Add support of Arbitrary

Author: greyblake

.github/workflows/ci.yml

@@ -36,6 +36,12 @@ jobs:
           command: test
           args: --features dev
 
+      - name: cargo test --features arbitrary
+        uses: actions-rs/cargo@v1
+        with:
+          command: test
+          args: --features arbitrary
+
   rustfmt:
     name: Rustfmt
     runs-on: ubuntu-latest

Cargo.toml

@@ -22,11 +22,13 @@ include = [
 hashbrown = "0.12.0"
 once_cell = "1.10.0"
 enum-map = { version = "2", optional = true }
+arbitrary = { version = "1", optional = true, features = ["derive"] }
 
 [dev-dependencies]
 serde_json = "1.0.39"
 bencher = "0.1.5"
 proptest = "0.9.1"
+arbtest = "0.1"
 
 [features]
 dev = []

Makefile

@@ -8,3 +8,5 @@ bench:
 	cargo bench --all-features
 test:
 	cargo test --all-features
+test-fuzz:
+	ARBTEST_BUDGET_MS=60000 cargo test --all-features --release

misc/lang.rs.erb

@@ -7,11 +7,9 @@ use std::str::FromStr;
 
 use crate::error::ParseError;
 
-#[cfg(feature = "enum-map")]
-use enum_map::Enum;
-
 /// Represents a language following [ISO 639-3](https://en.wikipedia.org/wiki/ISO_639-3) standard.
-#[cfg_attr(feature = "enum-map", derive(Enum))]
+#[cfg_attr(feature = "enum-map", derive(::enum_map::Enum))]
+#[cfg_attr(feature = "arbitrary", derive(::arbitrary::Arbitrary))]
 #[derive(PartialEq, Eq, Debug, Hash, Clone, Copy)]
 pub enum Lang {
     <% langs.each_with_index do |lang, index| %>

src/core/detector.rs

@@ -28,6 +28,7 @@ use crate::Lang;
 /// let lang = detector.detect_lang("Jen la trinkejo fermitis, ni iras tra mallumo kaj pluvo.");
 /// assert_eq!(lang, Some(Lang::Epo));
 /// ```
+#[cfg_attr(feature = "arbitrary", derive(::arbitrary::Arbitrary))]
 #[derive(Debug, Clone, Default)]
 pub struct Detector {
     options: Options,

src/core/filter_list.rs

@@ -1,5 +1,6 @@
 use crate::Lang;
 
+#[cfg_attr(feature = "arbitrary", derive(::arbitrary::Arbitrary))]
 #[derive(Debug, Clone)]
 pub enum FilterList {
     All,

src/core/method.rs

@@ -2,6 +2,7 @@ use crate::error::ParseError;
 use std::fmt;
 use std::str::FromStr;
 
+#[cfg_attr(feature = "arbitrary", derive(::arbitrary::Arbitrary))]
 #[derive(Debug, Clone, Copy, PartialEq, Eq)]
 pub enum Method {
     Trigram,

src/core/options.rs

@@ -1,5 +1,6 @@
 use super::{FilterList, Method};
 
+#[cfg_attr(feature = "arbitrary", derive(::arbitrary::Arbitrary))]
 #[derive(Debug, Clone)]
 pub struct Options {
     pub(crate) filter_list: FilterList,

src/lang.rs

@@ -7,11 +7,9 @@ use std::str::FromStr;
 
 use crate::error::ParseError;
 
-#[cfg(feature = "enum-map")]
-use enum_map::Enum;
-
 /// Represents a language following [ISO 639-3](https://en.wikipedia.org/wiki/ISO_639-3) standard.
-#[cfg_attr(feature = "enum-map", derive(Enum))]
+#[cfg_attr(feature = "enum-map", derive(::enum_map::Enum))]
+#[cfg_attr(feature = "arbitrary", derive(::arbitrary::Arbitrary))]
 #[derive(PartialEq, Eq, Debug, Hash, Clone, Copy)]
 pub enum Lang {
     /// Esperanto (Esperanto)

src/scripts/script.rs

@@ -5,11 +5,9 @@ use super::lang_mapping;
 use crate::error::ParseError;
 use crate::Lang;
 
-#[cfg(feature = "enum-map")]
-use enum_map::Enum;
-
 /// Represents a writing system (Latin, Cyrillic, Arabic, etc).
-#[cfg_attr(feature = "enum-map", derive(Enum))]
+#[cfg_attr(feature = "enum-map", derive(::enum_map::Enum))]
+#[cfg_attr(feature = "arbitrary", derive(::arbitrary::Arbitrary))]
 #[derive(PartialEq, Eq, Debug, Clone, Copy)]
 pub enum Script {
     // Keep this in alphabetic order (for C bindings)

tests/fuzzing.rs

@@ -0,0 +1,15 @@
+#[cfg(feature = "arbitrary")]
+#[test]
+fn test_fuzzing() {
+    use whatlang::Detector;
+    use ::arbitrary::{Unstructured, Arbitrary};
+
+    fn prop(u: &mut Unstructured) -> ::arbitrary::Result<()> {
+        let detector = Detector::arbitrary(u)?;
+        let input = String::arbitrary(u)?;
+        detector.detect(&input);
+        Ok(())
+    }
+
+    arbtest::builder().run(prop)
+}