallel` feature, with sequential fallbacks
+  for `--no-default-features`.
+- `cargo fmt --check`, `cargo clippy -- -D warnings`, `cargo doc --no-deps`
+  (with `RUSTDOCFLAGS=-D warnings`), and a dependency vulnerability audit
+  (`rustsec/audit-check`) to CI.
+- `regex_charclass` is re-exported at the crate root, so automata can be
+  built by hand (`Char`, `CharRange`) without adding a separately
+  version-matched dependency.
+- `NoHashHasher`, the crate-owned no-op hasher behind the `IntSet` state-id
+  sets (replaces the `nohash-hasher` dependency, keeping 0.x types out of
+  the public API).
+- `EngineError` implements `Clone`; `StringGenerator` implements `Debug` and
+  `FusedIterator`; `ExecutionProfileBuilder` implements `Debug` and `Clone`.
+
+### Changed
+- `Term::to_regex`/`to_pattern` and `FastAutomaton::to_regex` are now fallible
+  (`Result<_, EngineError>`) and honor the `ExecutionProfile` timeout, since
+  state elimination can grow super-polynomially on adversarial automata.
+  `Term`'s `Display` therefore renders a pattern only for regex-backed terms;
+  automaton-backed terms display as Graphviz DOT (use `to_pattern` for a
+  parseable pattern).
+- `ExecutionProfile` redesigned as an immutable, thread-local-aware config
+  built via the new `ExecutionProfileBuilder`, governing execution timeouts,
+  state-count limits, and an `implicit_determinization` toggle.
+- `union`/`intersection`/`concat` now take
+  `impl IntoIterator<Item = impl Borrow<Term>>` instead of `&[Term]`, so
+  `&[a, b]`, `[&a, &b]`, and `Vec<Term>` all work without cloning.
+- `repeat` now takes `impl RangeBounds<u32>` (e.g. `3..6`, `..=2`) instead of
+  explicit min/max parameters.
+- `generate_strings` now takes `(limit, offset)` for pagination instead of a
+  single `count`.
+- `is_empty`, `is_total`, and `is_empty_string` now return
+  `Result<bool, EngineError>` instead of `bool`.
+- `are_equivalent`/`is_subset_of` renamed to `equivalent`/`subset`.
+- `subtraction` renamed to `difference`, kept single-operand by design.
+- Renamed `FastAutomaton::as_dot` to `to_dot` (old printing `to_dot` is now
+  `print_dot`), matching the crate's `to_*` convention for allocating
+  conversions (`to_pattern`, `to_regex`, `to_automaton`, `to_range`).
+- Renamed `get_*` accessors to drop the `get_` prefix, per the Rust API
+  Guidelines' C-GETTER convention: `Term::get_length` to `length`,
+  `Term::get_cardinality` to `cardinality`, `FastAutomaton::get_length` to
+  `length`, `FastAutomaton::get_cardinality` to `cardinality`,
+  `FastAutomaton::get_number_of_states` to `number_of_states`,
+  `FastAutomaton::get_condition` to `condition`,
+  `FastAutomaton::get_start_state` to `start_state`,
+  `FastAutomaton::get_accept_states` to `accept_states`,
+  `FastAutomaton::get_spanning_set` to `spanning_set`,
+  `FastAutomaton::get_live_states` to `live_states`,
+  `FastAutomaton::get_spanning_bases` to `spanning_bases`,
+  `RegularExpression::get_length` to `length`,
+  `RegularExpression::get_cardinality` to `cardinality`,
+  `SpanningSet::get_spanning_ranges` to `spanning_ranges`,
+  `SpanningSet::get_number_of_spanning_ranges` to
+  `number_of_spanning_ranges`, `SpanningSet::get_spanning_range` to
+  `spanning_range`, `SpanningSet::get_rest` to `rest`,
+  `Condition::get_cardinality` to `cardinality`,
+  `Condition::get_binary_representation` to `binary_representation`,
+  `ConditionConverter::get_from_spanning_set`/`get_to_spanning_set` to
+  `from_spanning_set`/`to_spanning_set`.
+- Error messages follow the std convention (lowercase, no trailing
+  punctuation) so they compose cleanly when wrapped by callers.
+- Edition bumped to 2024 and `Cargo.toml` metadata (`description`,
+  `categories`) updated.
+
+### Removed
+- The `serde` feature and all serialization, FAIR (base85) encoding,
+  encryption, and compression support (`serde`, `ciborium`, `z85`,
+  `aes-gcm-siv`, `sha2`, `flate2` dependencies).
+- `Term::get_details` and the `Details` type.
+- The `tokenizer` module.
+- Unused `log`, `rand`, and `lazy_static` dependencies, and the `regex`
+  crate dependency (now dev-only, used by integration tests).
+- The `nohash-hasher` dependency (replaced by the crate-owned
+  `NoHashHasher`; see Added).
+- `EngineError` variants `AutomatonShouldBeDeterministic`, `TooMuchTerms`,
+  `ConditionIndexOutOfBound`, `TokenError`, and the `is_server_error` method.
+- The `max_number_of_terms` execution-profile limit (no longer enforced).
+
+## Earlier releases
+
+Releases prior to 1.0.0 (`v0.1.0` through `v0.3.1`) predate this changelog;
+see the [GitHub tags](https://github.com/RegexSolver/regexsolver/tags) and
+commit history for details.
+
+[Unreleased]: https://github.com/RegexSolver/regexsolver/compare/v0.3.1...HEAD