Skip to content

Offline, privacy-first grammar checker. Fast, open-source, Rust-powered

fuzz: add example fuzzing targets (#1949)

rnxpykead429ad0d01a
diff --git a/Cargo.lock b/Cargo.lock
index c9a244bc9..479ff0b1a 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -2224,6 +2224,18 @@ dependencies = [
"slab",
]
+[[package]]
+name = "fuzz"
+version = "0.0.0"
+dependencies = [
+ "harper-comments",
+ "harper-core",
+ "harper-html",
+ "harper-literate-haskell",
+ "harper-typst",
+ "libfuzzer-sys",
+]
+
[[package]]
name = "gemm"
version = "0.17.1"
@@ -3370,6 +3382,16 @@ version = "0.2.174"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "1171693293099992e19cddea4e8b849964e9846f4acee11b3948bcc337be8776"
+[[package]]
+name = "libfuzzer-sys"
+version = "0.4.10"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "5037190e1f70cbeef565bd267599242926f724d3b8a9f510fd7e0b540cfa4404"
+dependencies = [
+ "arbitrary",
+ "cc",
+]
+
[[package]]
name = "libloading"
version = "0.8.8"
diff --git a/Cargo.toml b/Cargo.toml
index 75ff6eea2..1ac053674 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -1,5 +1,5 @@
[workspace]
-members = ["harper-cli", "harper-core", "harper-ls", "harper-comments", "harper-wasm", "harper-tree-sitter", "harper-html", "harper-literate-haskell", "harper-typst", "harper-stats", "harper-pos-utils", "harper-brill", "harper-ink", "harper-python", "harper-jjdescription"]
+members = ["harper-cli", "harper-core", "harper-ls", "harper-comments", "harper-wasm", "harper-tree-sitter", "harper-html", "harper-literate-haskell", "harper-typst", "harper-stats", "harper-pos-utils", "harper-brill", "harper-ink", "harper-python", "harper-jjdescription", "fuzz"]
resolver = "2"
[profile.test]
diff --git a/fuzz/.gitignore b/fuzz/.gitignore
new file mode 100644
index 000000000..1a45eee77
--- /dev/null
+++ b/fuzz/.gitignore
@@ -0,0 +1,4 @@
+target
+corpus
+artifacts
+coverage
diff --git a/fuzz/Cargo.toml b/fuzz/Cargo.toml
new file mode 100644
index 000000000..c3631bbae
--- /dev/null
+++ b/fuzz/Cargo.toml
@@ -0,0 +1,52 @@
+[package]
+name = "fuzz"
+version = "0.0.0"
+publish = false
+edition = "2024"
+
+[package.metadata]
+cargo-fuzz = true
+
+[dependencies]
+libfuzzer-sys = "0.4"
+
+harper-core = { path = "../harper-core" }
+harper-typst = { path = "../harper-typst" }
+harper-literate-haskell = { path = "../harper-literate-haskell" }
+harper-html = { path = "../harper-html" }
+harper-comments = { path = "../harper-comments" }
+
+[[bin]]
+name = "fuzz_harper_typst"
+path = "fuzz_targets/fuzz_harper_typst.rs"
+test = false
+doc = false
+bench = false
+
+[[bin]]
+name = "fuzz_harper_literate_haskell"
+path = "fuzz_targets/fuzz_harper_literate_haskell.rs"
+test = false
+doc = false
+bench = false
+
+[[bin]]
+name = "fuzz_harper_html"
+path = "fuzz_targets/fuzz_harper_html.rs"
+test = false
+doc = false
+bench = false
+
+[[bin]]
+name = "fuzz_harper_comment"
+path = "fuzz_targets/fuzz_harper_comment.rs"
+test = false
+doc = false
+bench = false
+
+[[bin]]
+name = "fuzz_harper_core_markdown"
+path = "fuzz_targets/fuzz_harper_core_markdown.rs"
+test = false
+doc = false
+bench = false
diff --git a/fuzz/README.md b/fuzz/README.md
new file mode 100644
index 000000000..b62cf57e6
--- /dev/null
+++ b/fuzz/README.md
@@ -0,0 +1,34 @@
+# cargo-fuzz targets
+
+## Setup
+
+Follow the rust-fuzz [setup guide](https://rust-fuzz.github.io/book/cargo-fuzz/setup.html).
+You need a nightly toolchain and the cargo-fuzz plugin.
+
+Simple installation steps:
+
+- `rustup install nightly`
+- `cargo install cargo-fuzz`
+
+## Adding a new fuzzing target
+
+To add a new target, run `cargo fuzz add $TARGET_NAME`
+
+## Doing a fuzzing run
+
+If possible, prefill the `fuzz/corpus/$TARGET_NAME` directory with appropriate examples to speed up fuzzing.
+The fuzzer should be coverage aware, so providing a well formed input document to fuzzing targets only expecting a string as input can speed things up a lot.
+
+Then, run `cargo +nightly fuzz run $TARGET_NAME -- -timeout=$TIMEOUT`
+
+The timeout flag accepts a timeout in seconds, after which a long-running test case will be aborted.
+This should be set to a low number to quickly report endless loops / deep recursion in parsers.
+
+The normal fuzzing run will continue until a crash is found.
+
+Alternatively, if you want to run all the fuzzing targets at once: `cargo +nightly fuzz list | parallel -j0 cargo +nightly fuzz run {} -- -timeout=$TIMEOUT`
+
+## Minifying a test case
+
+Once the fuzzer finds a crash, we probably want to minify the result.
+This can be done with `CARGO_PROFILE_RELEASE_LTO=false cargo +nightly fuzz tmin $TARGET $TEST_CASE_PATH`
diff --git a/fuzz/fuzz_targets/fuzz_harper_comment.rs b/fuzz/fuzz_targets/fuzz_harper_comment.rs
new file mode 100644
index 000000000..b59045796
--- /dev/null
+++ b/fuzz/fuzz_targets/fuzz_harper_comment.rs
@@ -0,0 +1,76 @@
+#![no_main]
+
+use harper_core::parsers::{MarkdownOptions, StrParser};
+use libfuzzer_sys::arbitrary::{Arbitrary, Result, Unstructured};
+use libfuzzer_sys::fuzz_target;
+
+#[derive(Debug)]
+struct Language(String);
+
+const LANGUAGES: [&str; 32] = [
+ "cmake",
+ "cpp",
+ "csharp",
+ "c",
+ "dart",
+ "go",
+ "haskell",
+ "javascriptreact",
+ "javascript",
+ "java",
+ "kotlin",
+ "lua",
+ "nix",
+ "php",
+ "python",
+ "ruby",
+ "rust",
+ "scala",
+ "shellscript",
+ "solidity",
+ "swift",
+ "toml",
+ "typescriptreact",
+ "typescript",
+ "clojure",
+ "go",
+ "lua",
+ "java",
+ "javascriptreact",
+ "typescript",
+ "typescriptreact",
+ "solidity",
+];
+
+impl<'a> Arbitrary<'a> for Language {
+ fn arbitrary(u: &mut Unstructured<'a>) -> Result<Self> {
+ let &lang = u.choose(&LANGUAGES)?;
+ Ok(Language(lang.to_owned()))
+ }
+}
+
+#[derive(Debug)]
+struct Input {
+ language: Language,
+ text: String,
+}
+
+impl<'a> Arbitrary<'a> for Input {
+ fn arbitrary(u: &mut Unstructured<'a>) -> Result<Self> {
+ let (language, text) = Arbitrary::arbitrary(u)?;
+ Ok(Input { language, text })
+ }
+
+ fn arbitrary_take_rest(u: Unstructured<'a>) -> Result<Self> {
+ let (language, text) = Arbitrary::arbitrary_take_rest(u)?;
+ Ok(Input { language, text })
+ }
+}
+
+fuzz_target!(|data: Input| {
+ let opts = MarkdownOptions::default();
+ let parser = harper_comments::CommentParser::new_from_language_id(&data.language.0, opts);
+ if let Some(parser) = parser {
+ let _res = parser.parse_str(&data.text);
+ }
+});
diff --git a/fuzz/fuzz_targets/fuzz_harper_core_markdown.rs b/fuzz/fuzz_targets/fuzz_harper_core_markdown.rs
new file mode 100644
index 000000000..596a269d7
--- /dev/null
+++ b/fuzz/fuzz_targets/fuzz_harper_core_markdown.rs
@@ -0,0 +1,10 @@
+#![no_main]
+
+use harper_core::parsers::{Markdown, MarkdownOptions, StrParser};
+use libfuzzer_sys::fuzz_target;
+
+fuzz_target!(|data: &str| {
+ let opts = MarkdownOptions::default();
+ let parser = Markdown::new(opts);
+ let _res = parser.parse_str(data);
+});
diff --git a/fuzz/fuzz_targets/fuzz_harper_html.rs b/fuzz/fuzz_targets/fuzz_harper_html.rs
new file mode 100644
index 000000000..b0f11cf63
--- /dev/null
+++ b/fuzz/fuzz_targets/fuzz_harper_html.rs
@@ -0,0 +1,9 @@
+#![no_main]
+
+use harper_core::parsers::StrParser;
+use libfuzzer_sys::fuzz_target;
+
+fuzz_target!(|data: &str| {
+ let parser = harper_html::HtmlParser::default();
+ let _res = parser.parse_str(data);
+});
diff --git a/fuzz/fuzz_targets/fuzz_harper_literate_haskell.rs b/fuzz/fuzz_targets/fuzz_harper_literate_haskell.rs
new file mode 100644
index 000000000..ea3df937a
--- /dev/null
+++ b/fuzz/fuzz_targets/fuzz_harper_literate_haskell.rs
@@ -0,0 +1,9 @@
+#![no_main]
+
+// use harper_core::parsers::StrParser;
+use libfuzzer_sys::fuzz_target;
+
+fuzz_target!(|_data: &str| {
+ // TODO: figure out how to create a literate haskell parser
+ // let _res = typst.parse_str(&data);
+});
diff --git a/fuzz/fuzz_targets/fuzz_harper_typst.rs b/fuzz/fuzz_targets/fuzz_harper_typst.rs
new file mode 100644
index 000000000..166282c68
--- /dev/null
+++ b/fuzz/fuzz_targets/fuzz_harper_typst.rs
@@ -0,0 +1,9 @@
+#![no_main]
+
+use harper_core::parsers::StrParser;
+use libfuzzer_sys::fuzz_target;
+
+fuzz_target!(|data: &str| {
+ let typst = harper_typst::Typst;
+ let _res = typst.parse_str(data);
+});